From dd5f19d133441c281cccd353f321cf682edf532e Mon Sep 17 00:00:00 2001 From: Michael Ekstrand Date: Fri, 30 May 2025 18:00:31 -0400 Subject: [PATCH 01/59] move stats into subpackage --- src/lenskit/stats/__init__.py | 13 +++++++ src/lenskit/{stats.py => stats/_gini.py} | 43 +-------------------- src/lenskit/stats/_topn.py | 48 ++++++++++++++++++++++++ 3 files changed, 62 insertions(+), 42 deletions(-) create mode 100644 src/lenskit/stats/__init__.py rename src/lenskit/{stats.py => stats/_gini.py} (62%) create mode 100644 src/lenskit/stats/_topn.py diff --git a/src/lenskit/stats/__init__.py b/src/lenskit/stats/__init__.py new file mode 100644 index 000000000..bf7974185 --- /dev/null +++ b/src/lenskit/stats/__init__.py @@ -0,0 +1,13 @@ +# This file is part of LensKit. +# Copyright (C) 2018-2023 Boise State University. +# Copyright (C) 2023-2025 Drexel University. +# Licensed under the MIT license, see LICENSE.md for details. +# SPDX-License-Identifier: MIT + +from ._gini import gini +from ._topn import argtopn + +__all__ = [ + "gini", + "argtopn", +] diff --git a/src/lenskit/stats.py b/src/lenskit/stats/_gini.py similarity index 62% rename from src/lenskit/stats.py rename to src/lenskit/stats/_gini.py index 6b66a59cc..b456beb2f 100644 --- a/src/lenskit/stats.py +++ b/src/lenskit/stats/_gini.py @@ -4,10 +4,6 @@ # Licensed under the MIT license, see LICENSE.md for details. # SPDX-License-Identifier: MIT -""" -LensKit statistical computations. -""" - from __future__ import annotations import warnings @@ -15,7 +11,6 @@ import numpy as np from numpy.typing import ArrayLike -from lenskit.data.types import NPVector from lenskit.diagnostics import DataWarning @@ -59,40 +54,4 @@ def gini(xs: ArrayLike) -> float: warnings.warn( "Gini coefficient is not defined for non-positive totals", DataWarning, stacklevel=2 ) - return max(num / denom, 0) - - -def argtopn(xs: ArrayLike, n: int) -> NPVector[np.int64]: - """ - Compute the ordered positions of the top *n* elements. Similar to - :func:`torch.topk`, but works with NumPy arrays and only returns the - indices. - - .. deprecated:: 2025.3.0 - - This was never declared stable, but is now deprecated and will be - removed in 2026.1. - """ - if n == 0: - return np.empty(0, np.int64) - - xs = np.asarray(xs) - - N = len(xs) - invalid = np.isnan(xs) - if np.any(invalid): - mask = ~invalid - vxs = xs[mask] - remap = np.arange(N)[mask] - res = argtopn(vxs, n) - return remap[res] - - if n >= 0 and n < N: - parts = np.argpartition(-xs, n) - top_scores = xs[parts[:n]] - top_sort = np.argsort(-top_scores) - order = parts[top_sort] - else: - order = np.argsort(-xs) - - return order + return max(num / denom, 0.0) diff --git a/src/lenskit/stats/_topn.py b/src/lenskit/stats/_topn.py new file mode 100644 index 000000000..d6fa20159 --- /dev/null +++ b/src/lenskit/stats/_topn.py @@ -0,0 +1,48 @@ +# This file is part of LensKit. +# Copyright (C) 2018-2023 Boise State University. +# Copyright (C) 2023-2025 Drexel University. +# Licensed under the MIT license, see LICENSE.md for details. +# SPDX-License-Identifier: MIT + +from __future__ import annotations + +import numpy as np +from numpy.typing import ArrayLike + +from lenskit.data.types import NPVector + + +def argtopn(xs: ArrayLike, n: int) -> NPVector[np.int64]: + """ + Compute the ordered positions of the top *n* elements. Similar to + :func:`torch.topk`, but works with NumPy arrays and only returns the + indices. + + .. deprecated:: 2025.3.0 + + This was never declared stable, but is now deprecated and will be + removed in 2026.1. + """ + if n == 0: + return np.empty(0, np.int64) + + xs = np.asarray(xs) + + N = len(xs) + invalid = np.isnan(xs) + if np.any(invalid): + mask = ~invalid + vxs = xs[mask] + remap = np.arange(N)[mask] + res = argtopn(vxs, n) + return remap[res] # type: ignore + + if n >= 0 and n < N: + parts = np.argpartition(-xs, n) + top_scores = xs[parts[:n]] + top_sort = np.argsort(-top_scores) + order = parts[top_sort] + else: + order = np.argsort(-xs) + + return order # type: ignore From e3b8d65c60df7789c6dd19e30875d4dda05b9e07 Mon Sep 17 00:00:00 2001 From: Michael Ekstrand Date: Fri, 30 May 2025 18:01:03 -0400 Subject: [PATCH 02/59] rename stats tests --- tests/{math => stats}/__init__.py | 0 tests/{math => stats}/test_argtopn.py | 0 tests/{math => stats}/test_gini.py | 0 3 files changed, 0 insertions(+), 0 deletions(-) rename tests/{math => stats}/__init__.py (100%) rename tests/{math => stats}/test_argtopn.py (100%) rename tests/{math => stats}/test_gini.py (100%) diff --git a/tests/math/__init__.py b/tests/stats/__init__.py similarity index 100% rename from tests/math/__init__.py rename to tests/stats/__init__.py diff --git a/tests/math/test_argtopn.py b/tests/stats/test_argtopn.py similarity index 100% rename from tests/math/test_argtopn.py rename to tests/stats/test_argtopn.py diff --git a/tests/math/test_gini.py b/tests/stats/test_gini.py similarity index 100% rename from tests/math/test_gini.py rename to tests/stats/test_gini.py From 80ef1aa654df962d048def3836c7652d70af2db3 Mon Sep 17 00:00:00 2001 From: Michael Ekstrand Date: Fri, 30 May 2025 18:09:15 -0400 Subject: [PATCH 03/59] start on a test --- tests/stats/test_blb.py | 14 ++++++++++++++ 1 file changed, 14 insertions(+) create mode 100644 tests/stats/test_blb.py diff --git a/tests/stats/test_blb.py b/tests/stats/test_blb.py new file mode 100644 index 000000000..d2c0807bf --- /dev/null +++ b/tests/stats/test_blb.py @@ -0,0 +1,14 @@ +# This file is part of LensKit. +# Copyright (C) 2018-2023 Boise State University. +# Copyright (C) 2023-2025 Drexel University. +# Licensed under the MIT license, see LICENSE.md for details. +# SPDX-License-Identifier: MIT + +import hypothesis.extra.numpy as nph +import hypothesis.strategies as st +from hypothesis import given + + +@given(nph.arrays(shape=st.integers(10000, 1_000_000), dtype=nph.floating_dtypes(endianness="="))) +def test_blb_array(): + pass From e83c97bb0bd3cd8b6a2c57b5d488843e0edcfd31 Mon Sep 17 00:00:00 2001 From: Michael Ekstrand Date: Mon, 2 Jun 2025 09:38:54 -0400 Subject: [PATCH 04/59] define exports for lenskit.random --- src/lenskit/random.py | 17 +++++++++++++++++ 1 file changed, 17 insertions(+) diff --git a/src/lenskit/random.py b/src/lenskit/random.py index 35015e1b1..d66af3297 100644 --- a/src/lenskit/random.py +++ b/src/lenskit/random.py @@ -25,6 +25,23 @@ if TYPE_CHECKING: # avoid circular import from lenskit.data import RecQuery +__all__ = [ + "Generator", + "SeedLike", + "RNGLike", + "RNGInput", + "ConfiguredSeed", + "SeedDependency", + "DerivableSeed", + "load_seed", + "set_global_rng", + "init_global_rng", + "random_generator", + "make_seed", + "RNGFactory", + "derivable_rng", +] + SeedLike: TypeAlias = int | Sequence[int] | np.random.SeedSequence """ Type for RNG seeds (see `SPEC 7`_). From d20b438c58ba7a6c5c07277b89835c181ad0f14d Mon Sep 17 00:00:00 2001 From: Michael Ekstrand Date: Mon, 2 Jun 2025 10:35:17 -0400 Subject: [PATCH 05/59] first pass at implementing BLB --- src/lenskit/stats/__init__.py | 2 + src/lenskit/stats/_blb.py | 107 ++++++++++++++++++++++++++++++++++ tests/stats/test_blb.py | 78 +++++++++++++++++++++++-- 3 files changed, 183 insertions(+), 4 deletions(-) create mode 100644 src/lenskit/stats/_blb.py diff --git a/src/lenskit/stats/__init__.py b/src/lenskit/stats/__init__.py index bf7974185..1ddfa5e09 100644 --- a/src/lenskit/stats/__init__.py +++ b/src/lenskit/stats/__init__.py @@ -4,10 +4,12 @@ # Licensed under the MIT license, see LICENSE.md for details. # SPDX-License-Identifier: MIT +from ._blb import blb_summary from ._gini import gini from ._topn import argtopn __all__ = [ "gini", "argtopn", + "blb_summary", ] diff --git a/src/lenskit/stats/_blb.py b/src/lenskit/stats/_blb.py new file mode 100644 index 000000000..826d72a70 --- /dev/null +++ b/src/lenskit/stats/_blb.py @@ -0,0 +1,107 @@ +# This file is part of LensKit. +# Copyright (C) 2018-2023 Boise State University. +# Copyright (C) 2023-2025 Drexel University. +# Licensed under the MIT license, see LICENSE.md for details. +# SPDX-License-Identifier: MIT + +from __future__ import annotations + +import warnings +from collections.abc import Iterable +from typing import Any, Literal, Protocol, TypeAlias, TypeVar + +import numpy as np +from numpy.typing import NDArray + +from lenskit.diagnostics import DataWarning +from lenskit.random import Generator, RNGInput, random_generator + +F = TypeVar("F", bound=np.floating, covariant=True) + +SummaryStat: TypeAlias = Literal["mean"] + +# dummy assignment to typecheck that we have correctly typed weighted average +__dummy_avg: WeightedStatistic = np.average + + +class WeightedStatistic(Protocol): + """ + Callable interface for weighted statistics, required by the Bag of Little Bootstraps. + """ + + def __call__( + self, + a: NDArray[np.floating[Any]], + /, + *, + weights: NDArray[np.floating[Any] | np.integer[Any]] | None = None, + ) -> np.floating[Any]: ... + + +def blb_summary( + xs: NDArray[F], + stat: SummaryStat, + *, + s: int = 10, + r: int = 100, + b_factor: float = 0.6, + rng: RNGInput = None, +) -> dict[str, float]: + """ + Summarize one or more statistics using the Bag of Little Bootstraps :cite:p:`blb`. + """ + if stat != "mean": + raise ValueError(f"unsupported statistic {stat}") + + mask = np.isfinite(xs) + if ninf := int(np.sum(~mask)): + warnings.warn(f"ignoring {ninf} nonfinite values", DataWarning, stacklevel=2) + + xs = xs[mask] + est = np.average(xs).item() + n = len(xs) + b = int(n**b_factor) + + rng = random_generator(rng) + + ss_summaries = {} + for ss in _blb_subsets(xs, s, b, rng=rng): + ss_sum = _miniboot_ss(xs, ss, np.average, r, rng=rng) + _accum_summaries(ss_sum, dest=ss_summaries) + + return {"value": est} | {n: np.mean(xs) for n, xs in ss_summaries.items()} + + +def _blb_subsets(xs: NDArray[F], s: int, b: int, *, rng: Generator) -> Iterable[NDArray[np.int64]]: + for i in range(s): + yield rng.choice(len(xs), b, replace=False) + + +def _accum_summaries(values: dict[str, float], *, dest: dict[str, list[float]]): + for name, value in values.items(): + vs = dest.setdefault(name, []) + vs.append(value) + + +def _miniboot_ss( + xs: NDArray[F], ss: NDArray[np.int64], stat: WeightedStatistic, r: int, *, rng: Generator +) -> dict[str, float]: + b = len(ss) + n = len(xs) + xss = xs[ss] + + flat = np.full(b, 1.0 / b) + vals = [_miniboot_sample_stat(n, xss, flat, stat, rng) for _j in range(r)] + vals = np.array(vals) + mean = np.mean(vals).item() + lo, hi = np.quantile(vals, [0.025, 0.975]) + return {"mean": mean, "low": lo, "high": hi} + + +def _miniboot_sample_stat( + n: int, xss: NDArray[F], flat: NDArray[np.float64], stat: WeightedStatistic, rng: Generator +) -> float: + weights = rng.multinomial(n, flat) + assert weights.shape == (len(flat),) + assert np.sum(weights) == n + return stat(xss, weights=weights).item() diff --git a/tests/stats/test_blb.py b/tests/stats/test_blb.py index d2c0807bf..e04766799 100644 --- a/tests/stats/test_blb.py +++ b/tests/stats/test_blb.py @@ -4,11 +4,81 @@ # Licensed under the MIT license, see LICENSE.md for details. # SPDX-License-Identifier: MIT +from math import sqrt + +import numpy as np +from numpy.typing import NDArray + import hypothesis.extra.numpy as nph import hypothesis.strategies as st -from hypothesis import given +from hypothesis import assume, given +from pytest import approx, mark, warns + +from lenskit.data.types import NPVector +from lenskit.diagnostics import DataWarning +from lenskit.random import random_generator +from lenskit.stats import blb_summary + + +@given( + st.integers(1000, 1_000_000), + nph.floating_dtypes(endianness="="), + st.integers(5, 20), + st.integers(50, 300), + st.integers(0), +) +@mark.filterwarnings(r"error:.*ignoring \d+ nonfinite values") +def test_blb_array_normal(n, dtype, s: int, r: int, seed): + "Test BLB with arrays of standard normals." + rng = random_generator(seed) + xs = rng.standard_normal(n).astype(dtype) + mean = np.mean(xs) + n = len(xs) + std = np.std(xs) + ste = std / sqrt(n) + + summary = blb_summary(xs, "mean", s=s, r=r, rng=rng) + assert isinstance(summary, dict) + assert summary["value"] == approx(mean) + assert summary["mean"] == approx(mean, rel=0.01) + + assert summary["low"] == approx(mean - 1.96 * ste, rel=0.01) + assert summary["high"] == approx(mean + 1.96 * ste, rel=0.01) + + +@given( + nph.arrays(shape=st.integers(10000, 1_000_000), dtype=nph.floating_dtypes(endianness="=")), + st.integers(5, 20), + st.integers(50, 300), + st.integers(0), +) +def test_blb_array(xs: NDArray[np.floating], s: int, r: int, seed: int): + "Test BLB with more aggressive edge-case hunting." + xsf = xs[np.isfinite(xs)] + mean = np.mean(xsf) + # ignore grotesquely out-of-bounds cases (for now) + assume(np.isfinite(mean)) + n = len(xsf) + std = np.std(xsf) + ste = std / sqrt(n) + + if np.all(np.isfinite(xs)): + summary = blb_summary(xs, "mean", s=s, r=r, rng=seed) + else: + with warns(DataWarning, match=r"ignoring \d+ nonfinite"): + summary = blb_summary(xs, "mean", s=s, r=r, rng=seed) + assert isinstance(summary, dict) + assert summary["value"] == approx(mean, nan_ok=True) + assert summary["mean"] == approx(mean, rel=0.01, nan_ok=True) -@given(nph.arrays(shape=st.integers(10000, 1_000_000), dtype=nph.floating_dtypes(endianness="="))) -def test_blb_array(): - pass + if n == 0: + assert np.isnan(summary["low"]) + assert np.isnan(summary["high"]) + elif np.allclose(xs, np.min(xs)): + # standard error is zero + assert summary["low"] == approx(mean, rel=0.01) + assert summary["high"] == approx(mean, rel=0.01) + else: + assert summary["low"] == approx(mean - 1.96 * ste, rel=0.01) + assert summary["high"] == approx(mean + 1.96 * ste, rel=0.01) From 1a6468b3cebaa34d945d1c0b823e7c1dcfa61f30 Mon Sep 17 00:00:00 2001 From: Michael Ekstrand Date: Mon, 2 Jun 2025 10:44:44 -0400 Subject: [PATCH 06/59] add functions --- tests/stats/test_blb.py | 14 +++++++------- 1 file changed, 7 insertions(+), 7 deletions(-) diff --git a/tests/stats/test_blb.py b/tests/stats/test_blb.py index e04766799..0472fc27d 100644 --- a/tests/stats/test_blb.py +++ b/tests/stats/test_blb.py @@ -23,8 +23,8 @@ @given( st.integers(1000, 1_000_000), nph.floating_dtypes(endianness="="), - st.integers(5, 20), - st.integers(50, 300), + st.integers(10, 20), + st.integers(100, 300), st.integers(0), ) @mark.filterwarnings(r"error:.*ignoring \d+ nonfinite values") @@ -40,16 +40,16 @@ def test_blb_array_normal(n, dtype, s: int, r: int, seed): summary = blb_summary(xs, "mean", s=s, r=r, rng=rng) assert isinstance(summary, dict) assert summary["value"] == approx(mean) - assert summary["mean"] == approx(mean, rel=0.01) + assert summary["mean"] == approx(mean, rel=0.05) - assert summary["low"] == approx(mean - 1.96 * ste, rel=0.01) - assert summary["high"] == approx(mean + 1.96 * ste, rel=0.01) + assert summary["low"] == approx(mean - 1.96 * ste, rel=0.05) + assert summary["high"] == approx(mean + 1.96 * ste, rel=0.05) @given( nph.arrays(shape=st.integers(10000, 1_000_000), dtype=nph.floating_dtypes(endianness="=")), - st.integers(5, 20), - st.integers(50, 300), + st.integers(10, 20), + st.integers(100, 300), st.integers(0), ) def test_blb_array(xs: NDArray[np.floating], s: int, r: int, seed: int): From 1a59f6f96e3f7103f307f975193c3b771a0af1ab Mon Sep 17 00:00:00 2001 From: Michael Ekstrand Date: Mon, 2 Jun 2025 11:09:21 -0400 Subject: [PATCH 07/59] don't export Generator --- src/lenskit/random.py | 1 - 1 file changed, 1 deletion(-) diff --git a/src/lenskit/random.py b/src/lenskit/random.py index d66af3297..160ab7fb3 100644 --- a/src/lenskit/random.py +++ b/src/lenskit/random.py @@ -26,7 +26,6 @@ from lenskit.data import RecQuery __all__ = [ - "Generator", "SeedLike", "RNGLike", "RNGInput", From 98bca5ba7c6829a2cfdf3ff29c0ea52a8cb8bcde Mon Sep 17 00:00:00 2001 From: Michael Ekstrand Date: Mon, 2 Jun 2025 11:37:00 -0400 Subject: [PATCH 08/59] refactor BLB and simplify tests --- src/lenskit/stats/_blb.py | 195 ++++++++++++++++++++++++++++++-------- tests/stats/test_blb.py | 52 ++++++---- 2 files changed, 186 insertions(+), 61 deletions(-) diff --git a/src/lenskit/stats/_blb.py b/src/lenskit/stats/_blb.py index 826d72a70..f833e9845 100644 --- a/src/lenskit/stats/_blb.py +++ b/src/lenskit/stats/_blb.py @@ -7,14 +7,14 @@ from __future__ import annotations import warnings -from collections.abc import Iterable -from typing import Any, Literal, Protocol, TypeAlias, TypeVar +from typing import Any, ClassVar, Literal, Protocol, TypeAlias, TypedDict, TypeVar import numpy as np +import pandas as pd from numpy.typing import NDArray from lenskit.diagnostics import DataWarning -from lenskit.random import Generator, RNGInput, random_generator +from lenskit.random import RNGInput, random_generator F = TypeVar("F", bound=np.floating, covariant=True) @@ -42,8 +42,10 @@ def blb_summary( xs: NDArray[F], stat: SummaryStat, *, - s: int = 10, - r: int = 100, + ci_width: float = 0.95, + tol: float = 0.05, + s_w: int = 3, + r_w: int = 20, b_factor: float = 0.6, rng: RNGInput = None, ) -> dict[str, float]: @@ -59,49 +61,160 @@ def blb_summary( xs = xs[mask] est = np.average(xs).item() - n = len(xs) - b = int(n**b_factor) rng = random_generator(rng) + bootstrapper = _BLBootstrapper(np.average, ci_width, tol, s_w, r_w, b_factor, rng) - ss_summaries = {} - for ss in _blb_subsets(xs, s, b, rng=rng): - ss_sum = _miniboot_ss(xs, ss, np.average, r, rng=rng) - _accum_summaries(ss_sum, dest=ss_summaries) + boot_df = bootstrapper.summarize(xs) - return {"value": est} | {n: np.mean(xs) for n, xs in ss_summaries.items()} + return {"value": est} | boot_df.agg("mean").to_dict() -def _blb_subsets(xs: NDArray[F], s: int, b: int, *, rng: Generator) -> Iterable[NDArray[np.int64]]: - for i in range(s): - yield rng.choice(len(xs), b, replace=False) +class _BootResult(TypedDict): + mean: float + ci_min: float + ci_max: float + count: int -def _accum_summaries(values: dict[str, float], *, dest: dict[str, list[float]]): - for name, value in values.items(): - vs = dest.setdefault(name, []) - vs.append(value) +class _BLBootstrapper: + """ + Implementation of BLB computation. + """ + statistic: WeightedStatistic + ci_width: float -def _miniboot_ss( - xs: NDArray[F], ss: NDArray[np.int64], stat: WeightedStatistic, r: int, *, rng: Generator -) -> dict[str, float]: - b = len(ss) - n = len(xs) - xss = xs[ss] - - flat = np.full(b, 1.0 / b) - vals = [_miniboot_sample_stat(n, xss, flat, stat, rng) for _j in range(r)] - vals = np.array(vals) - mean = np.mean(vals).item() - lo, hi = np.quantile(vals, [0.025, 0.975]) - return {"mean": mean, "low": lo, "high": hi} - - -def _miniboot_sample_stat( - n: int, xss: NDArray[F], flat: NDArray[np.float64], stat: WeightedStatistic, rng: Generator -) -> float: - weights = rng.multinomial(n, flat) - assert weights.shape == (len(flat),) - assert np.sum(weights) == n - return stat(xss, weights=weights).item() + tolerance: float + s_window: int + r_window: int + b_factor: float + rng: np.random.Generator + + def __init__( + self, + stat: WeightedStatistic, + ci_width: float, + tol: float, + s_w: int, + r_w: int, + b_factor: float, + rng: np.random.Generator, + ): + self.statistic = stat + self.ci_width = ci_width + self.tolerance = tol + self.s_window = s_w + self.r_window = r_w + self.b_factor = b_factor + self.rng = rng + self.ss_stats = {} + + def summarize(self, xs: NDArray[F]): + results = [] + means = StatAccum(self.tolerance, self.s_window) + + count = 0 + for ss in self.blb_subsets(xs): + count += 1 + res = self.measure_subset(xs, ss) + results.append(res) + means.record(res["mean"]) + if means.converged(): + break + + return pd.DataFrame.from_records(results) + + def blb_subsets(self, xs: NDArray[F]): + b = int(len(xs) ** self.b_factor) + + while True: + yield self.rng.choice(len(xs), b, replace=False) + + def measure_subset(self, xs: NDArray[F], ss: NDArray[np.int64]) -> _BootResult: + b = len(ss) + n = len(xs) + xss = xs[ss] + + acc = StatAccum(self.tolerance, self.r_window) + + count = 0 + for weights in self.miniboot_weights(n, b): + count += 1 + stat = self.statistic(xss, weights=weights) + acc.record(stat) + if acc.converged(): + break + + [lo, hi] = np.quantile(acc.values, [0.025, 0.975]) + return {"mean": np.mean(acc.values).item(), "ci_min": lo, "ci_max": hi, "count": count} + + def miniboot_weights(self, n: int, b: int): + flat = np.full(b, 1.0 / b) + + while True: + yield self.rng.multinomial(n, flat) + + +class StatAccum: + INIT_SIZE: ClassVar[int] = 100 + ABS_TOL: ClassVar[float] = 1.0e-12 + + tolerance: float + window: int + + _len: int = 0 + _values: NDArray[np.float64] + _cum_means: NDArray[np.float64] + + def __init__(self, tol: float, w: int): + self.tolerance = tol + self.window = w + + self._values = np.zeros(self.INIT_SIZE) + self._cum_means = np.zeros(self.INIT_SIZE) + + @property + def values(self) -> NDArray[np.float64]: + return self._values[: self._len] + + def record(self, x: float | np.floating[Any]) -> None: + "Record a new value in the accumulator." + self._expand_if_needed() + i = self._len + self._len += 1 + + # record and update the cumulative mean + self._values[i] = x + self._cum_means[i] = np.mean(self.values) + + def mean(self) -> float | None: + "Get the mean of the accumulated values." + if self._len > 0: + return self._cum_means[self._len - 1] + else: + return None + + def converged(self) -> bool: + """ + Check for convergence. + """ + if self._len < self.window: + return False + + i_cur = self._len - 1 + i_start = self._len - self.window + current = self._cum_means[i_cur] + + # lower-bound tolerance for very small values + atol = max(current * self.tolerance, self.ABS_TOL) + + window = self._cum_means[i_start : self._len] + gaps = np.abs(window - current) + return np.all(gaps <= atol).item() + + def _expand_if_needed(self): + cap = len(self._values) + if cap == self._len: + self._values.resize(cap * 2) + self._cum_means.resize(cap * 2) diff --git a/tests/stats/test_blb.py b/tests/stats/test_blb.py index 0472fc27d..cbbc82de7 100644 --- a/tests/stats/test_blb.py +++ b/tests/stats/test_blb.py @@ -20,39 +20,51 @@ from lenskit.stats import blb_summary +def test_blb_single_array(rng: np.random.Generator): + "Quick one-array test to fail fast" + xs = rng.standard_normal(40_000) + 1.0 + mean = np.mean(xs) + ste = np.std(xs) / 200 + + summary = blb_summary(xs, "mean", rng=rng) + print(summary) + assert isinstance(summary, dict) + assert summary["value"] == approx(mean) + assert summary["mean"] == approx(mean, rel=0.05) + + assert summary["ci_min"] == approx(mean - 1.96 * ste, rel=0.05) + assert summary["ci_max"] == approx(mean + 1.96 * ste, rel=0.05) + + @given( st.integers(1000, 1_000_000), nph.floating_dtypes(endianness="="), - st.integers(10, 20), - st.integers(100, 300), st.integers(0), ) @mark.filterwarnings(r"error:.*ignoring \d+ nonfinite values") -def test_blb_array_normal(n, dtype, s: int, r: int, seed): - "Test BLB with arrays of standard normals." +def test_blb_array_normal(n, dtype, seed): + "Test BLB with arrays of normals." rng = random_generator(seed) - xs = rng.standard_normal(n).astype(dtype) + xs = rng.normal(1.0, 1.0, n).astype(dtype) mean = np.mean(xs) n = len(xs) std = np.std(xs) ste = std / sqrt(n) - summary = blb_summary(xs, "mean", s=s, r=r, rng=rng) + summary = blb_summary(xs, "mean", rng=rng) assert isinstance(summary, dict) assert summary["value"] == approx(mean) - assert summary["mean"] == approx(mean, rel=0.05) + assert summary["mean"] == approx(mean, rel=0.075) - assert summary["low"] == approx(mean - 1.96 * ste, rel=0.05) - assert summary["high"] == approx(mean + 1.96 * ste, rel=0.05) + assert summary["ci_min"] == approx(mean - 1.96 * ste, rel=0.075) + assert summary["ci_max"] == approx(mean + 1.96 * ste, rel=0.075) @given( nph.arrays(shape=st.integers(10000, 1_000_000), dtype=nph.floating_dtypes(endianness="=")), - st.integers(10, 20), - st.integers(100, 300), st.integers(0), ) -def test_blb_array(xs: NDArray[np.floating], s: int, r: int, seed: int): +def test_blb_array(xs: NDArray[np.floating], seed: int): "Test BLB with more aggressive edge-case hunting." xsf = xs[np.isfinite(xs)] mean = np.mean(xsf) @@ -63,22 +75,22 @@ def test_blb_array(xs: NDArray[np.floating], s: int, r: int, seed: int): ste = std / sqrt(n) if np.all(np.isfinite(xs)): - summary = blb_summary(xs, "mean", s=s, r=r, rng=seed) + summary = blb_summary(xs, "mean", rng=seed) else: with warns(DataWarning, match=r"ignoring \d+ nonfinite"): - summary = blb_summary(xs, "mean", s=s, r=r, rng=seed) + summary = blb_summary(xs, "mean", rng=seed) assert isinstance(summary, dict) assert summary["value"] == approx(mean, nan_ok=True) assert summary["mean"] == approx(mean, rel=0.01, nan_ok=True) if n == 0: - assert np.isnan(summary["low"]) - assert np.isnan(summary["high"]) + assert np.isnan(summary["ci_min"]) + assert np.isnan(summary["ci_max"]) elif np.allclose(xs, np.min(xs)): # standard error is zero - assert summary["low"] == approx(mean, rel=0.01) - assert summary["high"] == approx(mean, rel=0.01) + assert summary["ci_min"] == approx(mean, rel=0.01) + assert summary["ci_max"] == approx(mean, rel=0.01) else: - assert summary["low"] == approx(mean - 1.96 * ste, rel=0.01) - assert summary["high"] == approx(mean + 1.96 * ste, rel=0.01) + assert summary["ci_min"] == approx(mean - 1.96 * ste, rel=0.01) + assert summary["ci_max"] == approx(mean + 1.96 * ste, rel=0.01) From e79573ee0be1449670b5f45f7a00e48baa73075a Mon Sep 17 00:00:00 2001 From: Michael Ekstrand Date: Mon, 2 Jun 2025 11:40:52 -0400 Subject: [PATCH 09/59] rearrange parameters --- src/lenskit/stats/_blb.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/lenskit/stats/_blb.py b/src/lenskit/stats/_blb.py index f833e9845..4fbdd27a3 100644 --- a/src/lenskit/stats/_blb.py +++ b/src/lenskit/stats/_blb.py @@ -43,10 +43,10 @@ def blb_summary( stat: SummaryStat, *, ci_width: float = 0.95, + b_factor: float = 0.6, tol: float = 0.05, s_w: int = 3, r_w: int = 20, - b_factor: float = 0.6, rng: RNGInput = None, ) -> dict[str, float]: """ From c29f2f3780614b987474299c0bc7fc02364b3bf6 Mon Sep 17 00:00:00 2001 From: Michael Ekstrand Date: Mon, 2 Jun 2025 17:05:55 -0400 Subject: [PATCH 10/59] BLB seems to work --- notebooks/BLB.ipynb | 853 +++++++++++++++++++++++++++++++++++++ src/lenskit/stats/_blb.py | 164 ++++--- src/lenskit/stats/_topn.py | 5 +- 3 files changed, 957 insertions(+), 65 deletions(-) create mode 100644 notebooks/BLB.ipynb diff --git a/notebooks/BLB.ipynb b/notebooks/BLB.ipynb new file mode 100644 index 000000000..2f9a4d355 --- /dev/null +++ b/notebooks/BLB.ipynb @@ -0,0 +1,853 @@ +{ + "cells": [ + { + "cell_type": "markdown", + "id": "9e264225", + "metadata": {}, + "source": [ + "# Bag of Little Bootstraps analysis\n", + "\n", + "This notebook inspects our Bag of Little Bootstraps implementation to see how it is doing." + ] + }, + { + "cell_type": "markdown", + "id": "c3210118", + "metadata": {}, + "source": [ + "## Setup\n", + "\n", + "Load libraries:" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "id": "115b4f9e", + "metadata": {}, + "outputs": [], + "source": [ + "import matplotlib.pyplot as plt\n", + "import numpy as np\n", + "import pandas as pd\n", + "import seaborn as sns\n", + "# from scipy.stats import bootstrap" + ] + }, + { + "cell_type": "code", + "execution_count": 2, + "id": "5de05976", + "metadata": {}, + "outputs": [], + "source": [ + "from lenskit.stats._blb import _BLBootstrapper, blb_summary" + ] + }, + { + "cell_type": "code", + "execution_count": 3, + "id": "fada2c12", + "metadata": {}, + "outputs": [], + "source": [ + "rng = np.random.default_rng(20250602)" + ] + }, + { + "cell_type": "markdown", + "id": "8f0f4d2c", + "metadata": {}, + "source": [ + "## Initial Test — N=10,000" + ] + }, + { + "cell_type": "code", + "execution_count": 4, + "id": "85677bc9", + "metadata": {}, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "-0.0017 (-0.0036, 0.0003)\n" + ] + } + ], + "source": [ + "N = 1_000_000\n", + "data = rng.normal(0.0, 1.0, N)\n", + "mean = np.mean(data)\n", + "std = np.std(data)\n", + "ste = std / np.sqrt(N)\n", + "print(\"{:.4f} ({:.4f}, {:.4f})\".format(mean, mean - 1.96 * ste, mean + 1.96 * ste))" + ] + }, + { + "cell_type": "code", + "execution_count": 5, + "id": "4bb810c8", + "metadata": {}, + "outputs": [], + "source": [ + "# bootstrap([data], np.mean).confidence_interval" + ] + }, + { + "cell_type": "code", + "execution_count": 6, + "id": "d136fce6", + "metadata": {}, + "outputs": [ + { + "ename": "KeyboardInterrupt", + "evalue": "", + "output_type": "error", + "traceback": [ + "\u001b[31m---------------------------------------------------------------------------\u001b[39m", + "\u001b[31mKeyboardInterrupt\u001b[39m Traceback (most recent call last)", + "\u001b[36mCell\u001b[39m\u001b[36m \u001b[39m\u001b[32mIn[6]\u001b[39m\u001b[32m, line 2\u001b[39m\n\u001b[32m 1\u001b[39m blb = _BLBootstrapper(np.average, \u001b[32m0.95\u001b[39m, \u001b[32m0.05\u001b[39m, \u001b[32m3\u001b[39m, \u001b[32m20\u001b[39m, \u001b[32m0.6\u001b[39m, rng)\n\u001b[32m----> \u001b[39m\u001b[32m2\u001b[39m blb_df = \u001b[43mblb\u001b[49m\u001b[43m.\u001b[49m\u001b[43mrun_bootstraps\u001b[49m\u001b[43m(\u001b[49m\u001b[43mdata\u001b[49m\u001b[43m)\u001b[49m.samples\n\u001b[32m 3\u001b[39m _gstat = blb_df.groupby([\u001b[33m'\u001b[39m\u001b[33msubset\u001b[39m\u001b[33m'\u001b[39m])[\u001b[33m'\u001b[39m\u001b[33mstatistic\u001b[39m\u001b[33m'\u001b[39m]\n\u001b[32m 4\u001b[39m blb_df[\u001b[33m'\u001b[39m\u001b[33mcum_mean\u001b[39m\u001b[33m'\u001b[39m] = _gstat.cumsum() / (_gstat.cumcount() + \u001b[32m1\u001b[39m)\n", + "\u001b[36mFile \u001b[39m\u001b[32m~/LensKit/lkpy/src/lenskit/stats/_blb.py:140\u001b[39m, in \u001b[36m_BLBootstrapper.run_bootstraps\u001b[39m\u001b[34m(self, xs)\u001b[39m\n\u001b[32m 137\u001b[39m ubs = StatAccum(np.mean)\n\u001b[32m 139\u001b[39m \u001b[38;5;28;01mfor\u001b[39;00m i, ss \u001b[38;5;129;01min\u001b[39;00m \u001b[38;5;28menumerate\u001b[39m(\u001b[38;5;28mself\u001b[39m.blb_subsets(xs)):\n\u001b[32m--> \u001b[39m\u001b[32m140\u001b[39m res = \u001b[38;5;28;43mself\u001b[39;49m\u001b[43m.\u001b[49m\u001b[43mmeasure_subset\u001b[49m\u001b[43m(\u001b[49m\u001b[43mxs\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mss\u001b[49m\u001b[43m)\u001b[49m\n\u001b[32m 141\u001b[39m ss_frames[i] = res.samples\n\u001b[32m 142\u001b[39m means.record(res.mean)\n", + "\u001b[36mFile \u001b[39m\u001b[32m~/LensKit/lkpy/src/lenskit/stats/_blb.py:179\u001b[39m, in \u001b[36m_BLBootstrapper.measure_subset\u001b[39m\u001b[34m(self, xs, ss)\u001b[39m\n\u001b[32m 177\u001b[39m values.append(stat)\n\u001b[32m 178\u001b[39m means.record(stat)\n\u001b[32m--> \u001b[39m\u001b[32m179\u001b[39m \u001b[43mlbs\u001b[49m\u001b[43m.\u001b[49m\u001b[43mrecord\u001b[49m\u001b[43m(\u001b[49m\u001b[43mstat\u001b[49m\u001b[43m)\u001b[49m\n\u001b[32m 180\u001b[39m ubs.record(stat)\n\u001b[32m 182\u001b[39m \u001b[38;5;28;01mif\u001b[39;00m _check_convergence(means, svs, lbs, ubs, tol=\u001b[38;5;28mself\u001b[39m.tolerance, w=\u001b[38;5;28mself\u001b[39m.r_window):\n", + "\u001b[36mFile \u001b[39m\u001b[32m~/LensKit/lkpy/src/lenskit/stats/_blb.py:247\u001b[39m, in \u001b[36mStatAccum.record\u001b[39m\u001b[34m(self, x)\u001b[39m\n\u001b[32m 245\u001b[39m \u001b[38;5;66;03m# record and update the cumulative mean\u001b[39;00m\n\u001b[32m 246\u001b[39m \u001b[38;5;28mself\u001b[39m._values[i] = x\n\u001b[32m--> \u001b[39m\u001b[32m247\u001b[39m \u001b[38;5;28mself\u001b[39m._cum_stat[i] = \u001b[38;5;28;43mself\u001b[39;49m\u001b[43m.\u001b[49m\u001b[43m_stat_func\u001b[49m\u001b[43m(\u001b[49m\u001b[38;5;28;43mself\u001b[39;49m\u001b[43m.\u001b[49m\u001b[43mvalues\u001b[49m\u001b[43m)\u001b[49m\n", + "\u001b[36mFile \u001b[39m\u001b[32m~/LensKit/lkpy/src/lenskit/stats/_blb.py:170\u001b[39m, in \u001b[36m_BLBootstrapper.measure_subset..\u001b[39m\u001b[34m(a)\u001b[39m\n\u001b[32m 168\u001b[39m means = StatAccum(np.mean)\n\u001b[32m 169\u001b[39m svs = StatAccum(np.var)\n\u001b[32m--> \u001b[39m\u001b[32m170\u001b[39m lbs = StatAccum(\u001b[38;5;28;01mlambda\u001b[39;00m a: \u001b[43mnp\u001b[49m\u001b[43m.\u001b[49m\u001b[43mquantile\u001b[49m\u001b[43m(\u001b[49m\u001b[43ma\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[38;5;28;43mself\u001b[39;49m\u001b[43m.\u001b[49m\u001b[43m_ci_qmin\u001b[49m\u001b[43m)\u001b[49m)\n\u001b[32m 171\u001b[39m ubs = StatAccum(\u001b[38;5;28;01mlambda\u001b[39;00m a: np.quantile(a, \u001b[38;5;28mself\u001b[39m._ci_qmax))\n\u001b[32m 173\u001b[39m \u001b[38;5;28;01mfor\u001b[39;00m weights \u001b[38;5;129;01min\u001b[39;00m \u001b[38;5;28mself\u001b[39m.miniboot_weights(n, b):\n", + "\u001b[36mFile \u001b[39m\u001b[32m~/LensKit/lkpy/.venv/lib/python3.12/site-packages/numpy/lib/_function_base_impl.py:4537\u001b[39m, in \u001b[36mquantile\u001b[39m\u001b[34m(a, q, axis, out, overwrite_input, method, keepdims, weights, interpolation)\u001b[39m\n\u001b[32m 4534\u001b[39m \u001b[38;5;28;01mif\u001b[39;00m np.any(weights < \u001b[32m0\u001b[39m):\n\u001b[32m 4535\u001b[39m \u001b[38;5;28;01mraise\u001b[39;00m \u001b[38;5;167;01mValueError\u001b[39;00m(\u001b[33m\"\u001b[39m\u001b[33mWeights must be non-negative.\u001b[39m\u001b[33m\"\u001b[39m)\n\u001b[32m-> \u001b[39m\u001b[32m4537\u001b[39m \u001b[38;5;28;01mreturn\u001b[39;00m \u001b[43m_quantile_unchecked\u001b[49m\u001b[43m(\u001b[49m\n\u001b[32m 4538\u001b[39m \u001b[43m \u001b[49m\u001b[43ma\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mq\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43maxis\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mout\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43moverwrite_input\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mmethod\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mkeepdims\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mweights\u001b[49m\u001b[43m)\u001b[49m\n", + "\u001b[36mFile \u001b[39m\u001b[32m~/LensKit/lkpy/.venv/lib/python3.12/site-packages/numpy/lib/_function_base_impl.py:4550\u001b[39m, in \u001b[36m_quantile_unchecked\u001b[39m\u001b[34m(a, q, axis, out, overwrite_input, method, keepdims, weights)\u001b[39m\n\u001b[32m 4541\u001b[39m \u001b[38;5;28;01mdef\u001b[39;00m\u001b[38;5;250m \u001b[39m\u001b[34m_quantile_unchecked\u001b[39m(a,\n\u001b[32m 4542\u001b[39m q,\n\u001b[32m 4543\u001b[39m axis=\u001b[38;5;28;01mNone\u001b[39;00m,\n\u001b[32m (...)\u001b[39m\u001b[32m 4547\u001b[39m keepdims=\u001b[38;5;28;01mFalse\u001b[39;00m,\n\u001b[32m 4548\u001b[39m weights=\u001b[38;5;28;01mNone\u001b[39;00m):\n\u001b[32m 4549\u001b[39m \u001b[38;5;250m \u001b[39m\u001b[33;03m\"\"\"Assumes that q is in [0, 1], and is an ndarray\"\"\"\u001b[39;00m\n\u001b[32m-> \u001b[39m\u001b[32m4550\u001b[39m \u001b[38;5;28;01mreturn\u001b[39;00m \u001b[43m_ureduce\u001b[49m\u001b[43m(\u001b[49m\u001b[43ma\u001b[49m\u001b[43m,\u001b[49m\n\u001b[32m 4551\u001b[39m \u001b[43m \u001b[49m\u001b[43mfunc\u001b[49m\u001b[43m=\u001b[49m\u001b[43m_quantile_ureduce_func\u001b[49m\u001b[43m,\u001b[49m\n\u001b[32m 4552\u001b[39m \u001b[43m \u001b[49m\u001b[43mq\u001b[49m\u001b[43m=\u001b[49m\u001b[43mq\u001b[49m\u001b[43m,\u001b[49m\n\u001b[32m 4553\u001b[39m \u001b[43m \u001b[49m\u001b[43mweights\u001b[49m\u001b[43m=\u001b[49m\u001b[43mweights\u001b[49m\u001b[43m,\u001b[49m\n\u001b[32m 4554\u001b[39m \u001b[43m \u001b[49m\u001b[43mkeepdims\u001b[49m\u001b[43m=\u001b[49m\u001b[43mkeepdims\u001b[49m\u001b[43m,\u001b[49m\n\u001b[32m 4555\u001b[39m \u001b[43m \u001b[49m\u001b[43maxis\u001b[49m\u001b[43m=\u001b[49m\u001b[43maxis\u001b[49m\u001b[43m,\u001b[49m\n\u001b[32m 4556\u001b[39m \u001b[43m \u001b[49m\u001b[43mout\u001b[49m\u001b[43m=\u001b[49m\u001b[43mout\u001b[49m\u001b[43m,\u001b[49m\n\u001b[32m 4557\u001b[39m \u001b[43m \u001b[49m\u001b[43moverwrite_input\u001b[49m\u001b[43m=\u001b[49m\u001b[43moverwrite_input\u001b[49m\u001b[43m,\u001b[49m\n\u001b[32m 4558\u001b[39m \u001b[43m \u001b[49m\u001b[43mmethod\u001b[49m\u001b[43m=\u001b[49m\u001b[43mmethod\u001b[49m\u001b[43m)\u001b[49m\n", + "\u001b[36mFile \u001b[39m\u001b[32m~/LensKit/lkpy/.venv/lib/python3.12/site-packages/numpy/lib/_function_base_impl.py:3894\u001b[39m, in \u001b[36m_ureduce\u001b[39m\u001b[34m(a, func, keepdims, **kwargs)\u001b[39m\n\u001b[32m 3891\u001b[39m index_out = (\u001b[32m0\u001b[39m, ) * nd\n\u001b[32m 3892\u001b[39m kwargs[\u001b[33m'\u001b[39m\u001b[33mout\u001b[39m\u001b[33m'\u001b[39m] = out[(\u001b[38;5;28mEllipsis\u001b[39m, ) + index_out]\n\u001b[32m-> \u001b[39m\u001b[32m3894\u001b[39m r = \u001b[43mfunc\u001b[49m\u001b[43m(\u001b[49m\u001b[43ma\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43m*\u001b[49m\u001b[43m*\u001b[49m\u001b[43mkwargs\u001b[49m\u001b[43m)\u001b[49m\n\u001b[32m 3896\u001b[39m \u001b[38;5;28;01mif\u001b[39;00m out \u001b[38;5;129;01mis\u001b[39;00m \u001b[38;5;129;01mnot\u001b[39;00m \u001b[38;5;28;01mNone\u001b[39;00m:\n\u001b[32m 3897\u001b[39m \u001b[38;5;28;01mreturn\u001b[39;00m out\n", + "\u001b[36mFile \u001b[39m\u001b[32m~/LensKit/lkpy/.venv/lib/python3.12/site-packages/numpy/lib/_function_base_impl.py:4727\u001b[39m, in \u001b[36m_quantile_ureduce_func\u001b[39m\u001b[34m(a, q, weights, axis, out, overwrite_input, method)\u001b[39m\n\u001b[32m 4725\u001b[39m arr = a.copy()\n\u001b[32m 4726\u001b[39m wgt = weights\n\u001b[32m-> \u001b[39m\u001b[32m4727\u001b[39m result = \u001b[43m_quantile\u001b[49m\u001b[43m(\u001b[49m\u001b[43marr\u001b[49m\u001b[43m,\u001b[49m\n\u001b[32m 4728\u001b[39m \u001b[43m \u001b[49m\u001b[43mquantiles\u001b[49m\u001b[43m=\u001b[49m\u001b[43mq\u001b[49m\u001b[43m,\u001b[49m\n\u001b[32m 4729\u001b[39m \u001b[43m \u001b[49m\u001b[43maxis\u001b[49m\u001b[43m=\u001b[49m\u001b[43maxis\u001b[49m\u001b[43m,\u001b[49m\n\u001b[32m 4730\u001b[39m \u001b[43m \u001b[49m\u001b[43mmethod\u001b[49m\u001b[43m=\u001b[49m\u001b[43mmethod\u001b[49m\u001b[43m,\u001b[49m\n\u001b[32m 4731\u001b[39m \u001b[43m \u001b[49m\u001b[43mout\u001b[49m\u001b[43m=\u001b[49m\u001b[43mout\u001b[49m\u001b[43m,\u001b[49m\n\u001b[32m 4732\u001b[39m \u001b[43m \u001b[49m\u001b[43mweights\u001b[49m\u001b[43m=\u001b[49m\u001b[43mwgt\u001b[49m\u001b[43m)\u001b[49m\n\u001b[32m 4733\u001b[39m \u001b[38;5;28;01mreturn\u001b[39;00m result\n", + "\u001b[36mFile \u001b[39m\u001b[32m~/LensKit/lkpy/.venv/lib/python3.12/site-packages/numpy/lib/_function_base_impl.py:4842\u001b[39m, in \u001b[36m_quantile\u001b[39m\u001b[34m(arr, quantiles, axis, method, out, weights)\u001b[39m\n\u001b[32m 4838\u001b[39m previous_indexes, next_indexes = _get_indexes(arr,\n\u001b[32m 4839\u001b[39m virtual_indexes,\n\u001b[32m 4840\u001b[39m values_count)\n\u001b[32m 4841\u001b[39m \u001b[38;5;66;03m# --- Sorting\u001b[39;00m\n\u001b[32m-> \u001b[39m\u001b[32m4842\u001b[39m \u001b[43marr\u001b[49m\u001b[43m.\u001b[49m\u001b[43mpartition\u001b[49m\u001b[43m(\u001b[49m\n\u001b[32m 4843\u001b[39m \u001b[43m \u001b[49m\u001b[43mnp\u001b[49m\u001b[43m.\u001b[49m\u001b[43munique\u001b[49m\u001b[43m(\u001b[49m\u001b[43mnp\u001b[49m\u001b[43m.\u001b[49m\u001b[43mconcatenate\u001b[49m\u001b[43m(\u001b[49m\u001b[43m(\u001b[49m\u001b[43m[\u001b[49m\u001b[32;43m0\u001b[39;49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43m-\u001b[49m\u001b[32;43m1\u001b[39;49m\u001b[43m]\u001b[49m\u001b[43m,\u001b[49m\n\u001b[32m 4844\u001b[39m \u001b[43m \u001b[49m\u001b[43mprevious_indexes\u001b[49m\u001b[43m.\u001b[49m\u001b[43mravel\u001b[49m\u001b[43m(\u001b[49m\u001b[43m)\u001b[49m\u001b[43m,\u001b[49m\n\u001b[32m 4845\u001b[39m \u001b[43m \u001b[49m\u001b[43mnext_indexes\u001b[49m\u001b[43m.\u001b[49m\u001b[43mravel\u001b[49m\u001b[43m(\u001b[49m\u001b[43m)\u001b[49m\u001b[43m,\u001b[49m\n\u001b[32m 4846\u001b[39m \u001b[43m \u001b[49m\u001b[43m)\u001b[49m\u001b[43m)\u001b[49m\u001b[43m)\u001b[49m\u001b[43m,\u001b[49m\n\u001b[32m 4847\u001b[39m \u001b[43m \u001b[49m\u001b[43maxis\u001b[49m\u001b[43m=\u001b[49m\u001b[32;43m0\u001b[39;49m\u001b[43m)\u001b[49m\n\u001b[32m 4848\u001b[39m \u001b[38;5;28;01mif\u001b[39;00m supports_nans:\n\u001b[32m 4849\u001b[39m slices_having_nans = np.isnan(arr[-\u001b[32m1\u001b[39m, ...])\n", + "\u001b[31mKeyboardInterrupt\u001b[39m: " + ] + } + ], + "source": [ + "blb = _BLBootstrapper(np.average, 0.95, 0.05, 3, 20, 0.6, rng)\n", + "blb_df = blb.run_bootstraps(data).samples\n", + "_gstat = blb_df.groupby([\"subset\"])[\"statistic\"]\n", + "blb_df[\"cum_mean\"] = _gstat.cumsum() / (_gstat.cumcount() + 1)\n", + "blb_df = blb_df.reset_index()" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "id": "2e37fc8d", + "metadata": {}, + "outputs": [ + { + "data": { + "image/png": "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", + "text/plain": [ + "
" + ] + }, + "metadata": {}, + "output_type": "display_data" + } + ], + "source": [ + "plt.hlines([0.0], xmin=0, xmax=blb_df[\"iter\"].max(), label=\"True mean\", color=\"black\")\n", + "plt.hlines([mean], xmin=0, xmax=blb_df[\"iter\"].max(), label=\"Data mean\", color=\"magenta\", ls=\"--\")\n", + "plt.hlines(\n", + " [blb_df[\"statistic\"].mean()],\n", + " xmin=0,\n", + " xmax=blb_df[\"iter\"].max(),\n", + " color=\"red\",\n", + " label=\"BLB mean\",\n", + " ls=\"-.\",\n", + ")\n", + "for snum, sdf in blb_df.groupby(\"subset\"):\n", + " plt.plot(sdf[\"iter\"], sdf[\"cum_mean\"], color=\"steelblue\", alpha=0.5)\n", + "plt.legend()\n", + "plt.show()" + ] + }, + { + "cell_type": "markdown", + "id": "8ed0e077", + "metadata": {}, + "source": [ + "## Randomized Testing\n", + "\n", + "Now let's test a bunch of possible values." + ] + }, + { + "cell_type": "code", + "execution_count": null, + "id": "509893b4", + "metadata": {}, + "outputs": [ + { + "data": { + "text/plain": [ + "(200, 100000)" + ] + }, + "execution_count": 51, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "M = 200\n", + "N = 100_000\n", + "means = rng.normal(0, 10, size=M)\n", + "stds = rng.standard_exponential(size=M) + 0.1\n", + "\n", + "data = rng.normal(\n", + " np.broadcast_to(means.reshape((M, 1)), (M, N)), np.broadcast_to(stds.reshape((M, 1)), (M, N))\n", + ")\n", + "data.shape" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "id": "1a38c5ce", + "metadata": {}, + "outputs": [], + "source": [ + "data_means = np.mean(data, axis=1)\n", + "data_stds = np.std(data, axis=1)\n", + "param_stats = pd.DataFrame(\n", + " {\n", + " \"mean\": data_means,\n", + " \"ci_lower\": data_means - 1.96 * (data_stds / np.sqrt(N)),\n", + " \"ci_upper\": data_means - 1.96 * (data_stds / np.sqrt(N)),\n", + " }\n", + ")" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "id": "3a64a3c6", + "metadata": {}, + "outputs": [], + "source": [ + "# boots = [bootstrap([data[i, :]], np.mean, n_resamples=5000) for i in range(M)]\n", + "# boot_stats = pd.DataFrame.from_records(\n", + "# {\n", + "# \"mean\": np.mean(data[i, :]),\n", + "# \"ci_lower\": boot.confidence_interval.low,\n", + "# \"ci_upper\": boot.confidence_interval.high,\n", + "# }\n", + "# for i, boot in enumerate(boots)\n", + "# )\n", + "# boot_stats" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "id": "c59e3e12", + "metadata": {}, + "outputs": [ + { + "ename": "KeyboardInterrupt", + "evalue": "", + "output_type": "error", + "traceback": [ + "\u001b[31m---------------------------------------------------------------------------\u001b[39m", + "\u001b[31mKeyboardInterrupt\u001b[39m Traceback (most recent call last)", + "\u001b[36mCell\u001b[39m\u001b[36m \u001b[39m\u001b[32mIn[54]\u001b[39m\u001b[32m, line 1\u001b[39m\n\u001b[32m----> \u001b[39m\u001b[32m1\u001b[39m blbs = [\u001b[43mblb_summary\u001b[49m\u001b[43m(\u001b[49m\u001b[43mdata\u001b[49m\u001b[43m[\u001b[49m\u001b[43mi\u001b[49m\u001b[43m,\u001b[49m\u001b[43m:\u001b[49m\u001b[43m]\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[33;43m'\u001b[39;49m\u001b[33;43mmean\u001b[39;49m\u001b[33;43m'\u001b[39;49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mb_factor\u001b[49m\u001b[43m=\u001b[49m\u001b[32;43m0.8\u001b[39;49m\u001b[43m)\u001b[49m \u001b[38;5;28;01mfor\u001b[39;00m i \u001b[38;5;129;01min\u001b[39;00m \u001b[38;5;28mrange\u001b[39m(M)]\n\u001b[32m 2\u001b[39m blb_stats = pd.DataFrame.from_records(blbs)\n\u001b[32m 3\u001b[39m blb_stats\n", + "\u001b[36mFile \u001b[39m\u001b[32m~/LensKit/lkpy/src/lenskit/stats/_blb.py:69\u001b[39m, in \u001b[36mblb_summary\u001b[39m\u001b[34m(xs, stat, ci_width, b_factor, rel_tol, s_window, r_window, rng)\u001b[39m\n\u001b[32m 0\u001b[39m \n", + "\u001b[36mFile \u001b[39m\u001b[32m~/LensKit/lkpy/src/lenskit/stats/_blb.py:139\u001b[39m, in \u001b[36mrun_bootstraps\u001b[39m\u001b[34m(self, xs)\u001b[39m\n\u001b[32m 136\u001b[39m lbs = StatAccum(np.mean)\n\u001b[32m 137\u001b[39m ubs = StatAccum(np.mean)\n\u001b[32m--> \u001b[39m\u001b[32m139\u001b[39m \u001b[38;5;28;01mfor\u001b[39;00m i, ss \u001b[38;5;129;01min\u001b[39;00m \u001b[38;5;28menumerate\u001b[39m(\u001b[38;5;28mself\u001b[39m.blb_subsets(xs)):\n\u001b[32m 140\u001b[39m res = \u001b[38;5;28mself\u001b[39m.measure_subset(xs, ss)\n\u001b[32m 141\u001b[39m ss_frames[i] = res.samples\n", + "\u001b[36mFile \u001b[39m\u001b[32m~/LensKit/lkpy/src/lenskit/stats/_blb.py:168\u001b[39m, in \u001b[36mmeasure_subset\u001b[39m\u001b[34m(self, xs, ss)\u001b[39m\n\u001b[32m 165\u001b[39m xss = xs[ss]\n\u001b[32m 167\u001b[39m values = []\n\u001b[32m--> \u001b[39m\u001b[32m168\u001b[39m means = StatAccum(np.mean)\n\u001b[32m 169\u001b[39m svs = StatAccum(np.var)\n\u001b[32m 170\u001b[39m lbs = StatAccum(\u001b[38;5;28;01mlambda\u001b[39;00m a: np.quantile(a, \u001b[38;5;28mself\u001b[39m._ci_qmin))\n", + "\u001b[36mFile \u001b[39m\u001b[32m~/LensKit/lkpy/src/lenskit/stats/_blb.py:189\u001b[39m, in \u001b[36mminiboot_weights\u001b[39m\u001b[34m(self, n, b)\u001b[39m\n\u001b[32m 186\u001b[39m df.index.name = \u001b[33m\"\u001b[39m\u001b[33miter\u001b[39m\u001b[33m\"\u001b[39m\n\u001b[32m 187\u001b[39m \u001b[38;5;28;01mreturn\u001b[39;00m _BootResult(means.statistic, svs.statistic, lbs.statistic, ubs.statistic, df)\n\u001b[32m--> \u001b[39m\u001b[32m189\u001b[39m \u001b[38;5;28;01mdef\u001b[39;00m\u001b[38;5;250m \u001b[39m\u001b[34mminiboot_weights\u001b[39m(\u001b[38;5;28mself\u001b[39m, n: \u001b[38;5;28mint\u001b[39m, b: \u001b[38;5;28mint\u001b[39m):\n\u001b[32m 190\u001b[39m flat = np.full(b, \u001b[32m1.0\u001b[39m / b)\n\u001b[32m 192\u001b[39m \u001b[38;5;28;01mwhile\u001b[39;00m \u001b[38;5;28;01mTrue\u001b[39;00m:\n", + "\u001b[31mKeyboardInterrupt\u001b[39m: " + ] + } + ], + "source": [ + "blbs = [blb_summary(data[i, :], \"mean\", b_factor=0.8) for i in range(M)]\n", + "blb_stats = pd.DataFrame.from_records(blbs)\n", + "blb_stats" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "id": "b229ea22", + "metadata": {}, + "outputs": [ + { + "data": { + "text/html": [ + "
\n", + "\n", + "\n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + "
ParametricBootstrapBLBErrorRelError
quantitysamp
ci_lower02.597184-4.4230932.593609-0.0035750.001377
1-12.386757-4.425957-12.3863000.0004580.000037
213.514136-8.08019113.510683-0.0034530.000256
39.921847-25.3295039.9219690.0001220.000012
4-0.510517-2.748500-0.5099920.0005260.001029
.....................
mean1951.883268NaN1.8844180.0011500.000611
196-12.321563NaN-12.3212200.0003430.000028
19719.500813NaN19.5009180.0001050.000005
198-7.208713NaN-7.2085160.0001960.000027
1999.404558NaN9.4048170.0002590.000028
\n", + "

600 rows × 5 columns

\n", + "
" + ], + "text/plain": [ + " Parametric Bootstrap BLB Error RelError\n", + "quantity samp \n", + "ci_lower 0 2.597184 -4.423093 2.593609 -0.003575 0.001377\n", + " 1 -12.386757 -4.425957 -12.386300 0.000458 0.000037\n", + " 2 13.514136 -8.080191 13.510683 -0.003453 0.000256\n", + " 3 9.921847 -25.329503 9.921969 0.000122 0.000012\n", + " 4 -0.510517 -2.748500 -0.509992 0.000526 0.001029\n", + "... ... ... ... ... ...\n", + "mean 195 1.883268 NaN 1.884418 0.001150 0.000611\n", + " 196 -12.321563 NaN -12.321220 0.000343 0.000028\n", + " 197 19.500813 NaN 19.500918 0.000105 0.000005\n", + " 198 -7.208713 NaN -7.208516 0.000196 0.000027\n", + " 199 9.404558 NaN 9.404817 0.000259 0.000028\n", + "\n", + "[600 rows x 5 columns]" + ] + }, + "execution_count": 37, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "comb_stats = pd.DataFrame(\n", + " {\n", + " \"Parametric\": param_stats.unstack(),\n", + " # \"Bootstrap\": boot_stats.unstack(),\n", + " \"BLB\": blb_stats.drop(columns=[\"value\"]).unstack(),\n", + " }\n", + ")\n", + "comb_stats.index.rename([\"quantity\", \"samp\"], inplace=True)\n", + "comb_stats[\"Error\"] = comb_stats[\"BLB\"] - comb_stats[\"Parametric\"]\n", + "comb_stats[\"RelError\"] = (\n", + " np.abs(comb_stats[\"BLB\"] - comb_stats[\"Parametric\"]) / comb_stats[\"Parametric\"].abs()\n", + ")\n", + "comb_stats" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "id": "41e29fb8", + "metadata": {}, + "outputs": [ + { + "data": { + "text/html": [ + "
\n", + "\n", + "\n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + "
quantitysampParametricBootstrapBLBErrorRelErrorRealMeanRealSTDAbsMean
0ci_lower02.597184-4.4230932.593609-0.0035750.0013772.6011411.5925882.601141
1ci_lower1-12.386757-4.425957-12.3863000.0004580.000037-12.3857950.54763512.385795
2ci_lower213.514136-8.08019113.510683-0.0034530.00025613.5148521.42126513.514852
3ci_lower39.921847-25.3295039.9219690.0001220.0000129.9221370.1409949.922137
4ci_lower4-0.510517-2.748500-0.5099920.0005260.001029-0.5094120.3620410.509412
.................................
595mean1951.883268NaN1.8844180.0011500.0006111.8848012.1492751.884801
596mean196-12.321563NaN-12.3212200.0003430.000028-12.3192992.50934512.319299
597mean19719.500813NaN19.5009180.0001050.00000519.5011530.61237219.501153
598mean198-7.208713NaN-7.2085160.0001960.000027-7.2094430.2137607.209443
599mean1999.404558NaN9.4048170.0002590.0000289.4043991.0370949.404399
\n", + "

600 rows × 10 columns

\n", + "
" + ], + "text/plain": [ + " quantity samp Parametric Bootstrap BLB Error RelError \\\n", + "0 ci_lower 0 2.597184 -4.423093 2.593609 -0.003575 0.001377 \n", + "1 ci_lower 1 -12.386757 -4.425957 -12.386300 0.000458 0.000037 \n", + "2 ci_lower 2 13.514136 -8.080191 13.510683 -0.003453 0.000256 \n", + "3 ci_lower 3 9.921847 -25.329503 9.921969 0.000122 0.000012 \n", + "4 ci_lower 4 -0.510517 -2.748500 -0.509992 0.000526 0.001029 \n", + ".. ... ... ... ... ... ... ... \n", + "595 mean 195 1.883268 NaN 1.884418 0.001150 0.000611 \n", + "596 mean 196 -12.321563 NaN -12.321220 0.000343 0.000028 \n", + "597 mean 197 19.500813 NaN 19.500918 0.000105 0.000005 \n", + "598 mean 198 -7.208713 NaN -7.208516 0.000196 0.000027 \n", + "599 mean 199 9.404558 NaN 9.404817 0.000259 0.000028 \n", + "\n", + " RealMean RealSTD AbsMean \n", + "0 2.601141 1.592588 2.601141 \n", + "1 -12.385795 0.547635 12.385795 \n", + "2 13.514852 1.421265 13.514852 \n", + "3 9.922137 0.140994 9.922137 \n", + "4 -0.509412 0.362041 0.509412 \n", + ".. ... ... ... \n", + "595 1.884801 2.149275 1.884801 \n", + "596 -12.319299 2.509345 12.319299 \n", + "597 19.501153 0.612372 19.501153 \n", + "598 -7.209443 0.213760 7.209443 \n", + "599 9.404399 1.037094 9.404399 \n", + "\n", + "[600 rows x 10 columns]" + ] + }, + "execution_count": 38, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "comb_stats = comb_stats.join(\n", + " pd.Series(means, name=\"RealMean\", index=pd.Index(np.arange(M), name=\"samp\"))\n", + ")\n", + "comb_stats = comb_stats.join(\n", + " pd.Series(stds, name=\"RealSTD\", index=pd.Index(np.arange(M), name=\"samp\"))\n", + ")\n", + "comb_stats[\"AbsMean\"] = comb_stats[\"RealMean\"].abs()\n", + "comb_stats.reset_index(inplace=True)\n", + "comb_stats" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "id": "34c4fd67", + "metadata": {}, + "outputs": [ + { + "data": { + "image/png": "iVBORw0KGgoAAAANSUhEUgAABdEAAAHqCAYAAADrpwd3AAAAOnRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjEwLjEsIGh0dHBzOi8vbWF0cGxvdGxpYi5vcmcvc2/+5QAAAAlwSFlzAAAPYQAAD2EBqD+naQAAdmlJREFUeJzt3Ql4VOXZ//E7y2SZhJls7CQYDXUDlLoii2tF6wbS9i32b1WoWhWXaivigrVKEW1tK7hWRd5WcEPUWl+rFZVFtFXBgCAlFggYtoRkBrIv87/uB2ecLDPJJJn9+7muSGbOSTg5IL/n3M+W4HK5XAIAAAAAAAAAANpJbP8WAAAAAAAAAABQFNEBAAAAAAAAAPCBIjoAAAAAAAAAAD5QRAcAAAAAAAAAwAeK6AAAAAAAAAAA+EARHQAAAAAAAAAAHyiiAwAAAAAAAADgA0V0AAAAAAAAAAB8oIgOAAAAAAAAAIAPFNEBdMv7778vCQkJUlVVJdHk8ssvl4kTJ3bp3NNOO01uuummoF8TAADxmrUAAESTaM1mAD1HER1At4rJp5xyiuzcuVPsdrt5/eyzz0pWVpZEuj/96U/mWgEAiCRkLQAAkSWWshlAzyX3wvcAEIdSUlJkwIABEm3cjZ1o1NDQYO47ACA+kLXhRe4CAGIlmwH0HCPRgQhVXV0tP/3pTyUzM1MGDhwov//979v1hOs0sldffbXV12kvuPforxkzZsh3vvMdsVqtcuihh8pdd90ljY2NnuO//vWv5dhjj5W//OUvcsghh5gH3x//+Meyf/9+z5TsDz74wIwq099PP7Zu3dpqGpt+fsUVV4jD4fCco9/3N7/5jQwfPrzdz6a/n15HsHzxxRdy/vnni81mkz59+si4cePkq6++6vEU88rKSvNnkp2dbe7nueeeK5s3bzbHXC6X9O3bV15++eVWP6f+2bmtXLlSUlNTpaamxrzWe/ezn/3MfJ1e6xlnnCGff/55uz+bp556SgoLCyUtLa3b9wQA0B5ZG/6s1fvxxz/+sd2168/mpj/rY489ZnI3PT3d3GPvvNV7pec8//zzZoSg5qXeE72n3tavX2++h/559+/fXy699FIpLy/3HNc/++nTp5s//7y8PJkwYUK37w8AoHvI5u5xX9c//vEPGTVqlMlLfb7cs2eP/N///Z8ceeSRJrMvueQSz/OoamlpkTlz5pjnTf2aY445plXGNjc3y7Rp0zzHDz/8cHNPvLlz/3e/+535M8vNzZXrrruu1f0GYgFFdCBC/epXvzKh/dprr8nbb79tQvGzzz4L+Pvog602JjZs2GDC7s9//rP84Q9/aHWOPvRqI+SNN94wH/r73n///eaYfs3o0aPlyiuvNNPW9CM/P7/V1+sDqz4Aayi7z/nlL38pU6dOlY0bN8q///1vz7lr1qyR4uJi09jwRRtM/j5+/vOf+/zar7/+WsaPH2+K1cuWLZNPP/3UXEdTU5P0lDYOPvnkE3n99ddl9erVpnD+/e9/3zQOtMGiv6/+ObkL7vqz19bWypdffmne0/t6wgknmIac+uEPf+hp1Oh1fve735UzzzxT9u3b5/k9S0pKZMmSJfLKK6/I2rVre/wzAAC+RdZGXtb6okWHyZMnm87mn/zkJ6bQoT932z/PW265xfz8ej8vuOACqaioMMe02KHFBC0saJa/9dZbsnv3bvnRj37U6nssXLjQjDJctWqVPP7440H7eQAAHSObu5fNblrEnz9/vnz44Yeyfft2k3N6jYsWLZK///3v5p7OmzfPc74W0P/3f//XZJ52kP/iF7+Q//f//p+nI1qL7EOGDJGXXnrJ3MtZs2bJ7bffLi+++GKr3/e9994z91N/1SzVe8/Sbog1LOcCRKADBw7I008/LX/9619NUVVpEGl4BerOO+/0fK497BrqOlLr1ltv9byvwagBpw0NpSOz3n33XZk9e7bpkdeHSS38+pq2psf1PC0ke5+jQa+juBYsWGCKx0o/P/XUU81oAF86KxZrI8WXRx55xFyL/owWi8W8pyMQekpHnGvxXB+qtbGknnvuOdOQ0oaXFsR1hMQTTzxhji1fvtw8qOv90IbfEUccYX7Vn909Kv1f//qXKaJrEUJpz71+L+35v+qqqzxTybVRo6PVAQC9h6yNvKz1R3NWZ2+pe++9V9555x1TBHj00Uc95+goci20Kx25roVy/TPWPwctKGgu//a3v/Wc/8wzz5gc/89//uO5/mHDhskDDzwQ1J8FANAxsrn72ex23333yZgxY8znOoJ85syZprjt/n1/8IMfmEK3jtSvr683ufjPf/7TdBgoPU+fVfW5Vq9Xc/6ee+7xfH8dka4DyrSI7t0RrbO1NWuTkpLMs+95551n7qV2QgCxgiI6EIE05LR4etJJJ3ney8nJMVOnAvXCCy/Iww8/bL6nNkp0lFjb8NVGhbvhoHQKlhZ3e4OGpvbEP/TQQ5KYmGh6wNuOAGirqKio27+fNjx0Srn7ob636EiC5OTkVn8mOk1N/0zcI+G0kXHjjTfK3r17Tc+9FtXdRXRtwOhoAHejTUfS6Z+Hfg9vOnLdPR1eDR06lAI6AAQBWRt5WeuP++He+3XbYoP3OZrZxx9/vCejNXe1aKCFjbb0z81dRD/uuOOC9BMAADpDNnc/m91Gjhzp+VyXLnMvZ+P9ng7mcs961qVdvve977X6HvpnoB3P3p3n2vFcWlpqnlf1uC5N4+3oo482BXTve7lu3boe/zxAJKGIDkQx7fHWJUW8ea87pj3EOuVZe461J9w9akzXlfPW9iFYv6/2yvcGnUqtI62XLl1qeur1+rT325+OHnC96fQyX1OsdZ22cBkxYoRp5GkBXT90BIMW0efOnWum8unP7h7Frg05bVi4l3/x5r27e0ZGRkh/BgBAa2RtcLNWCwv+7m9v0dzV+6SZ3Jb3/iXkLgBEPrLZN++fSX8efz+jZqPSZV4GDx7c6jz3bGm9bzqKX++ddlZrp8ODDz4oH3/8sc/ft+3vA8QKiuhABDrssMNMCGkwFRQUeNbY1unG7uVAlI5Q1nXXvJcc8d4kREc+60jmO+64w/Petm3bAr4eDX3dUKQ75+hIsMsuu8xMX9NzdA3Tzh6+ezKNTXvedcqfNlJ6c4ScbsSioxf0z8RdCNd1Vjdt2iRHHXWUp6GgI/N0/T5dT27s2LGm51+nyel0OB0R53441/XPd+3aZe6PjoAAAIQWWRsZWdv2/jqdTtmyZUu78z766COz0Zz3a+9Rcu73dK12pZmta7XrEi/u3NU9RjRz9X4BACIP2dzz5VwCoc+xWizXEebe99ebeznTa6+91vOe98xpIJ7QggQikPZA6/IfuqmKLvfRr18/0wDQ0VredIMsXXdMe4Q1uHVdM++HWV3XUwNRe491LTbtYdae8EDpA6c2ZHQ3cr02HW3d0Tnak63rnumO3lo8dm+gqWuYahHaHcKd6ck0Nn1Y1jVStZGi67/pqAN9qD7xxBO7NQ3Q+15edNFFZlqeFsS1B/62224zPfb6vpsu4aKbmmnB3D2SQB/odf10/fN0O+uss8yfm+5irmuv6jTysrIy82c0adIk8/UAgOAhayMja/X+6nq0OmJPZ2LphmXe08HddEMzzUbtoNZM1anoum6uN51urn8eeh90yrwWXnQqvbruuuvMpnJTpkwxS6vp/dVp7Prn9tRTT3X4ewIAQots7vlyLoHQZ1odZa6bieqocc1Yh8NhrlUL9toJoPdS9+j6xz/+YdZD/8tf/mJmWevnQLxp/S8RgIihU6R0VLM+VGrBVQOt7TqdOqVKN8TS8y655BITgO7AVhdeeKEJRH3Y1TXLtEf+rrvuCvha9Pvqw6X2VGuvvzZI2tLead0t/H/+53/MOd6bcmnw6nHdYMR7fbtg0MbWsmXLTENGe9P1nulDc2+MStdRBPr9zj//fNNg0ymEb775Zqvvrb+nNuS0mO6mn7d9T0et69dqgV13aNciuhYjdISErlMHAAg+sjb8WatFeP0emq26CZl2LutIxLZ0Sr4WQ3QUvD7ML1682DMTzO3+++83H1rE0E3RdEPwvLw8c2zQoEGmKKB5fPbZZ5sl2G666SZTuG9bnAEAhA/ZHFq6Wbfemzlz5piC/znnnGM6HdxF8quvvlouvvhi8/Ppz6Czsb1HpQPxJMHVdiEpABFLi7DaCPjjH/8o0UT/mdEGhIbtzTffHO7LAQDAJ7I28mjHs44g1AJ7R3SEoD7sr1mzpt1GZwCA6Ec2A4gELOcCIKj27t1rRo7p+t864hoAAPQushYAgMhCNgOxhyI6gKDSdex0KvWTTz4p2dnZEil0Kl7baeDeNmzY4NnMBgCASEbWAgAQWSI1mwF0H8u5AIhLTU1NZvq3L7pBjO6oDgAAuoesBQAAQKygiA4AAAAAAAAAgA9sRQ8AAAAAAAAAgA8U0QEAAAAAAAAA8IEiehu6uo3T6TS/AgCA4CFzAQAIDTIXAICeoYjexv79+8Vut5tfAQBA8JC5AACEBpkLAEDPUEQHAAAAAAAAAMAHiugAAAAAAAAAAPhAER0AAAAAAAAAAB8oogMAAAAAAAAA4ANFdAAAAAAAAAAAfKCIDgAAAAAAAACADxTRAQAAAAAAAADwgSI6AAAAAAAAAAA+UEQHAAAAAAAAAMAHiugAAAAAAAAAAPhAER0AAAAAAAAAAB8oogMAAAAAAAAA4ANFdAAAAAAAAAAAfKCIDgAAAAAAAACAD8nhvgAAANC7HDUNUn6gQZx1jWJLt0heRorYrSnhviwAAGIOmQsAQHxkLkV0AABiSFlVrcxYUiwrNpd73hs/LE/unzxSBmWlh/XaAACIJWQuAADxk7ks5wIAQAz1zLdtWKjlm8vltiXF5jgAAOg5MhcAgPjKXIroAADECJ3a1rZh4d3A0OMAAKDnyFwAAOIrcymiAwAQI3RtOH/2d3IcAAB0DZkLAEB8ZS5FdAAAYoQtzeL3eJ9OjgMAgK4hcwEAiK/MpYgOAECMyMtMMZurdETf1+MAAKDnyFwAAOIrcymiAwAQI+zWFLM7edsGhr6eO3mkOQ4AAHqOzAUAIL4yN8HlcrlC8jtFCafTKXa7XRwOh9hstnBfDgAAAdPdyXVzFV0bTqe2ac98JD7Mk7kAgGhH5gIAEB+Zmxyy3wkAAISENiQi8QEeAIBYQ+YCABAfmctyLgAAAAAAAAAA+EARHQAAAAAAAAAAHyiiAwAAAAAAAADgA0V0AAAAAAAAAAB8oIgOAAAAAAAAAIAPFNEBAAAAAAAAAPCBIjoAAAAAAAAAAD5QRAcAAAAAAAAAwAeK6AAAAAAAAAAA+JDs6wAAAAguR02DlB9oEGddo9jSLZKXkSJ2a0q4LwsAgJhD5gIAEBqOGM1ciugAAIRBWVWtzFhSLCs2l3veGz8sT+6fPFIGZaWH9doAAIglZC4AAKFRFsOZy3IuAACEoWe+bcNCLd9cLrctKTbHAQBAz5G5AACEhiPGM5ciOgAAIaZT29o2LLwbGHocAAD0HJkLAEBolMd45lJEBwAgxHRtOH/2d3IcAAB0DZkLAEBoOGM8cymiAwAQYrY0i9/jfTo5DgAAuobMBQAgNGwxnrkU0QEACLG8zBSzuUpH9H09DgAAeo7MBQAgNPJiPHMpogMAEGJ2a4rZnbxtA0Nfz5080hwHAAA9R+YCABAa9hjP3ASXy+UK90VEEqfTKXa7XRwOh9hstnBfDgAghunu5Lq5iq4Np1PbtGc+2hsWgSBzAQChQuaSuQCA0HDEaOYmh/sCAACIV9qQiIXGBAAAkY7MBQAgNOwxmrkU0QEACEKvu+5Mbku3SF5GbDYgAAAINzIXAIDQcJC5FNEBAOgtZVW1MmNJsazYXN5q/TddF25QVnpYrw0AgFhC5gIAEBpk7kFsLAoAQC/1zLdtWKjlm8vltiXF5jgAAOg5MhcAgNAgc79FER0AgF6gU9vaNiy8Gxh6HAAA9ByZCwBAaJC536KIDgBAL9C14fzRnckBAEDPkbkAAIQGmfstiugAAPQCW5rF7/E+nRwHAABdQ+YCABAaZO632FgUAIBe2Jk8LzPFbK6iU9ra0vf1OAAA6HnuZqYlk7kAAAQBmRsDRfQ5c+bIK6+8Il9++aWkp6fLKaecInPnzpXDDz/cc05dXZ3ccsst8vzzz0t9fb1MmDBBHn30Uenfv39Yrx0AEPs7k8+dPNLsTq6bqyzv4JjdGj+NCwAAgpm73zuyn9w3cbjc+ep6MhcAgF5C5vqX4HK5XBIFzjnnHPnxj38sJ5xwgjQ1Ncntt98u69evlw0bNkhGRoY555prrpG///3v8uyzz4rdbpfp06dLYmKirFq1qsu/j9PpNF/rcDjEZrMF8ScCAERjr/z0xWs63FhFGxHzpowyn2vPva4Np1PbtGc+nhoWgSBzAQDdzV3zUD9phByoayJzu4DMBQD4Q+bG0Ej0t956q9VrLZT369dPPv30Uxk/frxpDDz99NOyaNEiOeOMM8w5CxYskCOPPFI++ugjOfnkk8N05QCAeNqZ/LB+mXHXmAAAINS5+87GPXLbuU0mdwEAQM+QuTG8sagWzVVOTo75VYvpjY2NctZZZ3nOOeKII6SgoEBWr14dtusEAMQOdiYHACB0yF0AAEKDzI2hkejeWlpa5KabbpIxY8bI8OHDzXu7du2SlJQUycrKanWuroeux3zRtdP1w3uaGwAAbruddVJZrRurNIktLVmmn1Ekz6zcIjUNzXG9M3l3kLkAgM6mku/ZXy9VtY2SkZIkWekWsaYkdZi5itz1jcwFAPhD5sZJEf26664z66GvXLmyVzYsveeee3rlugAAsaW0olpmLl0nq0oqPO+NLcqVh6eMkhsWr2nVwIi3ncm7g8wFAPjdzOzlYllR8u1U8nFFufL0ZcfLtIWftHuoJ3f9I3MBAL6QuTG+saibbhb62muvyfLly6WwsNDz/rJly+TMM8+UysrKVqPRhw4dakat/+IXv+hyD31+fj4brgBAnNMR6De/uLZVAd27kH5sQbbMX1bSamfygVnpYbjS6EHmAgB8bma2aE2rh3m3cUV58v0RA2Tm0vWe98jdzpG5AICOkLlxMBJda/3XX3+9LF26VN5///1WBXR13HHHicVikXfffVcmT55s3tu0aZOUlpbK6NGjfX7f1NRU8wEAgDddwqWjArpaWVIhd5x3lJx1RL+43Zm8O8hcAIDPzcw6eJhX+v6NZw2Tt24cJ3WNzeRuF5G5AICOkLlxUETXJVwWLVpkRqH36dPHs8653W6X9PR08+u0adPk5ptvNpuNau+6Ft21gH7yySeH+/IBAFFG10D3Z39dk5xYeHBzawAAELzNzHTN1uTEBDMLDAAAdB+ZGwdF9Mcee8z8etppp7V6f8GCBXL55Zebz//whz9IYmKiGYmuU9cmTJggjz76aFiuFwAQfdPatFdeGxW2dIvZRNSfzo4DAICuZW5mqv9MTU1OZEMzAAC6gcztPVFTAejK0u1paWnyyCOPmA8AAALaWGVJsazY/O20tuevOtmsfa5Lt7Sl72dnMKUNAIDeyNw5F4+QccPyWr3nNqYo14yKO34oI+IAAAgEmRvnG4sGm264okvDsOEKAMRPw0LXP9flW/qkJZsNRbWhoZ6/arT8+vX1rQrpWkD/7aQRUpCbEcarjg1kLgDEF1+ZW9PQLAsuP0Hmv1fS6qFeH+avP2OYHJJjlQFsaNYjZC4AxBcyt/dRRG+DxgUAxI/SimqZuXRdqw1EtUh+1/lHy0+e+si8XnrNKVLd0GwaH7qEi45A729LC+NVxw4yFwDiR2eZqw/1/3fjOGloahFHbaNYU5IkIyVZsqwWNjTrBWQuAMQPMjfOl3MBAKC3e+bbNiyUjjq/940vZO7kkTJt4SdSUd3ApioAAIQgc3XEHJkLAED3kbnBkxjE7w0AQERPbWvbsPBuYPSzpZrP2VQFAIDub2a2aZdTqmobyVwAAIKIzA0+iugAgLiia8FtLa+WA/VNfs87UNcs44flSV4m09kAAOhOh/X0RWtkwh9XiKOm0e+5ZC4AAN1H5oYGRXQAQFw1Ln754lq55KmPxZqa5PdcXf9cp7qxJhwAAIGPhtPNy1aUHNywLDONzAUAIBjI3NChiA4AiLPGxcGpbXuc9WZzlY7o+7qB6EB2JQcAIGDlBxpkxeaDD/OKzAUAIDjI3NChiA4AiPni+Vd7Dsh/9hyQK8YUyvQziszu41pQ193J2zYw9PVvJ42QQTQsAADoVubqptzPXH4CmQsAQJCQuaGX4HK5XOG+iEjidDrFbreLw+EQm80W7ssBAPRw+RYz+tyrZ35MUa4ppt+weI1pZOhUNt1cpaa+WTJTk03PPA2L0CBzASB2kLmRjcwFgNhB5oYHRfQ2aFwAQOz0zE9fvKZVw8K7gTGqIFvmLyvxvLfoZydJYV4GU9tCiMwFgPjLXH19/shBctp3+pK5IUTmAkBsIHPDh+VcAABxsTact1UlFTIqP8vzetywPDmsXyYNCwAAgpi5mrezJ46Q7w8fQOYCANANZG74JIfx9wYAIGicdY1+j9c3tZhfxw/LM1Pd+tvSQnRlAADEV+b2SbPIuzefKnmZKWK3poTsugAAiDVkbvhQRAcAxCRbmsXv8UNyrTQuAAAIQebmZqSYGV8AAKBnyNzwYTkXAEBM0uK4jjLviL4/OCvdNC4ooAMAENzM1eMAAKDnyNzwoYgOAIhJWhy/f/LIdg0M9/ItFM8BAOgdZC4AAKFB5oZPgsvlcoXx94847FoOANG1M/luZ704ahslIzVJ0i1JkpGaLP281jfXc3Tzlf11jWZ9OJZviRxkLgBEDzI3upG5ABA9yNzIxJroAICoVFZVKzNe/lxWlFR43htTlCvTTy+S+sZmyc/NMO9pQ4LGBAAA3UfmAgAQGmRu5KKIDgCIOtrrPuPl4lYNC7Xqm9fnjxgoZ1iSpL9XTz0AAAgcmQsAQGiQuZGNNdEBAFHXsNjpqJMVJeUdHtcGhk5zq6xuCPm1AQAQS8hcAABCg8yNfBTRAQBRNbVt+uI18t/yar/n1Te1iLOuKWTXBQBArCFzAQAIDTI3OlBEBwBEz9S2JcWyYnO5pCb7jy89bktjxTIAALqDzAUAIDTI3OhBER0AEBV053FtWKg126tkXFFuh+fppit7nHWSncEmKwAAdAeZCwBAaJC50YPuCwBAVHDWNXo+f2blFpl/ySgRSWi1ZtzBXcuHyZCsNDZbAQCgm8hcAABCg8yNHhTRAQBRwZZm8Xxe09As0xetkavGHyo3njVMmlpc0ictWayWJMlITTYbrgAAgO4hcwEACA0yN3okuFwuV7gvIpI4nU6x2+3icDjEZrOF+3IAAF5rxV2/eI0s/2aqm7fxw/Jk3pRRYrcytS2akLkAEBn5qlPJdSScLd0ied9MEydzYwuZCwDhR+ZGN9ZEBwBE3M7kG3c65eP/VsiXO53mtdKGw/2TR5qGhDd9PXfySBoWAAAEaMe+GvmizCmbdu+XyppGWbrma/nlS5+bkXBkLgAAvefrfTXy5rqdsrWiWnY66mRbRY28uX6X1JK5UYOR6G3QQw8A4bOtolpuX7pOVpVUeN4bW5QrsyeNkKG5Ga167/fXNUqfNIvkZabQsIhSZC4ARFbm6pqrV4wplBf+VSq/++Ex5j0yNzaQuQAQPmWVNbJtX43Mf6+kXe5ef8YwOXJAH/OazI1srIkOAIgIOuK87cO8WllSIXcsXScP/OAYGZSVbhoSNCYAAOi+3c46ubODzHW/HlWQbR7kD+uXSeYCANADOgisqraxXQFduV//duIIGZqXQeZGOJZzAQBEBEdtY7tGhXchXY8DAICeq6xukBU+MlezeFR+lhkJBwAAekY7pXWDUF/Puvp+dUNTyK8LgWMkOgAgIjZWaWxq8XsuD/MAAPRO5jY1+1/Rs76pxUwlBwAAPcvdiuqGTs/VvUgQ+SiiAwDCtnzLjCXFsuKbXcj/dv0Yv+fzMA8AQO9k7tOXHe/3/Kz0g2uxAgCAnufuyz8f7fd8ezrPutGA5VwAAGHplfduVKg9znqziWhH9H0aFgAA9E7mrtleZTYz85W5BTlW1mUFAKCXcndlSbmMK8rr8Pxxw/KkX5/UEF4huivB5XL5n8sXZ9i1HACC36jYV90g++ubzLQ1S1KirNi8V5Z8ukOeuuwEufeNL8wa6N4P87MnjZChuRlhvW70PjIXAMKTuX/9aJvcP3mkLFi1pdUarfqAP3vScCkgc2MOmQsAwc/cypoGqW5olv11TZ7MfXL5f83x+ZeMkgUrt7Tak0QL6A9MHikDs9LDeOXoKorobdC4AIDg2VlVK9sqamTee5vbPbRfMfYQufu1L+TXFx4tg7LSzRrouoSLjkDX14g9ZC4AhC9zb325WH58YoGcO3yAVNc3iy0tWbIzUqS/LS2s143gIHMBIHyZO33RGvN66thCk7u1Dc3mOVdHoDPzK3qwJjoAICR2O+tMz/wjbRoWakWJTnVzyUWjBsu0hZ/IuzefKicWdjzNHAAA9DxztYA+f1mJTDp2sBw9yB62awUAINYzV4vnmrnkbnRjTXQAQEg2Vvnli2tlp6Ou1fQ1b/r+qPwsGT8sj83MAADoJjIXAIDIy1xF7kY3iugAgNBsrFJSIfVNLZ2eP3fySKa0AQDQDWQuAACRl7l6XAvo5G50YzkXAEBQlR9o8OxMnprsv+92cHY6m6oAANBNZC4AAJGXuYV5GTJvyigK6FGOkegAgKBy1jV6Pl+zvUrGFHW81vm4olzJplEBAEC3kbkAAERe5uZkpFBAjwEU0QEAQWVLs3g+f2blFrliTGG7BobuWj570gjpb0sLwxUCABAbyFwAAEKDzI0/CS6XyxXui4gkTqdT7Ha7OBwOsdls4b4cAIi6deF0Wpv2ytvSLZKXcbC3/frFa2T5N1PdrClJZndy9+YqQ7LTJcuaQsMiDpG5ANB9ZC4CQeYCQPeRuVAU0dugcQEA3d+Z3Gys8k0jQrk3T9GguW1JsaeB4X2M9VjjF5kLAN1D5iJQZC4AdA+ZCzeK6G3QuACA7vXMT1+8plXDwrsRoZuoKO2931/XKH3SLJKXybpw8Y7MBYDAkbnoDjIXAAJH5sJbcqtXAAD0cGfytrRXXo8f1i+TxgQAAD1E5gIAEBpkLryxsSgAoFd3Ju+I9soDAICeI3MBAAgNMhfeKKIDAHp1Z/KO6LQ2AADQc2QuAAChQebCG0V0AECP6bpvuiZcR/R9PQ4AAHqOzAUAIDTIXHijiA4ACGhjlW3l1bKhzCH/3rpPNu/eb97TNeDunzyyXQPDvTM5a8QBABAYMhcAgNAgc9EVCS6Xy9WlM+MEu5YDQMfKKmukpqFZ7vnbF7KipMLz/rhvGhCDstJNQ4OdydFVZC4AdIzMRW8jcwGgY7uqamVfTYPMeXMjmQu/KKK3QeMCANor3VcjqzbvlTfW7ZRVXg0L7574eVNG0ZBAQMhcAGiPzEUwkLkA0J4WxysONMhdr68nc9EplnMBAPj1dWWNzHylWPrZ0jpsWKjlm8tNzzwAAOg+MhcAgNBx1DbKTmcdmYsuoYgOAPBpt7NOtlXUmEZFfVOL33N1ahsAAOgeMhcAgNCugb6jstYU0v0hc+FGER0A0KGyqlr5as8BqfqmUZGa7D8ydG04AAAQODIXAIDQZe70xWukZO/B3CVz0VUU0QEAHfbMz1hS3KpRsWZ7lYwpyu3wfF0rTjdXAQAAgSFzAQAIbeau2FxuZn1p7vrLXN1clMxFVBbRly9fLhdccIEMGjRIEhIS5NVXX211XPdInTVrlgwcOFDS09PlrLPOks2bN4ftegEgWum6b9qw8G5UPLNyi1wxprBdA8O9azmbrQAAEDgyFwCA0GaucufuhjJHh5k7tihX5kwaQebCI1miSHV1tRxzzDEydepUufjii9sdf+CBB+Thhx+WhQsXSmFhodx1110yYcIE2bBhg6SlpYXlmgEg2nrmtWFRUX1w8xTvRoXIFrlh8RqZOrZQpprXIkOy02WALY2GBQAAASJzAQAIT+Z65+4lJw2VRR9vk1EF2SZzdYS6Pd0iQ3OsMiTHGtbrRmSJqiL6ueeeaz46oqPQ//jHP8qdd94pF110kXnvf//3f6V///5mxPqPf/zjEF8tAESXrytrzIZmOp08/5vGgo6Ee3jKqHaNinRLkgzNtcrgbBoVAAAEiswFACB8mds2d48aZJdR+Vkmd7O0gE7uItqL6P5s2bJFdu3aZZZwcbPb7XLSSSfJ6tWrfRbR6+vrzYeb0+kMyfUCQCT1yu+raZBZr66XFSUV5r3pZxSZ6WsrSyo8I+G8GxWH9cuU/jZm+CAwZC4AiJTuq5E7XikmcxFUZC4A+M/cmobmVrmrhvXLlH59Upn1hehfE90fLaArHXnuTV+7j3Vkzpw5ptju/sjPzw/6tQJApNhRWSNfV9XJbke9zDj3SHn6suPNxinaK3/5mELTwNDGxfxlJTJt4Sfy/L9KpTAvg4d5dAuZCyCe7XbWycadTimrrCVzEXRkLoB41pXMVe7cXfjhVjlqoE2G9e9DAR0+Jbh0HZQopBuLLl26VCZOnGhef/jhhzJmzBgpKyszG4u6/ehHPzLnvvDCC13uodcGhsPhEJvNFoKfBADCY1tFtdy+dJ2s+qZXXmlj4q7zj5afPPWRaVBor/yEo/tLbUOL5GakmIYHjQp0F5kLIF6VVlTLTDIXIUTmAohXZC6CJWZGog8YMMD8unv37lbv62v3sY6kpqaaRoT3BwDEQ8982wK60mlt977xhcydPNLTK799X61pWOh0choW6AkyF0C8Zm7bh3lF5iKYyFwA8YjMRTDFTBG9sLDQFMvffffdVr3tH3/8sYwePTqs1wYAkaayuqFdw8K7gdHPlup5reuxas88AAAIHJkLAEBokLkIpqjaWPTAgQNSUlLSajPRtWvXSk5OjhQUFMhNN90k9913nwwbNswU1e+66y4ZNGiQZ8kXAIj3DUTLDzSIs65RGptb/J57oK7ZM+1NdyanZx4AgOBm7riiPDIXAIAAkbkIlagqon/yySdy+umne17ffPPN5tfLLrtMnn32Wbn11lulurparrrqKqmqqpKxY8fKW2+9JWlpbMYDIL6VVdXKjCXFsmJzuXn9t+vH+D0/My1Jxg3LkzmTRsjgbGuIrhIAgPjMXO20nn0xmQsAQCDIXIRS1G4sGiy6BIzuXs6GKwBiaV24r/YckKraRkmzJMlnpZXy3YIsszO5TmlrSxsVv500QuzpFnrmEVRkLoBYHA335vpd0q9PqtQ3tZjc1aniD7z1pc/M/c2FwyXVksjDPIKKzAUQa8hchFrMrIkOAGjv68qadgX0DWUOSU5IlFkXHG0aEt5Mr/ykEVKQm0EBHQCAAB/mdzrrzMN8QkKCbNjplOmLPpPH3i+ReycO95m5uZkpPMwDABAAMhfhwEj0NuihBxArduyrkRmvFLfaWGVMUa5cMaZQFn28TY4/JFsO72+TQVnpsr+uSWxpyZKdkSL9bSyBhdAgcwHE1HTyl4tlRcnB6eTemXvD4jVy6nfy5JrTisSSlEjmIizIXACxgsxFuETVmugAgK4v4TKzTQFduV+PKjhYQJ+28BN59+ZT5cTCnDBdKQAA0Z+5M17+XFb4yNypYwtl/rIS+cFx+XJIbgaZCwBAN5G5CCeWcwGAGOyZ1yVc2jYsvBsYo/KzzLpx44flmXXjAABA8DLXjcwFAKB7yFyEG0V0AIixteF0d3JdA90fLaBnpVtk7uSRrH0OAECQM1cNyU4ncwEA6AYyF5GAIjoAxNDUtjJHnUw5sUAKcqwy/YwisaYkdXiuFtAP65cpA7PSQ36dAADEU+amJieamV8DWIsVAICAkbmIFKyJDgAxYFtFtdy+dF2rNdB1B/KHp4wym6vUNDS3en9orpWNVQAACHLm6kZne/bXM/MLAIBuIHMRSSiiA0CUT2tz1Da2a1iolW02V1HjhuXJnEkjZHC2NSzXCwBAPGXuvRcNl2yrhYd5AAACQOYiElFEB4Ao3lhlxsvFcuu5h7drWHg3MGace4QcNdDmWcKFEegAAASGzAUAIDTIXEQq1kQHgCj0dWWN6ZnX9eCSEhL8rgu3fV+tPP+vUinMy6BhAQBAgMhcAABCg8xFJGMkOgDEwLpwY3ysC6e0UTFvyiimtQEAECAyFwCA0CBzEekYiQ4AUWR7Bw0Lpa8XrNpi1oXzppuu5GSk0LAAACBAZC4AAKFB5iIaUEQHgCiw21knG3c65UBDs8914fT9UflZrRoWsyeNYGobAAABIHMBAAidHZU1ZC6iQoLL5XKF+yIiidPpFLvdLg6HQ2w2W7gvBwBMr/yKknLTSMhMTZbqhmb5rLRSnlm5pd2UtpeuHi3NLpfY0ixiS0+WIdnWsF030BkyF0CkIXMRq8hcAJGau6WVtZKcmEDmIuKxJjoARLA9zjrZUVUrf1+3s0trw2WmJUu6JUmyrRamtgEAEAAyFwCA0CmtqJaZXVwDncxFJKCIDgAR7EB9k8x/r6TDteGUrg03f1mJZ1qbLS1ZBtMrDwBAwMhcAABC4+vKGrnDxxroisxFJGJNdACIUI6aBqnt4tpw7nXhaFgAABA4MhcAgNBl7raKGllB5iLKMBIdACJU+YEGsy6cP33SkuXv1481PfP5uRkhuzYAAGIJmQsAQOgyt6q20e85ui/J/904TjJTkshcRAyK6AAQoZx1jZKe4n/CkC3dYjZXGZSVHrLrAgAg1pC5AACELnNTk/1nblOLy3Res4EoIglFdACIoGlt2iuvjQp9UM+xpshbX+ySccPyZMXm8nbn6/sDbWlsrAIAQIDIXAAAwpe5H/xnr9lEtKNl1HQJl4LsdAroiDgU0QEgAhoV+2oaZNar61utCzd+WJ5MP6NIjhxoE3GJrCj59qF+XFGezJk0god5AAACQOYCABAZmTtyiN289i6ka6f17InDWcIFEYkiOgCEUem+Glm1ea+8sW5nu1745ZvL9TleLjxmkBxTkCWXjzlE6ptaJCvdIkNzrWyuAgBAAMhcAAAiK3NHFWTL1DGFZC6igv9FiAAAQbO9olpmvlIs/WxpHU5jUzql/Nj8LJl07GDJzUiRw/v3kaMH2WhYAAAQADIXAIDQIHMRqxiJDgBhsLOqVvbXN5led2uK/3+Kq+ub5NiC7JBdGwAAsYTMBQAgNMhcxDJGogNAiO2qqjXT1ea8uVGmLfxEahqa/J7fJ80SsmsDACCWkLkAAIQGmYtYRxEdAEJor7NOGppb5I5X13k2V1mzvcrsTN4R3VglL5ONzAAACBSZCwBA6DYRrWsicxHbKKIDQAintpUfqJftlbWt1oZ7ZuUWuWJMYbsGxriiXJkzaYTYrTQuAAAIBJkLAEDoOGob5esqMhexjTXRASAEdjvrZPf+eklKPNjA8FbT0Cw3LF4jU8cWmrXjMlKTzc7ktrRkGcTGKgAABITMBQAgdCPQdzvrpb65mcxFzGMkOgAEWem+Gvnli2tl4iOrZPu+WklNbv9PrzYw5i8rMWvHZVstMtCeRsMCAIAAkbkAAIQuc6cv+kzO/uNyMhdxgSI6AASxV760olpmvlLsWRdOGxadrQ03wJbG1DYAAAJA5gIAEBpkLuIVRXQACIKvK2tkR2Wt7HTUybSxh8r0M4rEmpJkGhYbyhwdrg03lrXhAAAIGJkLAEBokLmIZ6yJDgC9bMe+mla98u7NU+ZfMkpufblY7p88UhZ9vE1GFWSbteHqm1rEnm6RoTlWGZLD1DYAALqKzAUAILIzt4DMRYygiA4AvTy1rW3DQh18nSD/7+Shns1VRuVnmc1V0i0uGZprlcGsDQcAQJeRuQAARG7mpllcUpCdLvkU0BEjWM4FAHrJbmed7HTWtWtYuK0oKZexRXmezVWe/XCrDLClytGDbDzMAwAQADIXAIDIzdz+fVKlqF+m5OdmhPx6gWBhJDoA9ILtFdVSWlkryYkJ8szlJ8hnpZXyzMotpiHhranF5VkXbvakEWZqGwAA6DoyFwCA0CBzgW9RRAeAHjI7ky9dJ6u8euZ1M5WHp4wyU9q8Gxh9UpPluZ+dxLQ2AAC6gcwFACA0yFygNYroANDD3cnvaNOwUO7XuiacTmlzNzjSU5Ikr0+q9LelheV6AQCIVmQuAAChQeYC7bEmOgB0c2OV/+zaL/tqGnyuDacNDN1Uxd2wmH76MEm3JNGwAAAgAGQuAAChQeYCvjESHQACVFZVKzNe1p3Jy+XRn3zX77nWlGR5+rLjZY+zTgZnpcnArPSQXScAANGOzAUAIDTIXMA/iugAEGDDYmt5tUw5qUCuGFsotjT//4xmpiVJdoZFjhxok0E0LAAA6DIyFwCA0CBzgc6xnAsABLAzuTYsqmobJc2SZHYm/+/eAzKuKK/D83Vn8qw0ixwxgIYFAACBIHMBAAhd5n6waY/UNn67UWhTSwuZC7RBER0AutiwKK2sbfUwv6HMIX37pMn0M4pkXFFuq/PHDcuT304aIfm5GWG7ZgAAohGZCwBA6DYQddY1ST9bmiQkJMiGnU6ZvugzeXL5f+XuC48icwEvCS6XyxXui4gkTqdT7Ha7OBwOsdls4b4cABFgW0W13Ll0XauNVXQDlSvGFMqij7fJMflZcvrh/WTv/nrJSE2WlhaXDM21yuBsa1ivG4h0ZC6AtshcIDjIXACBZO4Ni9fIyYfmyMzvHymlFTVkLsBIdADofDTc7W0aFu4dyRes2iJHDbLL8EF22VFZK8+u2mrWjjusXyYNCwAAAkTmAgAQGjv21fjN3KljC2XZl3ultqGZzAW+QREdADrgqGmQreUHZHtlrWlIdETfH5WfJfVNLWJPt5jpbjnWFOlvSwv59QIAEM12O+vIXAAAQpS5pftqOs1cpUV0Mhc4yP92uwAQh8oqa8w6rCIJ3/zqmz7MZ6VbJDExQWxpFrOWHAAA6Hqn9b7qBqltaiFzAQCIoMxVmanJZC7wDUaiA4CX7ftqTK/87Dc3ytaKaklN9v/PpI6Gy85IkUNyrDQsAAAIwNf7amTjTqfc+dp62VpO5gIAECmZq8fHFeVJtpUCOuBGER0AvrGjskb21zXKI++VmCls2nBYs73KbK7SEW1U5GenyyB7mgzISg/59QIAEM2Z6wgwcwvIXAAAQpK5e/bXy5yLR8gg1kAHPCiiA8A3G6vMXFIsOx11ns1VtGHxZZnD7E7etoExtihX7r7waDMqzm5NCdNVAwAQO5m7wU/m3jdpuOTnZpC5AAAE+Tn3NxcdLad9p68MyaGADnhjTXQAcU83Vpn5SrFpVEw5aajn/WdWbpH5l4ySRR+VyqiCbJk6ptCzoVlORor0SUniYR4AgF7K3IenjJJFH29rlbm6BrrO+tICOgAACPJzrjVFLIkJMpBZX0A7jEQHIPHeM//VngOeXnnvteFqGppl+qI1MnyI3fTEZ6Qmy9BcqwzOSjPTyfvTsAAAoNcy94bFa+SoQXYZlZ9lHuaH5ljlsH6ZFNABAAj2c26O1Tzj5mamyBByF+gQRXQAcevryhqZ8Upxq53J264Npw2MP/5zs/zg8dVmDbkhWelySF4mI9ABAAhC5s5fViLTFn4ii/9VKllWi/RnMzMAAIL+nKuZW9g3k01EAT8oogOIO46aBtm8e7/sq2nwbKziPbWto7Xhxg3LkzmTRlA8BwCgG9PJq2oaA8rc2ZNGyGA2MwMAICBkLhA8Mbkm+iOPPCIPPvig7Nq1S4455hiZN2+enHjiieG+LAARoKyqVmYsKZYVm8vl0Z98t1WvvDY03NPJp44tNGvDqcHZ6WJLTWZncgAAupO7L3/uWYuVzAUAIDjIXCC4Ym4k+gsvvCA333yz3H333fLZZ5+ZIvqECRNkz5494b40ABEwAr10X43ces7hsvjKk6UwL6PDXnn3dPJnP9wqh+ZlSK41hYYFAADdeJivrG6Qa08fRuYCABBEZC4QfAkul8slMeSkk06SE044QebPn29et7S0SH5+vlx//fVy2223dfr1TqdT7Ha7OBwOsdlsIbhiAKGyvaJanHVN5qNPWrK0uFzyp3/+R979cq9YU5JMr7x7M7OsdIskJibIITlWGcAGokBQkLlA7CJzgchC5gKxi8wFQiOmlnNpaGiQTz/9VGbOnOl5LzExUc466yxZvXp1WK8NQHhtq6iW25euM1PZ3MYW5cq9E4eLyAbTwNBeeff7ui6cNjBYAx0AgMBmfe101snXlbWSkJAgn5VWmpFw3y3IInMBAOhlPOcCoRNTI9HLyspk8ODB8uGHH8ro0aM97996663ywQcfyMcff9zua+rr682Hdw+9jlynhx6IDXuddVLX2CyllbXiqG2UNEuS54Fep7NpQ+LWc46Q5KREcdY2SkuLS4bmWtlYBQgCMheIbbpk2h2vFMsKrwd5nUKuU8l1HVYtpLszd2t5tQzNsUqW1ULmAkFA5gKx32mtz69dec4lc4HeEXNrogdqzpw5Zlqb+0MbFgBiw46KajlQ32R65n/y1Mdy7XOfydRn/y1rSivl4SmjzNS2lSUVkpAgpgEy0JYmRw+y0bAAgoTMBWJ7KnnbArrSkXELVm0xU8ndmeuoaZTF/yqVvD6pZC4QJGQuELvKKmtkX3VDl55zyVyg98RUET0vL0+SkpJk9+7drd7X1wMGDOjwa3TpF+2Nd39s3749RFcLINjT2laUlMudr633+0CvDtQ1iz3dIkPzMpjWBgQRmQvEbuZur6xtl7feuatrsboz15aWLA9MHin9bWkhvlIgfpC5QOxm7gf/2dvl51wyF+g9MVVET0lJkeOOO07effddz3u6sai+9l7exVtqaqqZzub9ASC6bd9XY3rl+9nSWq0N5+uBXhsWORkUz4FgI3OB2M3cqtpGv+fpZmbuzM3OSJGBbGYGBBWZC8SeHd14ziVzgd4TU0V0dfPNN8uf//xnWbhwoWzcuFGuueYaqa6uliuuuCLclwYgBOvCbd69X6pqG0zjwf3A7ose17XitHFBzzwAAN3P3NRk/48VetyduYN4mAcAIODMrezGcy6ZC/SeZIkx//M//yN79+6VWbNmya5du+TYY4+Vt956S/r37x/uSwMQ5F553dBMR8LlZ6fL9DOKxGpJ8vs1uoSL7k6en5sRsusEACDafb2vRpZv3mtGwllTkuWZy08Ql8slZxzRV5Z9ubfd+bq56B5nHZkLAECAdlbVyvubNHNTPZmbl5li1j3XDUQ7wnMuEBwxV0RX06dPNx8A4mddOJ3W5j2lTR/Yzzqyv88H+nHD8uSQXCubqwAAEOBmZqWVNfLGup3tcnfW+Uebz71zV/P2NxceLWmWJKaTAwAQ4Aj0bRWauWWtMndcUa48fdnxMm3hJ+0K6Zq7Q3OtMoTnXKDXxWQRHUD82O2skzvbFNDVwddfyoxzjjDT2Vo1OoblyeyJwymgAwAQ4MO8zvia/15Jh7n7mze+kKljCuUnJw012ZuVbpEB9jRJtyTJAAroAAAEZF91g8x7b3O7zD24oWiC3HnekXL70vWtiuv3TRxOAR0IEoroAKL6YV4bFm13JXfTxkb5gQYZVZBtHurV4Ox0ybamsAY6AAABZu5OZ500tbj8bmameasj43QtVp1KroV0u5XNuwEACCRzK2sazShzX5m7oqRcbjxrmBmRrnTt88zUZMnPoYAOBAtFdABRqayqVj74z14p7GSdN10rbmxRnrS0uEyDgkYFAACB+bqyRvYeaJC6hmZJTEzwe64+wC/62UlkLgAAPcjcigP1Ykuz+D23ucVlZnsNyU4366DTaQ0EF0V0AFFZQK+sbpBD8zLElu6/YaHF88FZ6ZJlpVEBAECgdNPuO14p9sz6co9488VutchAWxqZCwBAgMhcILIlBvoF7733nvz+97+XVatWmddPPPGEFBQUSN++feXKK6+U2traYFwnAHg2EZ3x8udy3ryV8j9PfiR/X7fTTBnviL5fkGOVoXkZNCwAAAjQjjYP82rN9iqziWhHdM8RHuYBAAhB5hblmWVKyVwgQovof/7zn+V73/uePP7443LmmWfKnDlz5JZbbpHzzjtPfvSjH8mLL74o99xzT/CuFkBcrwtXWlFtNhH1blg8s3KLXD6msF0hXRsVuhbrEKaSAwAQsO0V1bK/vqndviOau1eMKWz3UK8F9Acmj+RhHgCAYGeu2XdkOPt8ASGW4HK5XF09efjw4XL11VfL9ddfL2+99ZZccMEF8tRTT8lll11mjr/00ksyc+ZMKSkpkWjldDrFbreLw+EQm80W7ssBICK7dPmWmkZpcbnMCPSO1j2fOrZQzhsxUKpqGiUzLUmq65vlyAF9eJgHIhiZC0Tuw3xpZa0kJiTIlD9/5DN3zz16gHnoT7MkyUB7Gg/zQAQjc4Hoz1xnXZPY0pIlOyPFbCQKIIJHov/3v/+VCy+80Hx+zjnnSEJCgpx44ome4yeddJJs3769968SQFyPQN9X0yD3vblBtu2r6fAc3bV8/rISU0DXhseDb22SoTlWCugAAARIZ33dtnSd/OSpj6Wmoclv7u5y1smj75VI/z6pFNABAAhy5j72fonkUEAHoqOIXldXJ+np3/7Pmpqaaj68Xzc1dfw/PgB0hxbG57y5UVaVVEhqcmKnG6v846ZxMm/KKBlIwwIAgIA7ru9Yus5kblfWP9cNvslcAAACR+YC0Sc5kJN15Pn+/fslLS1NdBUYfX3gwAEzNUy5fwWA3mhU7NlfLw3NLZ614dwNC3dDo+0a6FnpFnrlAQDoht3OOtlX3dBu35GHp4wyn3tnr3vfEd28GwAABIbMBeJgTfTExERTOHdzF9Lbvm5ubpZoxVpxQGTsTF66r0aqahtlaK5V/m/9LtOoUNqwWLBqS7uGxX2ThsvQ3IwwXjWAQJG5QGQ8yFfWNMjXlbXSr0+q/GPDbpO5On3cey3WUflZYk1JliyrRfqkJrNxNxBlyFwg/MhcILoFVET/4IMPunTeqaeeKtGKxgUQXtsqquV2r2ltSkef667kNyxeY167Gxb1TS1m7XPdXCWfAjoQdchcIPxrsepU8hU+Mtf9UO/23M9OkoLsdDIXiEJkLhBeZC4QZ0X0ztTU1MjatWvllFNOkWhF4wIIn7KqWtlaXm1GoKdZkuSz0kpPz7w2MEYVZJtNVdzGFuXKnEkjaFgAUYrMBcKbuVvKq8VB5gJxgcwFwofMBeJwTfTObN68WcaNGxfVy7kACN8SLjNfKW7XM6/Lt2jPvI5MnzqmsNXmKro2XD5T2wAACH7mThzOwzwAAAEic4HY0atFdADozgai2iPfdmqbci/posu3aM98RmqyPH/VyWYD0YH2NLFbU8J01QAARJ89zjqpbWyWO7uYuY/+5LtSmJchg8hcAACC+pxL5gKRLzHcFwAgfm3fVyM7KmvFUdfYrmHh3cDQ9c9Vc4vLTIOjgA4AQGB2VFRLxYF6OVDX1KXMra5vkkUfb5OcjBQyFwCAAOyqqpWyqjrZT+YCMYUiOoDwbSD6SrGcN2+lbN9X6/dc3UBU14XTRsXp3+lLwwIAgABHoDvrm+S+NzfKtn01nWauTjPXr/ntpBHS35YWsusEACAWRqDvq2mQ+97cQOYC8bycy+uvv+73+JYtW3p6PQDiwI7KmlZTyVOT/ffn6fIt900cIdlWCwV0AAACoA/mOsJtzpsb26276jtzh0tGSrL042EeAIDACujVDWQuEKMCKqJPnDix03MSEhJ6cj0AYtz2impxtpnWtmZ7lemBd68N5003VjlE14bLSg/xlQIAEBuZq9y5S+YCAND7vq6skaqaRvM5mQvEpoCWc2lpaen0o7m5OXhXCyCqe+U37XLKf/YckBaXS6afUSTWlCRz7JmVW+SKMYWmgeFtXFGezJk0goYFAADd2Hdke2WtmUrunbtkLgAAvWvHvhrZVlFD5gIxLsHlcrkC/aKKigrJzT34j8D27dvlz3/+s9TV1ckFF1wg48aNk2jmdDrFbreLw+EQm80W7ssBYkJZVa3MeLlYVpSUe97ThoQ2KG5YvEZqGppNI0N3J9fNVXR38pYWlxTkWmVItjWs1w4geMhcIDhKK6rlDq9l09rmrnJnrq7HOjTHKnarhcwFYhiZCwQHmQvEj4CK6OvWrTOFci2cDxs2TJ5//nk555xzpLq6WhITE82vL7/8cpeWfYlUNC6A3h+BPn3RZx3uSq6Ni1EF2TJ/WUmr984fMcj8WpCbEeKrBRBKZC4QnCVcblu6rsOp4x3lrm7craPh8slcIKaRuUDvI3OB+BLQci633nqrjBgxQpYvXy6nnXaanH/++XLeeeeZIK6srJSrr75a7r///uBdLYCos8tZ12EBXWljQ3vkvae13XvRcBlHAR0AgG5NJy+trO3wYb7D3B2WJ7N5mAcAIGBkLhB/AtpY9N///rcsW7ZMRo4cKcccc4w8+eSTcu2115pR6Or666+Xk08+OVjXCiDKRqBX1jR4NlfxpU+aRR79yXfNpirpyYmSZkmSgawNBwBAwJlbVlUnjlr/uatLpr149WjJSEliOjkAAN1crrR0Xw2ZC8SZgEai79u3TwYMGGA+z8zMlIyMDMnOzvYc18/379/f+1cJIOoaFdMXrZGv9lZL+jebh/qSZkmURR9vE6slyXxQQAcAoHuZW1XbKKnJ/pv3zS0u+WrPAemTlszDPAAAAdpWUS1bygPL3EwyF4i/IrpKSEjw+xpAfNPRcB9s2iuXjzlErCkHJ7u03Y3cTd9PTkww09qyrRbpTwEdAIBuZ64+zK/ZXuUzd3XZtPzsdDnzyH4smwYAQIB2VNbIjspaSUxI6HLmnnZ4XxlK5gLxt5yLuvzyyyU1NdV8XldXJz//+c/NiHRVX1/f+1cIIKrsq2mQN9aVmTXgnr7sePnX1gqZfnqROea9Xpw2Nq4/Y5hkW1MYfQ4AQC9krj7MbyhzyBVjCtvlrm5mNnvScIrnAAB0cw30ma8Um/2+yFwgPiW4XC5XV0++4oorunTeggULJFqxaznQPbuddWY91q8ra80Mlc9KK8WSlCCfb6+SS08+RHY6aqW/LU3qm1pMr/0eZ52M1d55GhZA3CJzge6PQHfWNsr2yloznVz3E9Em/cufbpfJx+WbZdKOGmQ3G5pp7trTLVKQnU7mAnGMzAW6h8wF0K2R6NFcHAcQPKUV1TJz6bp2I82njS2UkYOz5C8fbTUNCy2iq3RLkowb1leG5LAuHAAAgdhZVSvb9tXII8s2m9Fw3rk76/yj5Y//3NTqYT5LH+ZzrGQuAAABInMBdHskejyghx4IfF24bRUHdybXXnkdgf7Myi1S09BsGhcnHJIjTS0u07DQ3cn1nIH2NE9BHUD8InOBwHxdWSN79zdIRXW9Z9aXO3OV5u7Ub6aVk7kAvJG5QPAyV/claXG5pDAvQwaxVCkQswJeEx0AvHcmv3Ppuna98g9PGSU3LF5jRqZrw2Lawk/Msb9fP9b0zvMwDwBA72auPtR75y6ZCwBAaDL3uZ+dJEOy0ymgAzEuMdwXACA6bd9XI7e3aVgobUwsWLVFpo492Cuv09rUuGF5kmW1yGCmtgEAEJTMdecumQsAQOgyd2iuVYayBjoQ8yiiA+jW5irauPBeA92bvq/LtyjdRFQbFnMmjZDB2TzMAwAQrMxVOvqczAUAIHSZO4TMBeICy7kACKhRUX6gQeqbms3O5P6YXvmiPDmsb4bMnzJK7NaUkF0nAADxlrlKO60P65fJEi4AAASAzAXQFYxEB9AlZVW1Mn3xGjnzoQ9ka0WNGWHujz3dIrMnDZdD8jIpoAMAEMTMdc/6emDySB7mAQAIAJkLoKsoogPoUs/8jCXFsmJzuafhsGZ7ldlcpSPudeEKWBcOAICgZ+6heQdnfQ1kQzMAALqMzAUQCJZzAdBpw2Kno06mnFggV4wplM9KK2V9mUM2lDnMa+W9ZtzYoly5b+Jw1oUDACBEmctmZgAABIbMBRCoBJfL5Qr4q2KY0+kUu90uDodDbDZbuC8HCGujoqqmUe58tfXO5NorP21soSRIgvzlo61y1CC72VxF14bTjVXyc6zmAwA6Q+YC3yrdVyN3vFJM5gIICjIX+BaZC6A7GIkOoJ2dVbVmQ5X7/r6h3c7k7tcnHJLjaVjo+ufZ1hTJy0xh/XMAAAK0vaJaZi5dR+YCABBkZC6A7mJNdADtRqC//5+9sq+6oV3Dwk3fHzHYLvOXlcjCD7fKsH6ZZmdyGhYAAARmt7NOSitryVwAAIKMzAXQE4xEB9CqgK7F8+GDbbJ9X63fc3Va2/hheTJ38kgaFQAAdHPZtJrGZnHUNvo9l8wFAKD7yFwAvYGR6AA8S7hs3LVfZr223hTQdWdyf3RX8nnsSg4AQLcz9/ZX18mW8moyFwCAICFzAfQWiugApKyqVmoamuWRZZvN5irasFizvcpsrtKRccPyZKA9jZ55AAB6kLk6bZzMBQAgOMhcAL2JIjoQ50orquVXL38uWyuqPbuTa8NiQ5lDrhhT2K6BMbYoV+ZMGkHDAgCAEGTubDIXAICAkbkAehtrogNx3jPv3pn8JycN9bz/zMot8vCUUbLo420yqiBbpo4pNGvDZaVbZGiuVQZnW8N63QAAxHrm2tMtUpCdLvk5ZC4AAIEgcwEEAyPRgThuWFRWN3h2JvdeG06nvN2weI0cNcguo/KzTMNC14Y7epCNAjoAAEHO3MK8DCnqlyn5uRlhvGoAAKIPmQsgWBiJDsThzuR79tdL6b4ayUz99p8A99pw7saGNjDmLysxn+vu5Lq5ClPbAADoOjIXAIDQIHMBBBtFdCDOeuVnvFwsK0rKzeu/XT+m3dQ25W5guDdXmTt5JA0LAAACQOYCABAaZC6AUKCIDsRRz7x3w0LtcdabDVRWllR4prZNHVto1oZTQ7LTZYCN3ckBAAgEmQsAQGiQuQBCJcHlcrlC9rtFAafTKXa7XRwOh9hstnBfDtArdjvrZF91g1TVNEqftGTzesaSYnPsuZ+dLPe+8YVpYLhpg+O3k0ZIAevCAQgiMhexiMwFEInIXMQiMhdAKFFEb4PGBWLNtopquf2bncm9Gw93nX+0/OSpj8xrncY20J4ujtpGsaUlS3ZGigzKSg/jVQOIB2QuYg2ZCyBSkbmINWQugFD7dptiADFnR2VNu4aF0t547ZXXRkX5gQaZtvAT2emolcfeL5EcGhYAAASMzAUAIDTIXADhQBEdiFE6lW1/XVO7hoV3A6OfLdV8Pq4oVw7NyzA7kw+kYQEAQEDIXAAAQrcGOpkLIByipog+e/ZsOeWUU8RqtUpWVlaH55SWlsp5551nzunXr5/86le/kqamppBfKxAJu5P/8sW1sqW82u95B+qaZUxRrlx3xjDTM8/GKgAABIbMBQAgdCprGk32+kPmAojrInpDQ4P88Ic/lGuuuabD483NzaaArud9+OGHsnDhQnn22Wdl1qxZIb9WIJy0QTHj5c9lRUmFpCb7/1/cbrXI+SMHySE5VhoWAAAEiMwFACB0s7627D0gd726rtNzyVwAwRA1RfR77rlHfvGLX8iIESM6PP7222/Lhg0b5K9//asce+yxcu6558q9994rjzzyiCmsA/GgtKLajITTh3m1ZnuV6YHviG660iclSb4/fIAMYGobAAABIXMBAAhd5t784lr57ze5S+YCCIeoKaJ3ZvXq1abA3r9/f897EyZMMLuQf/HFF2G9NiAU68L9Z9d+2bzngCQlJnjef2blFrliTGG7BoY2LGZPGiH5uRn0zAMAEAAyFwCA0GfutLGHiiUp0W/mjivKI3MBBE2yxIhdu3a1KqAr92s95kt9fb35cNOiOxBtO5M7ahrFWdck/W1pYklKEGtKktQ0NJuPGxavkaljC2XqmEKpb2qRwrwM6ZOWLEOyreG+dABxhsxFtPu6skb2HmiQuoZm6dcnVdJTkslcABGJzEWsZa5IgrhcLnOso8zVZdV0E9GhuRnhvnQAMSqsRfTbbrtN5s6d6/ecjRs3yhFHHBG0a5gzZ45ZKgaIRtsqquX2peta7UyuO5A/c9kJMnXhvz0P9fOXlRw8NixPfvfDY0yxHQBCjcxFNCvdVyN3vFLsWb7FnblPX3a8TFv4CZkLIKKQuYi1zNVR53d8/0g564h+8s8v97TKXDV+WJ7MmzIqTFcMIB4kuNxdeWGwd+9eqaj49h/Fjhx66KGSkvLtNBzdLPSmm26SqqqqVufpBqKvv/66rF271vPeli1bzNd/9tlnMmrUqC730Ofn54vD4RCbzdaDnw4Iru0V1XJbmwK69zS2748YIDOXrv/2vWF58sDkkTKQdeEAhAmZi2i1Y1+NzGzzMO9G5gKIRGQuYjFztZB+5/ePlN/+35eyYnN5qwL6XHIXQCyPRO/bt6/56A2jR4+W2bNny549e6Rfv37mvXfeecc0EI466iifX5eammo+gGhbwkWXb+mogK5WlJTLzO8fKa9PHyM7KmvFnm4xU8ppVAAIJzIX0TqdvPxAQ4cP8+7MvfGsYfLmDeNka0U1mQsgIpC5iMXM1effnc56mXXBUVJaUWPey8+2Sn9bKmugAwi6qNlYtLS01Iwy11+bm5vN5/px4MABc/zss882xfJLL71UPv/8c/nHP/4hd955p1x33XU0HhBzm6u410D3e15to+zdXy+LPt4m+dnpMoiHeQAAAs7cbRU1UtfY7Pe8PfvrxVnXSOYCABDkzNXjFQcaZMGqLTKsX6Z8Z0AfCugAQiJqNhbV5VoWLlzoee1enuW9996T0047TZKSkuSNN96Qa665xoxKz8jIkMsuu0x+85vfhPGqgd6nPfNaQM9MS/J7ni0tWQbYUuX3PzqW9VgBAOhm5lbVNsqQbP9Fcd3MLMtqIXMBAOgmMhdApIuaIrquha4f/gwdOlTefPPNkF0TEOqeeW1YVFQ3mEZDY3OLjC3KlZUdTHXT9+1WiwzJtoblWgEAiJXMLcixSqol0Wwi6mt9Vh2JfvzQbEbCAQAQIDIXQLSImiI6EK92O+uksqZBvq6slYSEBPmstFKeWblFRh+aI/dOHC53vbq+VSFdC+izJ42ggA4AQDce5Hc568x+It6Ze8qhOXL3hcPlnte/MGugez/MX3/6MLOECw/zAAB0HZkLINokuFwuV7gvIpLoruV2u51dyxERSiuq5Y6l61r1wmvj4YoxhXLD4jWmkH7jWd+RhASRA3XNZgkX/cjPzQjrdQNAV5C5iCRlVbUy4+Xidg/s7sw95dBcuerUQyUxIUGaWlxiTUmS5MQEyUq3yCA6rgFEODIXkYTMBRCNomZjUSAeGxYz2xTQ3TuS6yYqU8cWyrtf7pXkpERPAV2XcKGADgBA4LO+Zrz8eauH+baZ+88v90hmqkVSkhMl4Zs1WQdnpfMwDwBAAMhcANGKIjoQoQX0yuoG05DoiL4/Kj/LfH6grkn69Uk1G7CwhAsAAIFn7ld7DnS49mq7zK1vkoyUZDliQB85fICN6eQAAASAzAUQzSiiAxG4NtyMJcXirGvye159U4v5NTcjRQ7rl0mjAgCAbmZuVW2j3/PIXAAAeobMBRDtKKIDETa1rcxRJ1NOLJAsq0Wmn1Fk1n/riE5pGz8sT/IyaVQAANCTzC3IsXaauePIXAAAuoXMBRALksN9AQAO2lZRLXe2WQN9bFGuPDxllNlcpaahudWmK3v218vcySPpmQcAINiZ66yTOZNGkLkAAASIzAUQKyiiAxFgx76adg0LtfKb17q5yvxlJebzcUV58puJR0uONYWGBQAAQc3cXLn7wqPFakliMzMAAAJE5gKIJRTRgQhoWGzbV+NzcxVtYNx+3lEy5rA8saUlS3ZGigzKSg/5dQIAEA+ZO/P7R8rYojzJSE2SPqnJkk2nNQAAASNzAcQa1kQHwry5ysxXisXRyeYqW8ur5bH3SySHAjoAAEHN3G0VNfLoeyWSl5Eqh+SxoRkAAIEicwHEIoroQBiVH2gwPfO6eYo/hXkZMm/KKBlIAR0AgG4hcwEACA0yF0AsoogOhJGz7mDP/JrtVWYTlY7oGug6Ap1eeQAAuo/MBQAgNMhcALGIIjoQRrY0i/n1mZVb5Ioxhe0aGNqwmD1puPS3pYXpCgEAiA1kLgAAoUHmAohFCS6XyxXui4gkTqdT7Ha7OBwOsdls4b4cxMFacdcvXiPLN5eLNSXJ7E4+Kj9L6ptaxJ5ukaE5VhmSw87kAGITmYtQInMBxDMyF6FE5gKIRRTR26BxgVArq6qV25YUmwaG27hheTJn0ggaFgBiGpmLUCNzAcQrMhehRuYCiDUU0dugcYFw9dTr5iv76xqlT5pF8jJZGw5A7CNzEQ5kLoB4ROYiHMhcALEkOdwXAMRDo0E3VrGlWyTPx8Yp+h6NCQAAuo/MBQAgNMhcAPGIIjoQpEbFLmed7KislYSEBPmstNJsqnL80Gy5f/JIGZSVHu5LBAAgJpC5AACEBpkLIJ6xnEsbTHNDb6z9NuPlYllR4rX2W1GuXDG2UKYvWmMaGPOmjKJHHkDcI3PRU2QuAHQNmYueInMBxDtGogO93DPftmGhVpRUaJ+VXDX+UPnjPzebqW80LgAA6D4yFwCA0CBzAUAkMdwXAMSSPfvr2zUs3PT9sUV55nPdWAUAAHQfmQsAQGiQuQBAER3oVVW1/hsNTS0HV0/SnckBAED3kbkAAIQGmQsAFNGBXpWRkuT3uDUlScYPy5O8TKa4AQDQE2QuAAChQeYCAEV0oFdlpCTLmKLcDo/p+8mJCTJ38kjWiQMAoIcyyFwAAEIig8wFADYWBXqyuYpunOKsaxRbukXyMlIky2qR688YZo6vMpusfNuw0PfzMlOlny0tjFcNAED0IXMBAAgNMhcAOpbgcrkOLl4Fw+l0it1uF4fDITabLdyXgwhtVDhqG2VHZa1ZGy7NkiSflVbKpp1Oueei4WZ6x3v/2Sv9+qRKfVOLpCYnmo1YTv9OXxmQlR7uyweAiEHmojNkLgD0DjIXnSFzAcA/iuht0LiAPzuraqWypkFmv7mxXQ/8FWMK5YV/lcrvfniMeU9773V3ct1cRdeGY2obALRG5sIfMhcAeg+ZC3/IXADoHMu5AAH0zL//n73yZnFZq4aFcr8eVZBtGhWH9cukMQEAQDeRuQAAhAaZCwBdw8aiQBdpo0Gnrq1o07DwbmCMys8yvfIAAKD7yFwAAEKDzAWArqGIDnSRbqyia7/5o8d1WhsAAOg+MhcAgNAgcwGgayiiA11kS7OYzVP8ydLdyzOZ3gYAQE+QuQAAhAaZCwBdQxEd8LEu3Fd7Dsia0kr5au8B81obDbr7uG6u0pGxRbkyNNfKGnEAAASAzAUAIHy5m5mWTOYCQBewsSjQRllVrcxYUiwrNpd73hs/LE/mTh4pp32nrxTmZZj3vDddGTcsT+ZMGiGDs61huWYAAKIRmQsAQHhz93tH9pNZ5x9F5gJAJxJcLpers5PiidPpFLvdLg6HQ2w2W7gvB2HolZ++eE2rRoX3Q/28KaPM51U1jVLd0CQ1Dc1iT7eYjVjomQeAwJC58Y3MBYDQIXPhL3e1kD570gipbWgmcwHAB0aiA212Ju+oUaGWby43xw/rl0lDAgCAHiJzAQCIjNx9Z+Meue3cJpO7AICOsSY64LUuXEV1gzxz+Qky/YwisaYktTtvf11jWK4PAIBYQeYCABD69c/rm5p9Zq4idwHAP0aiI+51tC6cbqry8JRRcsPiNWYqm1ufNEuYrhIAgOhH5gIAEHmZq8hdAPCPkeiQeO+Zb9uwcG+msmDVFpk6trDV+qx5mUwpBwCgO8hcAAAiL3MVuQsAnaOIjrjmb104bWCMys/yNCrmTh7JuqwAAHQTmQsAQGRlriJ3AaBrWM4Fcc3ZybpvOqXt3ZtPNb3yNCoAAOg+MhcAgMjIXHu6RV699hSTveQuAHQNRXTENVsn677lZqSwQzkAAL2AzAUAIDIyN9tK5gJAoFjOBXFNe911+lpHWBcOAIDeQ+YCABAaZC4A9D6K6IhrOm3t/skj2zUwWBcOAIDeReYCABAaZC4A9L4El8vlCsL3jVpOp1Psdrs4HA6x2WzhvhyEcPdy3Xxlf10j68IBQIiQufGJzAWA0CNz4xOZCwC9hzXREReNBt1YxZZukbyMjhsN+h6NCQAAuo/MBQAgNMhcAAg9iuiIWWVVtTJjSbGs2FzeavqaTmsblJUe1msDACCWkLkAAIQGmQsA4cGa6IjZnvm2DQu1fHO53Lak2BwHAAA9R+YCABAaZC4AhA9FdMQkndrWtmHh3cDQ4wAAoOfIXAAAQoPMBYDwoYiOmKRrw/mjG6sAAICeI3MBAAgNMhcAwociOmKSLc3i97juTA4AAHqOzAUAIDTIXAAIH4roiEl5mSlmc5WO6Pt6HAAA9ByZCwBAaJC5ABA+FNER1XY76+TLnU7515Z98uUup3mt7NYUszt52waGvp47eaQ5DgAAuo7MBQAgNMhcAIg8CS6XyxXui4gkTqdT7Ha7OBwOsdls4b4c+FFaUS0zl66TVSUVnvfGFuXKbyeNkILcDPNadyfXzVV0bTid2qY98zQsACAykLnRg8wFgOhG5kYPMhcAIhNF9DZoXEQH7Ym/+cW1rRoW3g2M3//oWOlvSwvLtQEAuobMjQ5kLgBEPzI3OpC5ABC5omI5l61bt8q0adOksLBQ0tPT5bDDDpO7775bGhoaWp1XXFws48aNk7S0NMnPz5cHHnggbNeM4KqsbuiwYaFWllSY4wAAoOfIXAAAQoPMBYDIlSxR4Msvv5SWlhZ54oknpKioSNavXy9XXnmlVFdXy+9+9ztPz/rZZ58tZ511ljz++OOybt06mTp1qmRlZclVV10V7h8BvcxZ19Sj4wAAoGvIXAAAQoPMBYDIFRVF9HPOOcd8uB166KGyadMmeeyxxzxF9Oeee86MTH/mmWckJSVFjj76aFm7dq089NBDFNFjkC0tuUfHAQBA15C5AACEBpkLAJErKpZz6Yiu5ZaTk+N5vXr1ahk/frwpoLtNmDDBFNsrKyt9fp/6+nozit37A5EvOyPFrAnXEX1fjwMAIguZG53IXACIPmRudCJzASByRWURvaSkRObNmydXX321571du3ZJ//79W53nfq3HfJkzZ47ZYMX9oWupI/LpZiq6O3nbBoZ713I2WwGAyEPmRicyFwCiD5kbnchcAIhcCS6XyxWu3/y2226TuXPn+j1n48aNcsQRR3hef/3113LqqafKaaedJk899ZTnfV0PXTce1XXT3TZs2GCWddFfjzzySJ899Prhpj302sBg1/Lo2b1cN1fRteF0apv2zNOwAIDIROZGNzIXAKIHmRvdyFwAiDxhXVDrlltukcsvv9zvObr+uVtZWZmcfvrpcsopp8iTTz7Z6rwBAwbI7t27W73nfq3HfElNTTUfiOBGQ3qyZFs7bjToezQmACA6kLmRicwFgNhD5kYmMhcAoldYi+h9+/Y1H12hI9C1gH7cccfJggULJDGx9Uo0o0ePljvuuEMaGxvFYrGY99555x05/PDDJTs7OyjXj+AoraiWmUvXyaqSinbT1wpyM8J6bQAAxBIyFwCA0CBzASC6RcWa6FpA1+VbCgoK5He/+53s3bvXrHPuvdb5JZdcYjYVnTZtmnzxxRfywgsvyJ/+9Ce5+eabw3rtCLxnvm3DQq0sqZDbl64zxwEAQM+RuQAAhAaZCwDRL6wj0btKR5TrZqL6MWTIkFbH3Eu662Ypb7/9tlx33XVmtHpeXp7MmjVLrrrqqjBdNbpDp7a1bVh4NzD0ONPaAADoOTIXAIDQIHMBIPpFRRFd103vbO10NXLkSFmxYkVIrgnBoWvD9eQ4AADoGjIXAIDQIHMBIPpFRREdsctR0yDlB3RjlUaxpVvM5ir+6M7kAAAgcGQuAADhyd0+nWQqmQsAkY9/qRE2ZVW1MmNJsazYXO55b87FI2RcUZ6sKPn2Pe9NV7IzUkJ8lQAARD8yFwCA8OXu9DOKTLbq0i1tkbkAEB2iYmNRxGavfNuHeXXvGxvkujOKZFxRbqv33buWs04cAACBIXMBAAhv7j6zcotcPqbQZKw3MhcAogcj0REWOq2t7cO8qmlolqnP/lvevGGs1DW2mLXhdGqb9szTsAAAIHBkLgAA4c1dzdwbFq+RqWML5Y7zjpL9ZC4ARB2K6AgLXRfOF21gVNU0yrEF2SG9JgAAYhGZCwBA+HNXM3f+shI564h+cmJhTsivCwDQMyzngrCwpVn8Hu/TyXEAANA1ZC4AAKFD7gJAbKKIjrDIy0yR8cPyOjym7+txAADQc2QuAAChQ+4CQGyiiI6wsFtT5P7JI9s1LvT13MkjzXEAANBzZC4AAKFD7gJAbEpwuVyucF9EJHE6nWK328XhcIjNZgv35cTFzuW68cr+ukYzrU175WlUAEB8IHNDi8wFgPhF5oYeuQsAsYWNRRFW2oigIQEAQPCRuQAAhA65CwCxheVcAAAAAAAAAADwgSI6AAAAAAAAAAA+sJwLgrLum7OuUWzpFsnLYAobAADBQOYCABAaZC4AgCI6ek1ZVa3MWFIsKzaXt9qBXHcmH5SVHtZrAwAglpC5AACEBpkLAFAs54Je65lv27BQyzeXy21Lis1xAADQc2QuAAChQeYCANwooqNX6NS2tg0L7waGHgcAAD1H5gIAEBpkLgDAjSI6eoWuDefP/k6OAwCAriFzAQAIDTIXAOBGER29wpZm8Xu8TyfHAQBA15C5AACEBpkLAHCjiI5ekZeZYjZX6Yi+r8cBAEDPkbkAAIQGmQsAcKOIjl5ht6aY3cnbNjD09dzJI81xAADQc2QuAAChQeYCANwSXC6Xy/MK4nQ6xW63i8PhEJvNFu7LiTq6O7lurqJrw+nUNu2Zp2EBAOgImdszZC4AoKvI3J4hcwEAyeG+AMQWbUjQmAAAIPjIXAAAQoPMBQCwnAsAAAAAAAAAAD5QRAcAAAAAAAAAwAeK6AAAAAAAAAAA+EARHQAAAAAAAAAAHyiiAwAAAAAAAADgA0V0AAAAAAAAAAB8oIgOAAAAAAAAAIAPFNEBAAAAAAAAAPCBIjoAAAAAAAAAAD5QRAcAAAAAAAAAwAeK6AAAAAAAAAAA+EARHQAAAAAAAAAAHyiiAwAAAAAAAADgA0V0AAAAAAAAAAB8oIgOAAAAAAAAAIAPFNEBAAAAAAAAAPCBIjoAAAAAAAAAAD5QRAcAAAAAAAAAwAeK6AAAAAAAAAAA+EARHQAAAAAAAAAAHyiiAwAAAAAAAADgA0V0AAAAAAAAAAB8oIgOAAAAAAAAAIAPFNEBAAAAAAAAAPCBIjoAAAAAAAAAAD5QRAcAAAAAAAAAwAeK6AAAAAAAAAAA+EARHQAAAAAAAAAAHyiiAwAAAAAAAADgA0V0AAAAAAAAAAB8oIgOAAAAAAAAAEC0F9EvvPBCKSgokLS0NBk4cKBceumlUlZW1uqc4uJiGTdunDknPz9fHnjggbBdLwAAAAAAAAAg+kVNEf3000+XF198UTZt2iRLliyRr776Sn7wgx94jjudTjn77LNl6NCh8umnn8qDDz4ov/71r+XJJ58M63UDAAAAAAAAAKJXgsvlckkUev3112XixIlSX18vFotFHnvsMbnjjjtk165dkpKSYs657bbb5NVXX5Uvv/yyy99Xi/F2u10cDofYbLYg/gQAAMQ3MhcAgNAgcwEAiJOR6N727dsnzz33nJxyyimmgK5Wr14t48eP9xTQ1YQJE8zI9crKyjBeLQAAAAAAAAAgWkVVEX3GjBmSkZEhubm5UlpaKq+99prnmI5A79+/f6vz3a/1mC86kl175b0/AABA7yNzAQAIDTIXAIAYKqLrcisJCQl+P7yXYvnVr34la9askbfffluSkpLkpz/9qfR0NZo5c+aYaW3uD92QFAAA9D4yFwCA0CBzAQCIoTXR9+7dKxUVFX7POfTQQ1st0eK2Y8cO0xD48MMPZfTo0aagrr3ruga623vvvSdnnHGGWf4lOzvbZw+9frjp99Dvy1pxAAD0LjIXAIDQIHMBAOhdyRJGffv2NR/d0dLSYn51Nwy0kK4bizY2NnrWSX/nnXfk8MMP91lAV6mpqeYDAAAEF5kLAEBokLkAAMThmugff/yxzJ8/X9auXSvbtm2TZcuWyZQpU+Swww4zxXN1ySWXmBHr06ZNky+++EJeeOEF+dOf/iQ333xzuC8fAAAAAAAAABCloqKIbrVa5ZVXXpEzzzzTjCzXQvnIkSPlgw8+8PSu6zpvulb6li1b5LjjjpNbbrlFZs2aJVdddVW4Lx8AAAAAAAAAEKXCuiZ6JNK14rQgz1pxAAAEF5kLAEBokLkAAMTBSHQAAAAAAAAAAOJuY9FY56hpkPIDDeKsaxRbukXyMlLEbk0J92UBABBzyFwAAEKH3AUAxBuK6EFSVlUrM5YUy4rN5Z73xg/Lk/snj5RBWelhvTYAAGIJmQsAQOiQuwCAeMRyLkHqlW/bqFDLN5fLbUuKzXEAANBzZC4AAKFD7gIA4hVF9CDQaW1tGxXejQs9DgAAeo7MBQAgdMhdAEC8oogeBLounD/7OzkOAAC6hswFACB0yF0AQLyiiB4EtjSL3+N9OjkOAAC6hswFACB0yF0AQLyiiB4EeZkpZmOVjuj7ehwAAPQcmQsAQOiQuwCAeEURPQjs1hSzM3nbxoW+njt5pDkOAAB6jswFACB0yF0AQLxKcLlcrnBfRCRxOp1it9vF4XCIzWbr0ffSncl1YxVdF06ntWmvPI0KAAAOInMBAIi+zFXkLgAg3iSH+wJimTYiaEgAABB8ZC4AAKFD7gIA4g3LuQAAAAAAAAAA4ANFdAAAAAAAAAAAfKCIDgAAAAAAAACADxTRAQAAAAAAAADwgSI6AAAAAAAAAAA+UEQHAAAAAAAAAMAHiugAAAAAAAAAAPhAER0AAAAAAAAAAB8oogMAAAAAAAAA4ANFdAAAAAAAAAAAfKCIDgAAAAAAAACADxTRAQAAAAAAAADwIdnXgXjlcrnMr06nM9yXAgBAROrTp48kJCT0+PuQuQAA+EfmAgAQGZlLEb2N/fv3m1/z8/PDfSkAAEQkh8MhNputx9+HzAUAwD8yFwCAyMjcBJe7SxpGS0uLlJWV9VqPf3fo6ABt3Gzfvr1XGkyxiHvkH/fHP+6Pf9wf/7g/vTcqjsyNDtwj/7g//nF//OP+dC7e71EsZa6K9z/PznB//OP++Mf98Y/707l4v0d9GIkemMTERBkyZIhEAv0LG49/aQPBPfKP++Mf98c/7o9/3J+eI3OjC/fIP+6Pf9wf/7g/neMexU7mKv48/eP++Mf98Y/74x/3p3Pco46xsSgAAAAAAAAAAD5QRAcAAAAAAAAAwAeK6BEoNTVV7r77bvMrOsY98o/74x/3xz/uj3/cn9jCn2fnuEf+cX/84/74x/3pHPcotvDn6R/3xz/uj3/cH/+4P53jHvnHxqIAAAAAAAAAAPjASHQAAAAAAAAAAHygiA4AAAAAAAAAgA8U0QEAAAAAAAAA8IEieoS58MILpaCgQNLS0mTgwIFy6aWXSllZWatziouLZdy4ceac/Px8eeCBByRebN26VaZNmyaFhYWSnp4uhx12mNn0oKGhodV58XyPZs+eLaeccopYrVbJysrq8JzS0lI577zzzDn9+vWTX/3qV9LU1CTx4pFHHpFDDjnE/P046aST5F//+pfEo+XLl8sFF1wggwYNkoSEBHn11VdbHdctM2bNmmX+LdL/38466yzZvHmzxIs5c+bICSecIH369DH/n0ycOFE2bdrU6py6ujq57rrrJDc3VzIzM2Xy5Mmye/fusF0zAkPm+kfmdo7M7RyZexCZ6x+ZGx/IXd/I3M6RuZ0jcw8ic/0jc7uPInqEOf300+XFF180f4GXLFkiX331lfzgBz/wHHc6nXL22WfL0KFD5dNPP5UHH3xQfv3rX8uTTz4p8eDLL7+UlpYWeeKJJ+SLL76QP/zhD/L444/L7bff7jkn3u+RNrR++MMfyjXXXNPh8ebmZtOw0PM+/PBDWbhwoTz77LMmROLBCy+8IDfffLNplH722WdyzDHHyIQJE2TPnj0Sb6qrq83Pr42tjmij/OGHHzb/j3388ceSkZFh7pUGajz44IMPTMPho48+knfeeUcaGxvNvy1639x+8YtfyN/+9jd56aWXzPn6IHjxxReH9brRdWSuf2Ru58hc/8jcb5G5/pG58YHc9Y3M7RyZ6x+Z+y0y1z8ytwdciGivvfaaKyEhwdXQ0GBeP/roo67s7GxXfX2955wZM2a4Dj/8cFe8euCBB1yFhYWe19yjgxYsWOCy2+3t3n/zzTddiYmJrl27dnnee+yxx1w2m63VPYtVJ554ouu6667zvG5ubnYNGjTINWfOHFc80zhYunSp53VLS4trwIABrgcffNDzXlVVlSs1NdW1ePFiVzzas2ePuU8ffPCB535YLBbXSy+95Dln48aN5pzVq1eH8UrRXWRu58jcjpG5HSNzO0bmdo7MjQ/krn9kbsfI3I6RuR0jcztH5nYdI9Ej2L59++S5554zU5YsFot5b/Xq1TJ+/HhJSUnxnKc9ZtqbX1lZKfHI4XBITk6O5zX3yD+9PyNGjJD+/fu3uj86skFHPcQyHZWgozZ0upZbYmKiea33Bd/asmWL7Nq1q9W9stvtZlpgvN4r/bdGuf+90b9L2mvvfY+OOOIIM005Xu9RNCNzu4bMDQyZS+Z2BZnbHpkb+8jdzpG5gSFzydyuIHPbI3O7jiJ6BJoxY4aZTqJrD+maXq+99prnmP7P7h0Kyv1aj8WbkpISmTdvnlx99dWe97hH/sXz/SkvLzfT/Dr6+WP9Zw+U+35wrw7S6bU33XSTjBkzRoYPH27e0/ugDzFt12SM13sUrcjcriNzAxfP94fM7ToytzUyN7aRu11D5gYunu8Pmdt1ZG5rZG5gKKKHwG233WY2M/D3oWuguenmF2vWrJG3335bkpKS5Kc//anZ+CCWBXqP1Ndffy3nnHOOWRftyiuvlFjWnfsDoPfomnHr16+X559/PtyXgk6QuZ0jc/0jc4HwInOjC7nrH5nrH5kLhBeZG5jkAM9HN9xyyy1y+eWX+z3n0EMP9Xyel5dnPr7zne/IkUceaXbd1gX/R48eLQMGDGi3I677tR6Ll3ukmxroxjQ6/a/tRiqxeI8CvT/+6D1ou0t3tN+frtL/r7Sx3tHfj1j/2QPlvh96b3TXcjd9feyxx0o8mT59urzxxhtml/chQ4a0ukc6dbKqqqpVLz1/n8KLzO0cmesfmds7yNyuI3O/ReZGH3LXPzLXPzK3d5C5XUfmfovMDRxF9BDo27ev+eju1ApVX19vftXGxR133GHWJ3KvHae76R5++OGSnZ0t8XCPtGdeGxbHHXecLFiwwKz15S0W71FP/g61pfdn9uzZZpfufv36ee6PzWaTo446SmKZTknSvzfvvvuuTJw40fP/mL7WAMG3CgsLTUDqvXE3JnQ9Qd29/JprrpF4oKOirr/+elm6dKm8//775p54079L+m+M3qPJkyeb93RNSp2arP+fITzI3M6Ruf6Rub2DzO06MpfMjWbkrn9krn9kbu8gc7uOzCVzeySATUgRZB999JFr3rx5rjVr1ri2bt3qevfdd12nnHKK67DDDnPV1dV5dsnt37+/69JLL3WtX7/e9fzzz7usVqvriSeecMWDHTt2uIqKilxnnnmm+Xznzp2eD7d4v0fbtm0zf4fuueceV2ZmpvlcP/bv32+ONzU1uYYPH+46++yzXWvXrnW99dZbrr59+7pmzpzpigf690F33n722WddGzZscF111VWurKysVru4xwv9O+H++6Fx8NBDD5nP9e+Quv/++829ee2111zFxcWuiy66yFVYWOiqra11xYNrrrnGZbfbXe+//36rf2tqamo85/z85z93FRQUuJYtW+b65JNPXKNHjzYfiHxkbufI3M6Ruf6Rud8ic/0jc2Mfuesfmds5Mtc/MvdbZK5/ZG73UUSPIPo/7+mnn+7Kyckx//gdcsgh5i+uhqi3zz//3DV27FhzzuDBg80/APFiwYIF5h/Bjj68xfM9uuyyyzq8P++9957nHG24nnvuua709HRXXl6e65ZbbnE1Nja64oU24DUQUlJSXCeeeKJp1Mcj/TvR0d8V/TukWlpaXHfddZdprOv/S9qo37Rpkyte+Pq3Rv8dctOG1rXXXuvKzs42DzGTJk1q9bCDyEXmdo7M7RyZ2zky9yAy1z8yN/aRu/6RuZ0jcztH5h5E5vpH5nZfgv6nZ2PZAQAAAAAAAACITa0X2QIAAAAAAAAAAB4U0QEAAAAAAAAA8IEiOgAAAAAAAAAAPlBEBwAAAAAAAADAB4roAAAAAAAAAAD4QBEdAAAAAAAAAAAfKKIDAAAAAAAAAOADRXQAAAAAAAAAAHygiA4A3zjttNPkpptuCvdlAAAQ88hcAABCg8wFegdFdACtXH755ZKQkGA+UlJSpKioSH7zm99IU1OTRCP9OV599dUunfvKK6/IvffeG/RrAgBAkblkLgAgNMhcMhfoqeQefwcAMeecc86RBQsWSH19vbz55pty3XXXicVikZkzZwb0fZqbm024JyZGdn9dQ0ODaUjl5OSE+1IAAHGGzAUAIDTIXAA9Edn/xwMIi9TUVBkwYIAMHTpUrrnmGjnrrLPk9ddfl4ceekhGjBghGRkZkp+fL9dee60cOHDA83XPPvusZGVlmXOPOuoo831KS0vl3//+t3zve9+TvLw8sdvtcuqpp8pnn33W6vfURsgTTzwh559/vlitVjnyyCNl9erVUlJSYqaf6e95yimnyFdffdXq61577TX57ne/K2lpaXLooYfKPffc4xlNcMghh5hfJ02aZL6/+/Wvf/1rOfbYY+Wpp56SwsJC87UdTXPTxtWMGTPMz6o/i45WePrpp4N45wEA8YbMPYjMBQAEG5l7EJkLdA9FdACdSk9PN73Y2tP+8MMPyxdffCELFy6UZcuWya233trq3JqaGpk7d64Jbj2vX79+sn//frnssstk5cqV8tFHH8mwYcPk+9//vnnfm04x++lPfypr166VI444Qi655BK5+uqrzciATz75RFwul0yfPt1z/ooVK8z5N954o2zYsME0TrSBM3v2bHNcGzVKRxvs3LnT81ppo2XJkiVmapv+fh3R77148WLzM2/cuNF8/8zMzF69twAAeCNzyVwAQGiQuWQuEBAXAHi57LLLXBdddJH5vKWlxfXOO++4UlNTXb/85S/bnfvSSy+5cnNzPa8XLFjg0n9W1q5d6/f3aG5udvXp08f1t7/9zfOeft2dd97peb169Wrz3tNPP+15b/Hixa60tDTP6zPPPNP129/+ttX3/stf/uIaOHBgq++7dOnSVufcfffdLovF4tqzZ0+r90899VTXjTfeaD7ftGmT+Vr9+QEACAYyl8wFAIQGmUvmAj3FmugA2nnjjTdMT3RjY6O0tLSYnnKdGvbPf/5T5syZI19++aU4nU4znayurs70yuvUNKVrro0cObLV99u9e7fceeed8v7778uePXvMGnL6NToFzpv31/Xv39/8qtPqvN/T309/b5vNJp9//rmsWrXK0yOv9Hu3vaaO6BS+vn37+jyuvfZJSUlmSh4AAMFC5pK5AIDQIHPJXKAnKKIDaOf000+Xxx57zDQUBg0aJMnJybJ161azjpuuHadhrpuT6LS1adOmmSlw7iDXKXG6Lps3neJWUVEhf/rTn0yo67pro0ePNl/nTTd1cXN/j47e0waP0nXqdG24iy++uN3P4F7/zRdde84f/TkAAAg2MpfMBQCEBplL5gI9QREdQIfBq5uLePv0009NqP/+97/37EL+4osvdun7aS/6o48+ataHU9u3b5fy8vIeX6dutLJp06Z21+pNGyfaax8oHRmgP+8HH3xgNpwBACAYyFwyFwAQGmQumQv0BEV0AF2iAa7T3ubNmycXXHCBaTA8/vjjXfpa3WDlL3/5ixx//PFmitqvfvWrXukBnzVrlhk1UFBQID/4wQ9Mo0envq1fv17uu+8+c47uVP7uu+/KmDFjzMiA7OzsLn1v/TodWTB16lSz4coxxxwj27ZtM9P0fvSjH/X42gEA8IXMJXMBAKFB5pK5QFcd7GYDgE5ouD700ENmR/Lhw4fLc889Z9aN64qnn35aKisrTY/6pZdeKjfccIPZzbynJkyYYNa1e/vtt+WEE06Qk08+Wf7whz+YqXRuOqLgnXfekfz8fBk1alRA31+n+mmj5dprrzW7qF955ZVSXV3d4+sGAMAfMpfMBQCEBplL5gJdlaC7i3b5bAAAAAAAAAAA4ggj0QEAAAAAAAAA8IEiOgAAAAAAAAAAPlBEBwAAAAAAAADAB4roAAAAAAAAAAD4QBEdAAAAAAAAAAAfKKIDAAAAAAAAAOADRXQAAAAAAAAAAHygiA4AAAAAAAAAgA8U0QEAAAAAAAAA8IEiOgAAAAAAAAAAPlBEBwAAAAAAAADAB4roAAAAAAAAAABIx/4/wajSM90MgRsAAAAASUVORK5CYII=", + "text/plain": [ + "
" + ] + }, + "metadata": {}, + "output_type": "display_data" + } + ], + "source": [ + "sns.relplot(comb_stats, x=\"Parametric\", y=\"BLB\", col=\"quantity\", kind=\"scatter\")\n", + "plt.show()" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "id": "4ce6bdd3", + "metadata": {}, + "outputs": [ + { + "data": { + "image/png": "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", + "text/plain": [ + "
" + ] + }, + "metadata": {}, + "output_type": "display_data" + } + ], + "source": [ + "sns.relplot(comb_stats, x=\"Bootstrap\", y=\"BLB\", col=\"quantity\", kind=\"scatter\")\n", + "plt.show()" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "id": "d22ac527", + "metadata": {}, + "outputs": [ + { + "data": { + "text/plain": [ + "" + ] + }, + "execution_count": 41, + "metadata": {}, + "output_type": "execute_result" + }, + { + "data": { + "image/png": "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", + "text/plain": [ + "
" + ] + }, + "metadata": {}, + "output_type": "display_data" + } + ], + "source": [ + "sns.displot(comb_stats, x=\"Error\", col=\"quantity\")" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "id": "9cc95f1d", + "metadata": {}, + "outputs": [ + { + "data": { + "text/plain": [ + "" + ] + }, + "execution_count": 42, + "metadata": {}, + "output_type": "execute_result" + }, + { + "data": { + "image/png": "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", + "text/plain": [ + "
" + ] + }, + "metadata": {}, + "output_type": "display_data" + } + ], + "source": [ + "sns.relplot(comb_stats, x=\"AbsMean\", y=\"RelError\", col=\"quantity\", kind=\"scatter\")" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "id": "82008957", + "metadata": {}, + "outputs": [ + { + "data": { + "text/plain": [ + "" + ] + }, + "execution_count": 43, + "metadata": {}, + "output_type": "execute_result" + }, + { + "data": { + "image/png": "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", + "text/plain": [ + "
" + ] + }, + "metadata": {}, + "output_type": "display_data" + } + ], + "source": [ + "sns.relplot(comb_stats, x=\"RealSTD\", y=\"RelError\", col=\"quantity\", kind=\"scatter\")" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "id": "8bcc2d96", + "metadata": {}, + "outputs": [], + "source": [] + } + ], + "metadata": { + "kernelspec": { + "display_name": ".venv", + "language": "python", + "name": "python3" + }, + "language_info": { + "codemirror_mode": { + "name": "ipython", + "version": 3 + }, + "file_extension": ".py", + "mimetype": "text/x-python", + "name": "python", + "nbconvert_exporter": "python", + "pygments_lexer": "ipython3", + "version": "3.12.10" + } + }, + "nbformat": 4, + "nbformat_minor": 5 +} diff --git a/src/lenskit/stats/_blb.py b/src/lenskit/stats/_blb.py index 4fbdd27a3..7e0ade008 100644 --- a/src/lenskit/stats/_blb.py +++ b/src/lenskit/stats/_blb.py @@ -7,7 +7,9 @@ from __future__ import annotations import warnings -from typing import Any, ClassVar, Literal, Protocol, TypeAlias, TypedDict, TypeVar +from collections.abc import Callable +from dataclasses import dataclass +from typing import Any, ClassVar, Literal, Protocol, TypeAlias, TypeVar import numpy as np import pandas as pd @@ -43,10 +45,10 @@ def blb_summary( stat: SummaryStat, *, ci_width: float = 0.95, - b_factor: float = 0.6, - tol: float = 0.05, - s_w: int = 3, - r_w: int = 20, + b_factor: float = 0.8, + rel_tol: float = 0.05, + s_window: int = 3, + r_window: int = 20, rng: RNGInput = None, ) -> dict[str, float]: """ @@ -63,18 +65,28 @@ def blb_summary( est = np.average(xs).item() rng = random_generator(rng) - bootstrapper = _BLBootstrapper(np.average, ci_width, tol, s_w, r_w, b_factor, rng) + bootstrapper = _BLBootstrapper(np.average, ci_width, rel_tol, s_window, r_window, b_factor, rng) - boot_df = bootstrapper.summarize(xs) + result = bootstrapper.run_bootstraps(xs) - return {"value": est} | boot_df.agg("mean").to_dict() + result = { + "value": est, + "mean": result.mean, + "sdist_var": result.sdist_var, + "ci_lower": result.ci_lower, + "ci_upper": result.ci_upper, + } + return result -class _BootResult(TypedDict): + +@dataclass +class _BootResult: mean: float - ci_min: float - ci_max: float - count: int + sdist_var: float + ci_lower: float + ci_upper: float + samples: pd.DataFrame class _BLBootstrapper: @@ -84,6 +96,8 @@ class _BLBootstrapper: statistic: WeightedStatistic ci_width: float + _ci_qmin: float + _ci_qmax: float tolerance: float s_window: int @@ -110,20 +124,34 @@ def __init__( self.rng = rng self.ss_stats = {} - def summarize(self, xs: NDArray[F]): - results = [] - means = StatAccum(self.tolerance, self.s_window) + alpha = 1 - ci_width + self._ci_qmin = 0.5 * alpha + self._ci_qmax = 1 - 0.5 * alpha + + def run_bootstraps(self, xs: NDArray[F]) -> _BootResult: + ss_frames = {} + + means = StatAccum(np.mean) + sv = StatAccum(np.mean) + lbs = StatAccum(np.mean) + ubs = StatAccum(np.mean) - count = 0 - for ss in self.blb_subsets(xs): - count += 1 + for i, ss in enumerate(self.blb_subsets(xs)): res = self.measure_subset(xs, ss) - results.append(res) - means.record(res["mean"]) - if means.converged(): + ss_frames[i] = res.samples + means.record(res.mean) + lbs.record(res.ci_lower) + ubs.record(res.ci_upper) + if _check_convergence(means, sv, lbs, ubs, tol=self.tolerance, w=self.s_window): break - return pd.DataFrame.from_records(results) + return _BootResult( + means.statistic, + sv.statistic, + lbs.statistic, + ubs.statistic, + pd.concat(ss_frames, names=["subset"]), + ) def blb_subsets(self, xs: NDArray[F]): b = int(len(xs) ** self.b_factor) @@ -136,18 +164,27 @@ def measure_subset(self, xs: NDArray[F], ss: NDArray[np.int64]) -> _BootResult: n = len(xs) xss = xs[ss] - acc = StatAccum(self.tolerance, self.r_window) + values = [] + means = StatAccum(np.mean) + svs = StatAccum(np.var) + lbs = StatAccum(lambda a: np.quantile(a, self._ci_qmin)) + ubs = StatAccum(lambda a: np.quantile(a, self._ci_qmax)) - count = 0 for weights in self.miniboot_weights(n, b): - count += 1 + assert weights.shape == (b,) + assert np.sum(weights) == n stat = self.statistic(xss, weights=weights) - acc.record(stat) - if acc.converged(): + values.append(stat) + means.record(stat) + lbs.record(stat) + ubs.record(stat) + + if _check_convergence(means, svs, lbs, ubs, tol=self.tolerance, w=self.r_window): break - [lo, hi] = np.quantile(acc.values, [0.025, 0.975]) - return {"mean": np.mean(acc.values).item(), "ci_min": lo, "ci_max": hi, "count": count} + df = pd.DataFrame({"statistic": values}) + df.index.name = "iter" + return _BootResult(means.statistic, svs.statistic, lbs.statistic, ubs.statistic, df) def miniboot_weights(self, n: int, b: int): flat = np.full(b, 1.0 / b) @@ -156,28 +193,49 @@ def miniboot_weights(self, n: int, b: int): yield self.rng.multinomial(n, flat) +def _check_convergence(*arrays: StatAccum, tol: float, w: int) -> bool: + gaps = np.zeros(w) + for arr in arrays: + if len(arr) < w + 1: + return False + stats = arr.stat_history + cur = stats[-1] + gaps += np.abs(stats[-(w + 1) : -1] - cur) / np.abs(cur) + + gaps /= len(arrays) + return np.all(gaps < tol).item() + + class StatAccum: INIT_SIZE: ClassVar[int] = 100 - ABS_TOL: ClassVar[float] = 1.0e-12 - tolerance: float - window: int + _stat_func: Callable[[NDArray[np.floating[Any]]], np.floating[Any]] _len: int = 0 _values: NDArray[np.float64] - _cum_means: NDArray[np.float64] + _cum_stat: NDArray[np.float64] - def __init__(self, tol: float, w: int): - self.tolerance = tol - self.window = w + def __init__(self, stat: Callable[[NDArray[np.floating[Any]]], np.floating[Any]]): + self._stat_func = stat self._values = np.zeros(self.INIT_SIZE) - self._cum_means = np.zeros(self.INIT_SIZE) + self._cum_stat = np.zeros(self.INIT_SIZE) @property def values(self) -> NDArray[np.float64]: return self._values[: self._len] + @property + def statistic(self) -> float: + if self._len: + return self._cum_stat[-1] + else: + return np.nan + + @property + def stat_history(self) -> NDArray[np.float64]: + return self._cum_stat[: self._len] + def record(self, x: float | np.floating[Any]) -> None: "Record a new value in the accumulator." self._expand_if_needed() @@ -186,35 +244,13 @@ def record(self, x: float | np.floating[Any]) -> None: # record and update the cumulative mean self._values[i] = x - self._cum_means[i] = np.mean(self.values) - - def mean(self) -> float | None: - "Get the mean of the accumulated values." - if self._len > 0: - return self._cum_means[self._len - 1] - else: - return None - - def converged(self) -> bool: - """ - Check for convergence. - """ - if self._len < self.window: - return False - - i_cur = self._len - 1 - i_start = self._len - self.window - current = self._cum_means[i_cur] - - # lower-bound tolerance for very small values - atol = max(current * self.tolerance, self.ABS_TOL) - - window = self._cum_means[i_start : self._len] - gaps = np.abs(window - current) - return np.all(gaps <= atol).item() + self._cum_stat[i] = self._stat_func(self.values) def _expand_if_needed(self): cap = len(self._values) if cap == self._len: self._values.resize(cap * 2) - self._cum_means.resize(cap * 2) + self._cum_stat.resize(cap * 2) + + def __len__(self): + return self._len diff --git a/src/lenskit/stats/_topn.py b/src/lenskit/stats/_topn.py index d6fa20159..9c76ed552 100644 --- a/src/lenskit/stats/_topn.py +++ b/src/lenskit/stats/_topn.py @@ -6,10 +6,13 @@ from __future__ import annotations +from typing import TYPE_CHECKING + import numpy as np from numpy.typing import ArrayLike -from lenskit.data.types import NPVector +if TYPE_CHECKING: + from lenskit.data.types import NPVector def argtopn(xs: ArrayLike, n: int) -> NPVector[np.int64]: From bfde242e03c9ce33a1fb7a3d801057ce9ec93679 Mon Sep 17 00:00:00 2001 From: Michael Ekstrand Date: Tue, 3 Jun 2025 10:06:04 -0400 Subject: [PATCH 11/59] document output --- src/lenskit/stats/_blb.py | 49 ++++++++++++++++++++++++++++++++------- 1 file changed, 40 insertions(+), 9 deletions(-) diff --git a/src/lenskit/stats/_blb.py b/src/lenskit/stats/_blb.py index 7e0ade008..78bebd5ff 100644 --- a/src/lenskit/stats/_blb.py +++ b/src/lenskit/stats/_blb.py @@ -51,8 +51,39 @@ def blb_summary( r_window: int = 20, rng: RNGInput = None, ) -> dict[str, float]: - """ - Summarize one or more statistics using the Bag of Little Bootstraps :cite:p:`blb`. + r""" + Summarize one or more statistics using the Bag of Little Bootstraps + :cite:p:`blb`. + + This is a direct, sequential implementation of Bag of Little Bootstraps as + described in the original paper :cite:p:`blb`, with automatic + convergence-based termination. + + Args: + xs: + The array of values to summarize. + stat: + The statistic to compute. The Bag of Little Bootstraps requires + statistics to support weighted computation (this is what allows it + to speed up the bootstrap procedure). + ci_width: + The width of the confidence interval to estimat.e + b_factor: + The shrinking factor :math:`\gamma` to use to derive subsample + sizes. Each subsample has size :math:`N^{\gamma}`. + rel_tol: + The relative tolerance for detecting convergence. + s_window: + The window length for detecting convergence in the outer subset loop + (and minimum number of subsets). + r_window: + The window length for detecting convergence in the inner replication + loop (and minimum number of replicates per subset). + rng: + The RNG or seed for randomization. + + Returns: + A dictionary of statistical results of the statistic. """ if stat != "mean": raise ValueError(f"unsupported statistic {stat}") @@ -70,9 +101,9 @@ def blb_summary( result = bootstrapper.run_bootstraps(xs) result = { - "value": est, - "mean": result.mean, - "sdist_var": result.sdist_var, + "estimate": est, + "rep_mean": result.mean, + "rep_var": result.rep_var, "ci_lower": result.ci_lower, "ci_upper": result.ci_upper, } @@ -83,7 +114,7 @@ def blb_summary( @dataclass class _BootResult: mean: float - sdist_var: float + rep_var: float ci_lower: float ci_upper: float samples: pd.DataFrame @@ -132,7 +163,7 @@ def run_bootstraps(self, xs: NDArray[F]) -> _BootResult: ss_frames = {} means = StatAccum(np.mean) - sv = StatAccum(np.mean) + vars = StatAccum(np.mean) lbs = StatAccum(np.mean) ubs = StatAccum(np.mean) @@ -142,12 +173,12 @@ def run_bootstraps(self, xs: NDArray[F]) -> _BootResult: means.record(res.mean) lbs.record(res.ci_lower) ubs.record(res.ci_upper) - if _check_convergence(means, sv, lbs, ubs, tol=self.tolerance, w=self.s_window): + if _check_convergence(means, vars, lbs, ubs, tol=self.tolerance, w=self.s_window): break return _BootResult( means.statistic, - sv.statistic, + vars.statistic, lbs.statistic, ubs.statistic, pd.concat(ss_frames, names=["subset"]), From b91bb81cb5bf0949ff99610abebdacd9b1bec6a9 Mon Sep 17 00:00:00 2001 From: Michael Ekstrand Date: Tue, 3 Jun 2025 10:14:53 -0400 Subject: [PATCH 12/59] try to fix BLB tests --- tests/stats/test_blb.py | 40 ++++++++++++++++++++++------------------ 1 file changed, 22 insertions(+), 18 deletions(-) diff --git a/tests/stats/test_blb.py b/tests/stats/test_blb.py index cbbc82de7..7575e3b0c 100644 --- a/tests/stats/test_blb.py +++ b/tests/stats/test_blb.py @@ -36,30 +36,34 @@ def test_blb_single_array(rng: np.random.Generator): assert summary["ci_max"] == approx(mean + 1.96 * ste, rel=0.05) -@given( - st.integers(1000, 1_000_000), - nph.floating_dtypes(endianness="="), - st.integers(0), -) +@mark.parametrize("size", [1000, 10000, 1000000]) @mark.filterwarnings(r"error:.*ignoring \d+ nonfinite values") -def test_blb_array_normal(n, dtype, seed): +def test_blb_array_normal(rng: np.random.Generator, size: int): "Test BLB with arrays of normals." - rng = random_generator(seed) - xs = rng.normal(1.0, 1.0, n).astype(dtype) - mean = np.mean(xs) - n = len(xs) - std = np.std(xs) - ste = std / sqrt(n) - summary = blb_summary(xs, "mean", rng=rng) - assert isinstance(summary, dict) - assert summary["value"] == approx(mean) - assert summary["mean"] == approx(mean, rel=0.075) + TRUE_MEAN = 1.0 + results = [] + + # Test: for 1000 runs, do approx. 95% of confidence intervals contain the + # true mean? + + for i in range(1000): + xs = rng.normal(1.0, 1.0, size) + mean = np.mean(xs) + + summary = blb_summary(xs, "mean", rng=rng) + assert isinstance(summary, dict) + assert summary["estimate"] == approx(mean) + assert summary["rep_mean"] == approx(mean, rel=0.075) + + results.append(summary) - assert summary["ci_min"] == approx(mean - 1.96 * ste, rel=0.075) - assert summary["ci_max"] == approx(mean + 1.96 * ste, rel=0.075) + n_good = len([r for r in results if r["ci_lower"] <= TRUE_MEAN <= r["ci_upper"]]) + # leave a little wiggle room + assert n_good >= 925 +@mark.skip("need to find better parameters") @given( nph.arrays(shape=st.integers(10000, 1_000_000), dtype=nph.floating_dtypes(endianness="=")), st.integers(0), From 6206f0b5cbcd1bc655ace3cab4af87d56bc41212 Mon Sep 17 00:00:00 2001 From: Michael Ekstrand Date: Tue, 3 Jun 2025 10:34:28 -0400 Subject: [PATCH 13/59] mostly-working tests --- src/lenskit/stats/_blb.py | 55 +++++++++++++++++++++++++-------------- tests/stats/test_blb.py | 9 +++---- 2 files changed, 40 insertions(+), 24 deletions(-) diff --git a/src/lenskit/stats/_blb.py b/src/lenskit/stats/_blb.py index 78bebd5ff..f5e94db6d 100644 --- a/src/lenskit/stats/_blb.py +++ b/src/lenskit/stats/_blb.py @@ -14,14 +14,18 @@ import numpy as np import pandas as pd from numpy.typing import NDArray +from structlog.stdlib import BoundLogger from lenskit.diagnostics import DataWarning +from lenskit.logging import get_logger, trace from lenskit.random import RNGInput, random_generator F = TypeVar("F", bound=np.floating, covariant=True) SummaryStat: TypeAlias = Literal["mean"] +_log = get_logger(__name__) + # dummy assignment to typecheck that we have correctly typed weighted average __dummy_avg: WeightedStatistic = np.average @@ -102,7 +106,7 @@ def blb_summary( result = { "estimate": est, - "rep_mean": result.mean, + "rep_mean": result.rep_mean, "rep_var": result.rep_var, "ci_lower": result.ci_lower, "ci_upper": result.ci_upper, @@ -113,7 +117,7 @@ def blb_summary( @dataclass class _BootResult: - mean: float + rep_mean: float rep_var: float ci_lower: float ci_upper: float @@ -125,6 +129,7 @@ class _BLBootstrapper: Implementation of BLB computation. """ + _log: BoundLogger statistic: WeightedStatistic ci_width: float _ci_qmin: float @@ -158,8 +163,11 @@ def __init__( alpha = 1 - ci_width self._ci_qmin = 0.5 * alpha self._ci_qmax = 1 - 0.5 * alpha + self._log = _log.bind(stat=stat.__name__) def run_bootstraps(self, xs: NDArray[F]) -> _BootResult: + self._log = self._log.bind(n=len(xs)) + self._log.debug("starting bootstrap") ss_frames = {} means = StatAccum(np.mean) @@ -168,12 +176,15 @@ def run_bootstraps(self, xs: NDArray[F]) -> _BootResult: ubs = StatAccum(np.mean) for i, ss in enumerate(self.blb_subsets(xs)): + self._log = self._log.bind(subset=i) + trace(self._log, "starting subset") res = self.measure_subset(xs, ss) ss_frames[i] = res.samples - means.record(res.mean) + means.record(res.rep_mean) + vars.record(res.rep_var) lbs.record(res.ci_lower) ubs.record(res.ci_upper) - if _check_convergence(means, vars, lbs, ubs, tol=self.tolerance, w=self.s_window): + if self._check_convergence(means, vars, lbs, ubs, tol=self.tolerance, w=self.s_window): break return _BootResult( @@ -186,6 +197,7 @@ def run_bootstraps(self, xs: NDArray[F]) -> _BootResult: def blb_subsets(self, xs: NDArray[F]): b = int(len(xs) ** self.b_factor) + self._log = self._log.bind(b=b) while True: yield self.rng.choice(len(xs), b, replace=False) @@ -197,25 +209,29 @@ def measure_subset(self, xs: NDArray[F], ss: NDArray[np.int64]) -> _BootResult: values = [] means = StatAccum(np.mean) - svs = StatAccum(np.var) + vars = StatAccum(np.var) lbs = StatAccum(lambda a: np.quantile(a, self._ci_qmin)) ubs = StatAccum(lambda a: np.quantile(a, self._ci_qmax)) - for weights in self.miniboot_weights(n, b): + for i, weights in enumerate(self.miniboot_weights(n, b)): + self._log = self._log.bind(rep=i) + trace(self._log, "starting replicate") assert weights.shape == (b,) assert np.sum(weights) == n stat = self.statistic(xss, weights=weights) values.append(stat) means.record(stat) + vars.record(stat) lbs.record(stat) ubs.record(stat) - if _check_convergence(means, svs, lbs, ubs, tol=self.tolerance, w=self.r_window): + if self._check_convergence(means, vars, lbs, ubs, tol=self.tolerance, w=self.r_window): break df = pd.DataFrame({"statistic": values}) df.index.name = "iter" - return _BootResult(means.statistic, svs.statistic, lbs.statistic, ubs.statistic, df) + self._log = self._log.unbind("rep") + return _BootResult(means.statistic, vars.statistic, lbs.statistic, ubs.statistic, df) def miniboot_weights(self, n: int, b: int): flat = np.full(b, 1.0 / b) @@ -223,18 +239,19 @@ def miniboot_weights(self, n: int, b: int): while True: yield self.rng.multinomial(n, flat) + def _check_convergence(self, *arrays: StatAccum, tol: float, w: int) -> bool: + gaps = np.zeros(w) + for arr in arrays: + if len(arr) < w + 1: + return False -def _check_convergence(*arrays: StatAccum, tol: float, w: int) -> bool: - gaps = np.zeros(w) - for arr in arrays: - if len(arr) < w + 1: - return False - stats = arr.stat_history - cur = stats[-1] - gaps += np.abs(stats[-(w + 1) : -1] - cur) / np.abs(cur) + stats = arr.stat_history + cur = arr.statistic + gaps += np.abs(stats[-(w + 1) : -1] - cur) / np.abs(cur) - gaps /= len(arrays) - return np.all(gaps < tol).item() + gaps /= len(arrays) + trace(self._log, "max gap: %.3f", np.max(gaps)) + return np.all(gaps < tol).item() class StatAccum: @@ -259,7 +276,7 @@ def values(self) -> NDArray[np.float64]: @property def statistic(self) -> float: if self._len: - return self._cum_stat[-1] + return self._cum_stat[self._len - 1] else: return np.nan diff --git a/tests/stats/test_blb.py b/tests/stats/test_blb.py index 7575e3b0c..4434a3a41 100644 --- a/tests/stats/test_blb.py +++ b/tests/stats/test_blb.py @@ -29,11 +29,11 @@ def test_blb_single_array(rng: np.random.Generator): summary = blb_summary(xs, "mean", rng=rng) print(summary) assert isinstance(summary, dict) - assert summary["value"] == approx(mean) - assert summary["mean"] == approx(mean, rel=0.05) + assert summary["estimate"] == approx(mean) + assert summary["rep_mean"] == approx(mean, rel=0.05) - assert summary["ci_min"] == approx(mean - 1.96 * ste, rel=0.05) - assert summary["ci_max"] == approx(mean + 1.96 * ste, rel=0.05) + assert summary["ci_lower"] == approx(mean - 1.96 * ste, rel=0.05) + assert summary["ci_upper"] == approx(mean + 1.96 * ste, rel=0.05) @mark.parametrize("size", [1000, 10000, 1000000]) @@ -54,7 +54,6 @@ def test_blb_array_normal(rng: np.random.Generator, size: int): summary = blb_summary(xs, "mean", rng=rng) assert isinstance(summary, dict) assert summary["estimate"] == approx(mean) - assert summary["rep_mean"] == approx(mean, rel=0.075) results.append(summary) From f46344bf845998a6657bb3032c35713e0c0618f6 Mon Sep 17 00:00:00 2001 From: Michael Ekstrand Date: Tue, 3 Jun 2025 10:43:37 -0400 Subject: [PATCH 14/59] BLB window test is slow --- tests/stats/test_blb.py | 1 + 1 file changed, 1 insertion(+) diff --git a/tests/stats/test_blb.py b/tests/stats/test_blb.py index 4434a3a41..34be91cd1 100644 --- a/tests/stats/test_blb.py +++ b/tests/stats/test_blb.py @@ -36,6 +36,7 @@ def test_blb_single_array(rng: np.random.Generator): assert summary["ci_upper"] == approx(mean + 1.96 * ste, rel=0.05) +@mark.slow @mark.parametrize("size", [1000, 10000, 1000000]) @mark.filterwarnings(r"error:.*ignoring \d+ nonfinite values") def test_blb_array_normal(rng: np.random.Generator, size: int): From ac4922f3d0e6ff1ccff178cc4310828552e0c6cf Mon Sep 17 00:00:00 2001 From: Michael Ekstrand Date: Tue, 3 Jun 2025 10:48:53 -0400 Subject: [PATCH 15/59] simplify tests? --- tests/stats/test_blb.py | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/tests/stats/test_blb.py b/tests/stats/test_blb.py index 34be91cd1..47f9df7aa 100644 --- a/tests/stats/test_blb.py +++ b/tests/stats/test_blb.py @@ -48,8 +48,8 @@ def test_blb_array_normal(rng: np.random.Generator, size: int): # Test: for 1000 runs, do approx. 95% of confidence intervals contain the # true mean? - for i in range(1000): - xs = rng.normal(1.0, 1.0, size) + for i in range(100): + xs = rng.normal(TRUE_MEAN, 1.0, size) mean = np.mean(xs) summary = blb_summary(xs, "mean", rng=rng) @@ -60,7 +60,7 @@ def test_blb_array_normal(rng: np.random.Generator, size: int): n_good = len([r for r in results if r["ci_lower"] <= TRUE_MEAN <= r["ci_upper"]]) # leave a little wiggle room - assert n_good >= 925 + assert 90 <= n_good <= 99 @mark.skip("need to find better parameters") From 75f88e4cb7b91f42ac785d44e54f1d27677aa606 Mon Sep 17 00:00:00 2001 From: Michael Ekstrand Date: Tue, 3 Jun 2025 10:52:27 -0400 Subject: [PATCH 16/59] split out UB and LB tests --- tests/stats/test_blb.py | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) diff --git a/tests/stats/test_blb.py b/tests/stats/test_blb.py index 47f9df7aa..037bd1d2a 100644 --- a/tests/stats/test_blb.py +++ b/tests/stats/test_blb.py @@ -58,9 +58,11 @@ def test_blb_array_normal(rng: np.random.Generator, size: int): results.append(summary) - n_good = len([r for r in results if r["ci_lower"] <= TRUE_MEAN <= r["ci_upper"]]) + n_lb_good = len([r for r in results if r["ci_lower"] <= TRUE_MEAN]) + n_ub_good = len([r for r in results if TRUE_MEAN <= r["ci_upper"]]) # leave a little wiggle room - assert 90 <= n_good <= 99 + assert 90 <= n_lb_good <= 99 + assert 90 <= n_ub_good <= 99 @mark.skip("need to find better parameters") From 9cf2feca27d841f971f4f828188d8334ae299820 Mon Sep 17 00:00:00 2001 From: Michael Ekstrand Date: Tue, 3 Jun 2025 10:57:37 -0400 Subject: [PATCH 17/59] use percentages for tests --- tests/stats/test_blb.py | 11 +++++++---- 1 file changed, 7 insertions(+), 4 deletions(-) diff --git a/tests/stats/test_blb.py b/tests/stats/test_blb.py index 037bd1d2a..3d2964e1b 100644 --- a/tests/stats/test_blb.py +++ b/tests/stats/test_blb.py @@ -37,7 +37,7 @@ def test_blb_single_array(rng: np.random.Generator): @mark.slow -@mark.parametrize("size", [1000, 10000, 1000000]) +@mark.parametrize("size", [1000, 10000, 100000]) @mark.filterwarnings(r"error:.*ignoring \d+ nonfinite values") def test_blb_array_normal(rng: np.random.Generator, size: int): "Test BLB with arrays of normals." @@ -48,7 +48,8 @@ def test_blb_array_normal(rng: np.random.Generator, size: int): # Test: for 1000 runs, do approx. 95% of confidence intervals contain the # true mean? - for i in range(100): + NTRIALS = 200 + for i in range(NTRIALS): xs = rng.normal(TRUE_MEAN, 1.0, size) mean = np.mean(xs) @@ -59,10 +60,12 @@ def test_blb_array_normal(rng: np.random.Generator, size: int): results.append(summary) n_lb_good = len([r for r in results if r["ci_lower"] <= TRUE_MEAN]) + pct_lb_good = (n_lb_good / NTRIALS) * 100 n_ub_good = len([r for r in results if TRUE_MEAN <= r["ci_upper"]]) + pct_ub_good = (n_ub_good / NTRIALS) * 100 # leave a little wiggle room - assert 90 <= n_lb_good <= 99 - assert 90 <= n_ub_good <= 99 + assert 90 <= pct_lb_good <= 99 + assert 90 <= pct_ub_good <= 99 @mark.skip("need to find better parameters") From 118c67ca53fc6fdea76a81fdaf1fc6a12fd562fd Mon Sep 17 00:00:00 2001 From: Michael Ekstrand Date: Tue, 3 Jun 2025 10:58:01 -0400 Subject: [PATCH 18/59] tighten BLB tolerance for test --- tests/stats/test_blb.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tests/stats/test_blb.py b/tests/stats/test_blb.py index 3d2964e1b..ad4cc9300 100644 --- a/tests/stats/test_blb.py +++ b/tests/stats/test_blb.py @@ -53,7 +53,7 @@ def test_blb_array_normal(rng: np.random.Generator, size: int): xs = rng.normal(TRUE_MEAN, 1.0, size) mean = np.mean(xs) - summary = blb_summary(xs, "mean", rng=rng) + summary = blb_summary(xs, "mean", rng=rng, rel_tol=0.02) assert isinstance(summary, dict) assert summary["estimate"] == approx(mean) From 413ef5e7b310cb6e538a4166e78294064a550eb1 Mon Sep 17 00:00:00 2001 From: Michael Ekstrand Date: Tue, 3 Jun 2025 12:23:19 -0400 Subject: [PATCH 19/59] use tracer in BLB --- src/lenskit/stats/_blb.py | 21 ++++++++++----------- 1 file changed, 10 insertions(+), 11 deletions(-) diff --git a/src/lenskit/stats/_blb.py b/src/lenskit/stats/_blb.py index f5e94db6d..9bdff5004 100644 --- a/src/lenskit/stats/_blb.py +++ b/src/lenskit/stats/_blb.py @@ -14,10 +14,9 @@ import numpy as np import pandas as pd from numpy.typing import NDArray -from structlog.stdlib import BoundLogger from lenskit.diagnostics import DataWarning -from lenskit.logging import get_logger, trace +from lenskit.logging import Tracer, get_logger, get_tracer from lenskit.random import RNGInput, random_generator F = TypeVar("F", bound=np.floating, covariant=True) @@ -129,7 +128,7 @@ class _BLBootstrapper: Implementation of BLB computation. """ - _log: BoundLogger + _tracer: Tracer statistic: WeightedStatistic ci_width: float _ci_qmin: float @@ -163,11 +162,11 @@ def __init__( alpha = 1 - ci_width self._ci_qmin = 0.5 * alpha self._ci_qmax = 1 - 0.5 * alpha - self._log = _log.bind(stat=stat.__name__) + self._tracer = get_tracer(_log, stat=stat.__name__) # type: ignore def run_bootstraps(self, xs: NDArray[F]) -> _BootResult: - self._log = self._log.bind(n=len(xs)) - self._log.debug("starting bootstrap") + self._tracer.add_bindings(n=len(xs)) + _log.debug("starting bootstrap", stat=self.statistic.__name__, n=len(xs)) # type: ignore ss_frames = {} means = StatAccum(np.mean) @@ -176,8 +175,8 @@ def run_bootstraps(self, xs: NDArray[F]) -> _BootResult: ubs = StatAccum(np.mean) for i, ss in enumerate(self.blb_subsets(xs)): - self._log = self._log.bind(subset=i) - trace(self._log, "starting subset") + self._tracer.add_bindings(subset=i) + self._tracer.trace(self._log, "starting subset") res = self.measure_subset(xs, ss) ss_frames[i] = res.samples means.record(res.rep_mean) @@ -214,8 +213,8 @@ def measure_subset(self, xs: NDArray[F], ss: NDArray[np.int64]) -> _BootResult: ubs = StatAccum(lambda a: np.quantile(a, self._ci_qmax)) for i, weights in enumerate(self.miniboot_weights(n, b)): - self._log = self._log.bind(rep=i) - trace(self._log, "starting replicate") + self._tracer.add_bindings(rep=i) + self._tracer.trace(self._log, "starting replicate") assert weights.shape == (b,) assert np.sum(weights) == n stat = self.statistic(xss, weights=weights) @@ -250,7 +249,7 @@ def _check_convergence(self, *arrays: StatAccum, tol: float, w: int) -> bool: gaps += np.abs(stats[-(w + 1) : -1] - cur) / np.abs(cur) gaps /= len(arrays) - trace(self._log, "max gap: %.3f", np.max(gaps)) + self._tracer.trace(self._log, "max gap: %.3f", np.max(gaps)) return np.all(gaps < tol).item() From 282ecf71f456cd2c386439e7c5c8b0b7b5d02b73 Mon Sep 17 00:00:00 2001 From: Michael Ekstrand Date: Tue, 3 Jun 2025 12:33:13 -0400 Subject: [PATCH 20/59] clean up tracing types --- src/lenskit/stats/_blb.py | 14 ++++++-------- 1 file changed, 6 insertions(+), 8 deletions(-) diff --git a/src/lenskit/stats/_blb.py b/src/lenskit/stats/_blb.py index 9bdff5004..6320da470 100644 --- a/src/lenskit/stats/_blb.py +++ b/src/lenskit/stats/_blb.py @@ -176,7 +176,7 @@ def run_bootstraps(self, xs: NDArray[F]) -> _BootResult: for i, ss in enumerate(self.blb_subsets(xs)): self._tracer.add_bindings(subset=i) - self._tracer.trace(self._log, "starting subset") + self._tracer.trace("starting subset") res = self.measure_subset(xs, ss) ss_frames[i] = res.samples means.record(res.rep_mean) @@ -196,7 +196,7 @@ def run_bootstraps(self, xs: NDArray[F]) -> _BootResult: def blb_subsets(self, xs: NDArray[F]): b = int(len(xs) ** self.b_factor) - self._log = self._log.bind(b=b) + self._tracer.add_bindings(b=b) while True: yield self.rng.choice(len(xs), b, replace=False) @@ -206,7 +206,6 @@ def measure_subset(self, xs: NDArray[F], ss: NDArray[np.int64]) -> _BootResult: n = len(xs) xss = xs[ss] - values = [] means = StatAccum(np.mean) vars = StatAccum(np.var) lbs = StatAccum(lambda a: np.quantile(a, self._ci_qmin)) @@ -214,11 +213,10 @@ def measure_subset(self, xs: NDArray[F], ss: NDArray[np.int64]) -> _BootResult: for i, weights in enumerate(self.miniboot_weights(n, b)): self._tracer.add_bindings(rep=i) - self._tracer.trace(self._log, "starting replicate") + self._tracer.trace("starting replicate") assert weights.shape == (b,) assert np.sum(weights) == n stat = self.statistic(xss, weights=weights) - values.append(stat) means.record(stat) vars.record(stat) lbs.record(stat) @@ -227,9 +225,9 @@ def measure_subset(self, xs: NDArray[F], ss: NDArray[np.int64]) -> _BootResult: if self._check_convergence(means, vars, lbs, ubs, tol=self.tolerance, w=self.r_window): break - df = pd.DataFrame({"statistic": values}) + df = pd.DataFrame({"statistic": means.values}) df.index.name = "iter" - self._log = self._log.unbind("rep") + self._tracer.remove_bindings("rep") return _BootResult(means.statistic, vars.statistic, lbs.statistic, ubs.statistic, df) def miniboot_weights(self, n: int, b: int): @@ -249,7 +247,7 @@ def _check_convergence(self, *arrays: StatAccum, tol: float, w: int) -> bool: gaps += np.abs(stats[-(w + 1) : -1] - cur) / np.abs(cur) gaps /= len(arrays) - self._tracer.trace(self._log, "max gap: %.3f", np.max(gaps)) + self._tracer.trace("max gap: %.3f", np.max(gaps)) return np.all(gaps < tol).item() From 8bd54aa586cefc04045a3bdc25a1cce8eff3d72a Mon Sep 17 00:00:00 2001 From: Michael Ekstrand Date: Tue, 3 Jun 2025 13:50:31 -0400 Subject: [PATCH 21/59] Use background thread to generate BLB subsets --- src/lenskit/stats/_blb.py | 123 +++++++++++++++++++++++++++++++------- 1 file changed, 101 insertions(+), 22 deletions(-) diff --git a/src/lenskit/stats/_blb.py b/src/lenskit/stats/_blb.py index 6320da470..8e85ae62c 100644 --- a/src/lenskit/stats/_blb.py +++ b/src/lenskit/stats/_blb.py @@ -7,16 +7,18 @@ from __future__ import annotations import warnings -from collections.abc import Callable +from collections import deque +from collections.abc import Callable, Generator from dataclasses import dataclass -from typing import Any, ClassVar, Literal, Protocol, TypeAlias, TypeVar +from threading import Condition, Lock, Thread +from typing import Any, ClassVar, Deque, Literal, Protocol, TypeAlias, TypeVar import numpy as np import pandas as pd from numpy.typing import NDArray from lenskit.diagnostics import DataWarning -from lenskit.logging import Tracer, get_logger, get_tracer +from lenskit.logging import Tracer, get_logger, get_tracer, trace from lenskit.random import RNGInput, random_generator F = TypeVar("F", bound=np.floating, covariant=True) @@ -139,6 +141,7 @@ class _BLBootstrapper: r_window: int b_factor: float rng: np.random.Generator + _rep_generator: ReplicateGenerator def __init__( self, @@ -165,7 +168,10 @@ def __init__( self._tracer = get_tracer(_log, stat=stat.__name__) # type: ignore def run_bootstraps(self, xs: NDArray[F]) -> _BootResult: - self._tracer.add_bindings(n=len(xs)) + n = len(xs) + b = int(n**self.b_factor) + + self._tracer.add_bindings(n=n, b=b) _log.debug("starting bootstrap", stat=self.statistic.__name__, n=len(xs)) # type: ignore ss_frames = {} @@ -174,17 +180,23 @@ def run_bootstraps(self, xs: NDArray[F]) -> _BootResult: lbs = StatAccum(np.mean) ubs = StatAccum(np.mean) - for i, ss in enumerate(self.blb_subsets(xs)): - self._tracer.add_bindings(subset=i) - self._tracer.trace("starting subset") - res = self.measure_subset(xs, ss) - ss_frames[i] = res.samples - means.record(res.rep_mean) - vars.record(res.rep_var) - lbs.record(res.ci_lower) - ubs.record(res.ci_upper) - if self._check_convergence(means, vars, lbs, ubs, tol=self.tolerance, w=self.s_window): - break + self._rep_generator = ReplicateGenerator(n, b, self.rng) + self._tracer.trace("let's go!") + + with self._rep_generator: + for i, ss in enumerate(self.blb_subsets(n, b)): + self._tracer.add_bindings(subset=i) + self._tracer.trace("starting subset") + res = self.measure_subset(xs, ss) + ss_frames[i] = res.samples + means.record(res.rep_mean) + vars.record(res.rep_var) + lbs.record(res.ci_lower) + ubs.record(res.ci_upper) + if self._check_convergence( + means, vars, lbs, ubs, tol=self.tolerance, w=self.s_window + ): + break return _BootResult( means.statistic, @@ -194,12 +206,9 @@ def run_bootstraps(self, xs: NDArray[F]) -> _BootResult: pd.concat(ss_frames, names=["subset"]), ) - def blb_subsets(self, xs: NDArray[F]): - b = int(len(xs) ** self.b_factor) - self._tracer.add_bindings(b=b) - + def blb_subsets(self, n: int, b: int): while True: - yield self.rng.choice(len(xs), b, replace=False) + yield self.rng.choice(n, b, replace=False) def measure_subset(self, xs: NDArray[F], ss: NDArray[np.int64]) -> _BootResult: b = len(ss) @@ -211,7 +220,8 @@ def measure_subset(self, xs: NDArray[F], ss: NDArray[np.int64]) -> _BootResult: lbs = StatAccum(lambda a: np.quantile(a, self._ci_qmin)) ubs = StatAccum(lambda a: np.quantile(a, self._ci_qmax)) - for i, weights in enumerate(self.miniboot_weights(n, b)): + loop = self._rep_generator.subsets() + for i, weights in enumerate(loop): self._tracer.add_bindings(rep=i) self._tracer.trace("starting replicate") assert weights.shape == (b,) @@ -223,7 +233,7 @@ def measure_subset(self, xs: NDArray[F], ss: NDArray[np.int64]) -> _BootResult: ubs.record(stat) if self._check_convergence(means, vars, lbs, ubs, tol=self.tolerance, w=self.r_window): - break + loop.close() df = pd.DataFrame({"statistic": means.values}) df.index.name = "iter" @@ -251,6 +261,75 @@ def _check_convergence(self, *arrays: StatAccum, tol: float, w: int) -> bool: return np.all(gaps < tol).item() +class ReplicateGenerator: + """ + Generate the subset samples for a bootstrap in a background thread. + """ + + n: int + b: int + + _rng: np.random.Generator + _flat: NDArray[np.float64] + _lock: Lock + _notify: Condition + _running: bool = True + _queue: Deque + _thread: Thread + + def __init__(self, n: int, b: int, rng: np.random.Generator): + self.n = n + self.b = b + self._rng = rng.spawn(1)[0] + self._queue = deque() + self._flat = np.full(b, 1.0 / b) + self._lock = Lock() + self._notify = Condition(self._lock) + + def subsets(self) -> Generator[NDArray[np.int64], None, None]: + while True: + with self._notify: + while self._thread.is_alive() and len(self._queue) == 0: + self._notify.wait() + + try: + val = self._queue.popleft() + self._notify.notify_all() + except IndexError: + break # things have shut down, loop is over + except GeneratorExit: + break # we've been asked to close + + yield val + + def _generate(self): + with self._notify: + while True: + # check if we need to wake up + while self._running and len(self._queue) >= 5: + trace(_log, "waiting for queue", len=len(self._queue)) + self._notify.wait() + + # are we done? + if not self._running: + break + + # generate a new value + val = self._rng.multinomial(self.n, self._flat) + self._queue.append(val) + self._notify.notify_all() + + def __enter__(self): + self._thread = Thread(target=self._generate) + self._thread.start() + return self + + def __exit__(self, *args: Any): + with self._notify: + self._running = False + self._notify.notify_all() + + class StatAccum: INIT_SIZE: ClassVar[int] = 100 From cad1ca9ee1692db42ab3b3cdbdfcdc8ded358918 Mon Sep 17 00:00:00 2001 From: Michael Ekstrand Date: Fri, 6 Jun 2025 16:53:28 -0400 Subject: [PATCH 22/59] update defaults and refactor CI width --- src/lenskit/stats/_blb.py | 8 ++++---- src/lenskit/stats/_distributions.py | 24 ++++++++++++++++++++++++ 2 files changed, 28 insertions(+), 4 deletions(-) create mode 100644 src/lenskit/stats/_distributions.py diff --git a/src/lenskit/stats/_blb.py b/src/lenskit/stats/_blb.py index 8e85ae62c..5cfee9778 100644 --- a/src/lenskit/stats/_blb.py +++ b/src/lenskit/stats/_blb.py @@ -21,6 +21,8 @@ from lenskit.logging import Tracer, get_logger, get_tracer, trace from lenskit.random import RNGInput, random_generator +from ._distributions import ci_quantiles + F = TypeVar("F", bound=np.floating, covariant=True) SummaryStat: TypeAlias = Literal["mean"] @@ -50,7 +52,7 @@ def blb_summary( stat: SummaryStat, *, ci_width: float = 0.95, - b_factor: float = 0.8, + b_factor: float = 0.7, rel_tol: float = 0.05, s_window: int = 3, r_window: int = 20, @@ -162,9 +164,7 @@ def __init__( self.rng = rng self.ss_stats = {} - alpha = 1 - ci_width - self._ci_qmin = 0.5 * alpha - self._ci_qmax = 1 - 0.5 * alpha + self._ci_qmin, self._ci_qmax = ci_quantiles(ci_width) self._tracer = get_tracer(_log, stat=stat.__name__) # type: ignore def run_bootstraps(self, xs: NDArray[F]) -> _BootResult: diff --git a/src/lenskit/stats/_distributions.py b/src/lenskit/stats/_distributions.py new file mode 100644 index 000000000..ed2e36df0 --- /dev/null +++ b/src/lenskit/stats/_distributions.py @@ -0,0 +1,24 @@ +# This file is part of LensKit. +# Copyright (C) 2018-2023 Boise State University. +# Copyright (C) 2023-2025 Drexel University. +# Licensed under the MIT license, see LICENSE.md for details. +# SPDX-License-Identifier: MIT + +""" +Distribution utilities. +""" + +from typing import Annotated + +from annotated_types import Gt, Lt +from pydantic import validate_call + + +@validate_call +def ci_quantiles(width: Annotated[float, Gt(0), Lt(1)]) -> tuple[float, float]: + r""" + Convert a confidence interval width to CI quantile bounds. + """ + + margin = 0.5 * (1 - width) + return margin, 1 - margin From 5631d70ace6dde208636f26167631613471b66fe Mon Sep 17 00:00:00 2001 From: Michael Ekstrand Date: Fri, 6 Jun 2025 16:53:39 -0400 Subject: [PATCH 23/59] test for new CI thing --- tests/stats/test_ci_utils.py | 22 ++++++++++++++++++++++ 1 file changed, 22 insertions(+) create mode 100644 tests/stats/test_ci_utils.py diff --git a/tests/stats/test_ci_utils.py b/tests/stats/test_ci_utils.py new file mode 100644 index 000000000..3318d135c --- /dev/null +++ b/tests/stats/test_ci_utils.py @@ -0,0 +1,22 @@ +# This file is part of LensKit. +# Copyright (C) 2018-2023 Boise State University. +# Copyright (C) 2023-2025 Drexel University. +# Licensed under the MIT license, see LICENSE.md for details. +# SPDX-License-Identifier: MIT + +""" +Test confidence interval utilities. +""" + +import hypothesis.strategies as st +from hypothesis import given +from pytest import approx + +from lenskit.stats._distributions import ci_quantiles + + +@given(st.floats(0, 1, exclude_max=True, exclude_min=True)) +def test_ci_bounds(width: float): + qlo, qhi = ci_quantiles(width) + assert qhi - qlo == approx(width) + assert 1 - qhi == approx(qlo) From c243524b37a30961c9cfd24e34995d2c2a2a72c9 Mon Sep 17 00:00:00 2001 From: Michael Ekstrand Date: Fri, 6 Jun 2025 16:54:00 -0400 Subject: [PATCH 24/59] more BLB testing --- notebooks/BLB.ipynb | 484 +++++++++++++++++++++++++++++++++++++--- tests/stats/test_blb.py | 13 +- 2 files changed, 456 insertions(+), 41 deletions(-) diff --git a/notebooks/BLB.ipynb b/notebooks/BLB.ipynb index 2f9a4d355..763c17f72 100644 --- a/notebooks/BLB.ipynb +++ b/notebooks/BLB.ipynb @@ -22,7 +22,7 @@ }, { "cell_type": "code", - "execution_count": null, + "execution_count": 1, "id": "115b4f9e", "metadata": {}, "outputs": [], @@ -31,7 +31,7 @@ "import numpy as np\n", "import pandas as pd\n", "import seaborn as sns\n", - "# from scipy.stats import bootstrap" + "from scipy.stats import bootstrap" ] }, { @@ -64,7 +64,7 @@ }, { "cell_type": "code", - "execution_count": 4, + "execution_count": 38, "id": "85677bc9", "metadata": {}, "outputs": [ @@ -72,13 +72,14 @@ "name": "stdout", "output_type": "stream", "text": [ - "-0.0017 (-0.0036, 0.0003)\n" + "0.0952 (0.0756, 0.1148)\n" ] } ], "source": [ - "N = 1_000_000\n", - "data = rng.normal(0.0, 1.0, N)\n", + "N = 10_000\n", + "TRUE_MEAN = 0.1\n", + "data = rng.normal(TRUE_MEAN, 1.0, N)\n", "mean = np.mean(data)\n", "std = np.std(data)\n", "ste = std / np.sqrt(N)\n", @@ -87,43 +88,34 @@ }, { "cell_type": "code", - "execution_count": 5, + "execution_count": 39, "id": "4bb810c8", "metadata": {}, - "outputs": [], + "outputs": [ + { + "data": { + "text/plain": [ + "ConfidenceInterval(low=np.float64(0.07528506138908903), high=np.float64(0.11462418644027217))" + ] + }, + "execution_count": 39, + "metadata": {}, + "output_type": "execute_result" + } + ], "source": [ - "# bootstrap([data], np.mean).confidence_interval" + "boot_res = bootstrap([data], np.mean)\n", + "boot_res.confidence_interval" ] }, { "cell_type": "code", - "execution_count": 6, + "execution_count": 47, "id": "d136fce6", "metadata": {}, - "outputs": [ - { - "ename": "KeyboardInterrupt", - "evalue": "", - "output_type": "error", - "traceback": [ - "\u001b[31m---------------------------------------------------------------------------\u001b[39m", - "\u001b[31mKeyboardInterrupt\u001b[39m Traceback (most recent call last)", - "\u001b[36mCell\u001b[39m\u001b[36m \u001b[39m\u001b[32mIn[6]\u001b[39m\u001b[32m, line 2\u001b[39m\n\u001b[32m 1\u001b[39m blb = _BLBootstrapper(np.average, \u001b[32m0.95\u001b[39m, \u001b[32m0.05\u001b[39m, \u001b[32m3\u001b[39m, \u001b[32m20\u001b[39m, \u001b[32m0.6\u001b[39m, rng)\n\u001b[32m----> \u001b[39m\u001b[32m2\u001b[39m blb_df = \u001b[43mblb\u001b[49m\u001b[43m.\u001b[49m\u001b[43mrun_bootstraps\u001b[49m\u001b[43m(\u001b[49m\u001b[43mdata\u001b[49m\u001b[43m)\u001b[49m.samples\n\u001b[32m 3\u001b[39m _gstat = blb_df.groupby([\u001b[33m'\u001b[39m\u001b[33msubset\u001b[39m\u001b[33m'\u001b[39m])[\u001b[33m'\u001b[39m\u001b[33mstatistic\u001b[39m\u001b[33m'\u001b[39m]\n\u001b[32m 4\u001b[39m blb_df[\u001b[33m'\u001b[39m\u001b[33mcum_mean\u001b[39m\u001b[33m'\u001b[39m] = _gstat.cumsum() / (_gstat.cumcount() + \u001b[32m1\u001b[39m)\n", - "\u001b[36mFile \u001b[39m\u001b[32m~/LensKit/lkpy/src/lenskit/stats/_blb.py:140\u001b[39m, in \u001b[36m_BLBootstrapper.run_bootstraps\u001b[39m\u001b[34m(self, xs)\u001b[39m\n\u001b[32m 137\u001b[39m ubs = StatAccum(np.mean)\n\u001b[32m 139\u001b[39m \u001b[38;5;28;01mfor\u001b[39;00m i, ss \u001b[38;5;129;01min\u001b[39;00m \u001b[38;5;28menumerate\u001b[39m(\u001b[38;5;28mself\u001b[39m.blb_subsets(xs)):\n\u001b[32m--> \u001b[39m\u001b[32m140\u001b[39m res = \u001b[38;5;28;43mself\u001b[39;49m\u001b[43m.\u001b[49m\u001b[43mmeasure_subset\u001b[49m\u001b[43m(\u001b[49m\u001b[43mxs\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mss\u001b[49m\u001b[43m)\u001b[49m\n\u001b[32m 141\u001b[39m ss_frames[i] = res.samples\n\u001b[32m 142\u001b[39m means.record(res.mean)\n", - "\u001b[36mFile \u001b[39m\u001b[32m~/LensKit/lkpy/src/lenskit/stats/_blb.py:179\u001b[39m, in \u001b[36m_BLBootstrapper.measure_subset\u001b[39m\u001b[34m(self, xs, ss)\u001b[39m\n\u001b[32m 177\u001b[39m values.append(stat)\n\u001b[32m 178\u001b[39m means.record(stat)\n\u001b[32m--> \u001b[39m\u001b[32m179\u001b[39m \u001b[43mlbs\u001b[49m\u001b[43m.\u001b[49m\u001b[43mrecord\u001b[49m\u001b[43m(\u001b[49m\u001b[43mstat\u001b[49m\u001b[43m)\u001b[49m\n\u001b[32m 180\u001b[39m ubs.record(stat)\n\u001b[32m 182\u001b[39m \u001b[38;5;28;01mif\u001b[39;00m _check_convergence(means, svs, lbs, ubs, tol=\u001b[38;5;28mself\u001b[39m.tolerance, w=\u001b[38;5;28mself\u001b[39m.r_window):\n", - "\u001b[36mFile \u001b[39m\u001b[32m~/LensKit/lkpy/src/lenskit/stats/_blb.py:247\u001b[39m, in \u001b[36mStatAccum.record\u001b[39m\u001b[34m(self, x)\u001b[39m\n\u001b[32m 245\u001b[39m \u001b[38;5;66;03m# record and update the cumulative mean\u001b[39;00m\n\u001b[32m 246\u001b[39m \u001b[38;5;28mself\u001b[39m._values[i] = x\n\u001b[32m--> \u001b[39m\u001b[32m247\u001b[39m \u001b[38;5;28mself\u001b[39m._cum_stat[i] = \u001b[38;5;28;43mself\u001b[39;49m\u001b[43m.\u001b[49m\u001b[43m_stat_func\u001b[49m\u001b[43m(\u001b[49m\u001b[38;5;28;43mself\u001b[39;49m\u001b[43m.\u001b[49m\u001b[43mvalues\u001b[49m\u001b[43m)\u001b[49m\n", - "\u001b[36mFile \u001b[39m\u001b[32m~/LensKit/lkpy/src/lenskit/stats/_blb.py:170\u001b[39m, in \u001b[36m_BLBootstrapper.measure_subset..\u001b[39m\u001b[34m(a)\u001b[39m\n\u001b[32m 168\u001b[39m means = StatAccum(np.mean)\n\u001b[32m 169\u001b[39m svs = StatAccum(np.var)\n\u001b[32m--> \u001b[39m\u001b[32m170\u001b[39m lbs = StatAccum(\u001b[38;5;28;01mlambda\u001b[39;00m a: \u001b[43mnp\u001b[49m\u001b[43m.\u001b[49m\u001b[43mquantile\u001b[49m\u001b[43m(\u001b[49m\u001b[43ma\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[38;5;28;43mself\u001b[39;49m\u001b[43m.\u001b[49m\u001b[43m_ci_qmin\u001b[49m\u001b[43m)\u001b[49m)\n\u001b[32m 171\u001b[39m ubs = StatAccum(\u001b[38;5;28;01mlambda\u001b[39;00m a: np.quantile(a, \u001b[38;5;28mself\u001b[39m._ci_qmax))\n\u001b[32m 173\u001b[39m \u001b[38;5;28;01mfor\u001b[39;00m weights \u001b[38;5;129;01min\u001b[39;00m \u001b[38;5;28mself\u001b[39m.miniboot_weights(n, b):\n", - "\u001b[36mFile \u001b[39m\u001b[32m~/LensKit/lkpy/.venv/lib/python3.12/site-packages/numpy/lib/_function_base_impl.py:4537\u001b[39m, in \u001b[36mquantile\u001b[39m\u001b[34m(a, q, axis, out, overwrite_input, method, keepdims, weights, interpolation)\u001b[39m\n\u001b[32m 4534\u001b[39m \u001b[38;5;28;01mif\u001b[39;00m np.any(weights < \u001b[32m0\u001b[39m):\n\u001b[32m 4535\u001b[39m \u001b[38;5;28;01mraise\u001b[39;00m \u001b[38;5;167;01mValueError\u001b[39;00m(\u001b[33m\"\u001b[39m\u001b[33mWeights must be non-negative.\u001b[39m\u001b[33m\"\u001b[39m)\n\u001b[32m-> \u001b[39m\u001b[32m4537\u001b[39m \u001b[38;5;28;01mreturn\u001b[39;00m \u001b[43m_quantile_unchecked\u001b[49m\u001b[43m(\u001b[49m\n\u001b[32m 4538\u001b[39m \u001b[43m \u001b[49m\u001b[43ma\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mq\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43maxis\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mout\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43moverwrite_input\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mmethod\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mkeepdims\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mweights\u001b[49m\u001b[43m)\u001b[49m\n", - "\u001b[36mFile \u001b[39m\u001b[32m~/LensKit/lkpy/.venv/lib/python3.12/site-packages/numpy/lib/_function_base_impl.py:4550\u001b[39m, in \u001b[36m_quantile_unchecked\u001b[39m\u001b[34m(a, q, axis, out, overwrite_input, method, keepdims, weights)\u001b[39m\n\u001b[32m 4541\u001b[39m \u001b[38;5;28;01mdef\u001b[39;00m\u001b[38;5;250m \u001b[39m\u001b[34m_quantile_unchecked\u001b[39m(a,\n\u001b[32m 4542\u001b[39m q,\n\u001b[32m 4543\u001b[39m axis=\u001b[38;5;28;01mNone\u001b[39;00m,\n\u001b[32m (...)\u001b[39m\u001b[32m 4547\u001b[39m keepdims=\u001b[38;5;28;01mFalse\u001b[39;00m,\n\u001b[32m 4548\u001b[39m weights=\u001b[38;5;28;01mNone\u001b[39;00m):\n\u001b[32m 4549\u001b[39m \u001b[38;5;250m \u001b[39m\u001b[33;03m\"\"\"Assumes that q is in [0, 1], and is an ndarray\"\"\"\u001b[39;00m\n\u001b[32m-> \u001b[39m\u001b[32m4550\u001b[39m \u001b[38;5;28;01mreturn\u001b[39;00m \u001b[43m_ureduce\u001b[49m\u001b[43m(\u001b[49m\u001b[43ma\u001b[49m\u001b[43m,\u001b[49m\n\u001b[32m 4551\u001b[39m \u001b[43m \u001b[49m\u001b[43mfunc\u001b[49m\u001b[43m=\u001b[49m\u001b[43m_quantile_ureduce_func\u001b[49m\u001b[43m,\u001b[49m\n\u001b[32m 4552\u001b[39m \u001b[43m \u001b[49m\u001b[43mq\u001b[49m\u001b[43m=\u001b[49m\u001b[43mq\u001b[49m\u001b[43m,\u001b[49m\n\u001b[32m 4553\u001b[39m \u001b[43m \u001b[49m\u001b[43mweights\u001b[49m\u001b[43m=\u001b[49m\u001b[43mweights\u001b[49m\u001b[43m,\u001b[49m\n\u001b[32m 4554\u001b[39m \u001b[43m \u001b[49m\u001b[43mkeepdims\u001b[49m\u001b[43m=\u001b[49m\u001b[43mkeepdims\u001b[49m\u001b[43m,\u001b[49m\n\u001b[32m 4555\u001b[39m \u001b[43m \u001b[49m\u001b[43maxis\u001b[49m\u001b[43m=\u001b[49m\u001b[43maxis\u001b[49m\u001b[43m,\u001b[49m\n\u001b[32m 4556\u001b[39m \u001b[43m \u001b[49m\u001b[43mout\u001b[49m\u001b[43m=\u001b[49m\u001b[43mout\u001b[49m\u001b[43m,\u001b[49m\n\u001b[32m 4557\u001b[39m \u001b[43m \u001b[49m\u001b[43moverwrite_input\u001b[49m\u001b[43m=\u001b[49m\u001b[43moverwrite_input\u001b[49m\u001b[43m,\u001b[49m\n\u001b[32m 4558\u001b[39m \u001b[43m \u001b[49m\u001b[43mmethod\u001b[49m\u001b[43m=\u001b[49m\u001b[43mmethod\u001b[49m\u001b[43m)\u001b[49m\n", - "\u001b[36mFile \u001b[39m\u001b[32m~/LensKit/lkpy/.venv/lib/python3.12/site-packages/numpy/lib/_function_base_impl.py:3894\u001b[39m, in \u001b[36m_ureduce\u001b[39m\u001b[34m(a, func, keepdims, **kwargs)\u001b[39m\n\u001b[32m 3891\u001b[39m index_out = (\u001b[32m0\u001b[39m, ) * nd\n\u001b[32m 3892\u001b[39m kwargs[\u001b[33m'\u001b[39m\u001b[33mout\u001b[39m\u001b[33m'\u001b[39m] = out[(\u001b[38;5;28mEllipsis\u001b[39m, ) + index_out]\n\u001b[32m-> \u001b[39m\u001b[32m3894\u001b[39m r = \u001b[43mfunc\u001b[49m\u001b[43m(\u001b[49m\u001b[43ma\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43m*\u001b[49m\u001b[43m*\u001b[49m\u001b[43mkwargs\u001b[49m\u001b[43m)\u001b[49m\n\u001b[32m 3896\u001b[39m \u001b[38;5;28;01mif\u001b[39;00m out \u001b[38;5;129;01mis\u001b[39;00m \u001b[38;5;129;01mnot\u001b[39;00m \u001b[38;5;28;01mNone\u001b[39;00m:\n\u001b[32m 3897\u001b[39m \u001b[38;5;28;01mreturn\u001b[39;00m out\n", - "\u001b[36mFile \u001b[39m\u001b[32m~/LensKit/lkpy/.venv/lib/python3.12/site-packages/numpy/lib/_function_base_impl.py:4727\u001b[39m, in \u001b[36m_quantile_ureduce_func\u001b[39m\u001b[34m(a, q, weights, axis, out, overwrite_input, method)\u001b[39m\n\u001b[32m 4725\u001b[39m arr = a.copy()\n\u001b[32m 4726\u001b[39m wgt = weights\n\u001b[32m-> \u001b[39m\u001b[32m4727\u001b[39m result = \u001b[43m_quantile\u001b[49m\u001b[43m(\u001b[49m\u001b[43marr\u001b[49m\u001b[43m,\u001b[49m\n\u001b[32m 4728\u001b[39m \u001b[43m \u001b[49m\u001b[43mquantiles\u001b[49m\u001b[43m=\u001b[49m\u001b[43mq\u001b[49m\u001b[43m,\u001b[49m\n\u001b[32m 4729\u001b[39m \u001b[43m \u001b[49m\u001b[43maxis\u001b[49m\u001b[43m=\u001b[49m\u001b[43maxis\u001b[49m\u001b[43m,\u001b[49m\n\u001b[32m 4730\u001b[39m \u001b[43m \u001b[49m\u001b[43mmethod\u001b[49m\u001b[43m=\u001b[49m\u001b[43mmethod\u001b[49m\u001b[43m,\u001b[49m\n\u001b[32m 4731\u001b[39m \u001b[43m \u001b[49m\u001b[43mout\u001b[49m\u001b[43m=\u001b[49m\u001b[43mout\u001b[49m\u001b[43m,\u001b[49m\n\u001b[32m 4732\u001b[39m \u001b[43m \u001b[49m\u001b[43mweights\u001b[49m\u001b[43m=\u001b[49m\u001b[43mwgt\u001b[49m\u001b[43m)\u001b[49m\n\u001b[32m 4733\u001b[39m \u001b[38;5;28;01mreturn\u001b[39;00m result\n", - "\u001b[36mFile \u001b[39m\u001b[32m~/LensKit/lkpy/.venv/lib/python3.12/site-packages/numpy/lib/_function_base_impl.py:4842\u001b[39m, in \u001b[36m_quantile\u001b[39m\u001b[34m(arr, quantiles, axis, method, out, weights)\u001b[39m\n\u001b[32m 4838\u001b[39m previous_indexes, next_indexes = _get_indexes(arr,\n\u001b[32m 4839\u001b[39m virtual_indexes,\n\u001b[32m 4840\u001b[39m values_count)\n\u001b[32m 4841\u001b[39m \u001b[38;5;66;03m# --- Sorting\u001b[39;00m\n\u001b[32m-> \u001b[39m\u001b[32m4842\u001b[39m \u001b[43marr\u001b[49m\u001b[43m.\u001b[49m\u001b[43mpartition\u001b[49m\u001b[43m(\u001b[49m\n\u001b[32m 4843\u001b[39m \u001b[43m \u001b[49m\u001b[43mnp\u001b[49m\u001b[43m.\u001b[49m\u001b[43munique\u001b[49m\u001b[43m(\u001b[49m\u001b[43mnp\u001b[49m\u001b[43m.\u001b[49m\u001b[43mconcatenate\u001b[49m\u001b[43m(\u001b[49m\u001b[43m(\u001b[49m\u001b[43m[\u001b[49m\u001b[32;43m0\u001b[39;49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43m-\u001b[49m\u001b[32;43m1\u001b[39;49m\u001b[43m]\u001b[49m\u001b[43m,\u001b[49m\n\u001b[32m 4844\u001b[39m \u001b[43m \u001b[49m\u001b[43mprevious_indexes\u001b[49m\u001b[43m.\u001b[49m\u001b[43mravel\u001b[49m\u001b[43m(\u001b[49m\u001b[43m)\u001b[49m\u001b[43m,\u001b[49m\n\u001b[32m 4845\u001b[39m \u001b[43m \u001b[49m\u001b[43mnext_indexes\u001b[49m\u001b[43m.\u001b[49m\u001b[43mravel\u001b[49m\u001b[43m(\u001b[49m\u001b[43m)\u001b[49m\u001b[43m,\u001b[49m\n\u001b[32m 4846\u001b[39m \u001b[43m \u001b[49m\u001b[43m)\u001b[49m\u001b[43m)\u001b[49m\u001b[43m)\u001b[49m\u001b[43m,\u001b[49m\n\u001b[32m 4847\u001b[39m \u001b[43m \u001b[49m\u001b[43maxis\u001b[49m\u001b[43m=\u001b[49m\u001b[32;43m0\u001b[39;49m\u001b[43m)\u001b[49m\n\u001b[32m 4848\u001b[39m \u001b[38;5;28;01mif\u001b[39;00m supports_nans:\n\u001b[32m 4849\u001b[39m slices_having_nans = np.isnan(arr[-\u001b[32m1\u001b[39m, ...])\n", - "\u001b[31mKeyboardInterrupt\u001b[39m: " - ] - } - ], + "outputs": [], "source": [ - "blb = _BLBootstrapper(np.average, 0.95, 0.05, 3, 20, 0.6, rng)\n", + "blb = _BLBootstrapper(np.average, 0.95, 0.01, 3, 20, 0.7, rng)\n", "blb_df = blb.run_bootstraps(data).samples\n", "_gstat = blb_df.groupby([\"subset\"])[\"statistic\"]\n", "blb_df[\"cum_mean\"] = _gstat.cumsum() / (_gstat.cumcount() + 1)\n", @@ -132,13 +124,13 @@ }, { "cell_type": "code", - "execution_count": null, + "execution_count": 48, "id": "2e37fc8d", "metadata": {}, "outputs": [ { "data": { - "image/png": "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", + "image/png": "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", "text/plain": [ "
" ] @@ -148,7 +140,7 @@ } ], "source": [ - "plt.hlines([0.0], xmin=0, xmax=blb_df[\"iter\"].max(), label=\"True mean\", color=\"black\")\n", + "plt.hlines([TRUE_MEAN], xmin=0, xmax=blb_df[\"iter\"].max(), label=\"True mean\", color=\"black\")\n", "plt.hlines([mean], xmin=0, xmax=blb_df[\"iter\"].max(), label=\"Data mean\", color=\"magenta\", ls=\"--\")\n", "plt.hlines(\n", " [blb_df[\"statistic\"].mean()],\n", @@ -164,6 +156,426 @@ "plt.show()" ] }, + { + "cell_type": "code", + "execution_count": 49, + "id": "8bda69d7", + "metadata": {}, + "outputs": [ + { + "data": { + "text/html": [ + "
\n", + "\n", + "\n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + "
meanstderrci_lowerci_upper
subset
00.0588310.0095070.0405310.078314
10.0226630.0101400.0044030.042291
20.0724630.0093550.0542760.088416
30.0868480.0088890.0686950.106284
40.1021740.0098070.0839230.121398
50.0982830.0101680.0788210.115761
60.0730010.0099740.0562030.092642
70.1003490.0094260.0827320.117557
80.1905460.0098000.1745660.211778
90.1210830.0098700.1020330.138631
100.1296870.0097790.1105670.147924
110.1131600.0097040.0928460.131451
120.1372200.0094110.1193330.154639
13-0.0011840.009929-0.0203540.018983
140.1103250.0093120.0940950.127731
150.1237170.0105770.1012890.144286
160.0285250.0103090.0073170.048091
170.0996340.0090370.0835950.114753
180.1209310.0098230.0994000.137266
190.1260810.0101830.1065160.147253
200.1157040.0101110.0971870.134352
210.1393770.0093690.1201330.158314
220.0924800.0092050.0733830.109897
230.1773440.0094080.1631700.196364
240.1151440.0106870.0953120.134840
250.0711060.0089660.0555310.089163
260.1090640.0093960.0934410.128746
\n", + "
" + ], + "text/plain": [ + " mean stderr ci_lower ci_upper\n", + "subset \n", + "0 0.058831 0.009507 0.040531 0.078314\n", + "1 0.022663 0.010140 0.004403 0.042291\n", + "2 0.072463 0.009355 0.054276 0.088416\n", + "3 0.086848 0.008889 0.068695 0.106284\n", + "4 0.102174 0.009807 0.083923 0.121398\n", + "5 0.098283 0.010168 0.078821 0.115761\n", + "6 0.073001 0.009974 0.056203 0.092642\n", + "7 0.100349 0.009426 0.082732 0.117557\n", + "8 0.190546 0.009800 0.174566 0.211778\n", + "9 0.121083 0.009870 0.102033 0.138631\n", + "10 0.129687 0.009779 0.110567 0.147924\n", + "11 0.113160 0.009704 0.092846 0.131451\n", + "12 0.137220 0.009411 0.119333 0.154639\n", + "13 -0.001184 0.009929 -0.020354 0.018983\n", + "14 0.110325 0.009312 0.094095 0.127731\n", + "15 0.123717 0.010577 0.101289 0.144286\n", + "16 0.028525 0.010309 0.007317 0.048091\n", + "17 0.099634 0.009037 0.083595 0.114753\n", + "18 0.120931 0.009823 0.099400 0.137266\n", + "19 0.126081 0.010183 0.106516 0.147253\n", + "20 0.115704 0.010111 0.097187 0.134352\n", + "21 0.139377 0.009369 0.120133 0.158314\n", + "22 0.092480 0.009205 0.073383 0.109897\n", + "23 0.177344 0.009408 0.163170 0.196364\n", + "24 0.115144 0.010687 0.095312 0.134840\n", + "25 0.071106 0.008966 0.055531 0.089163\n", + "26 0.109064 0.009396 0.093441 0.128746" + ] + }, + "execution_count": 49, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "blb_sdf = (\n", + " blb_df.groupby(\"subset\")[\"statistic\"]\n", + " .apply(\n", + " lambda x: pd.Series(\n", + " {\n", + " \"mean\": x.mean(),\n", + " \"stderr\": x.std(),\n", + " \"ci_lower\": np.quantile(x, 0.025),\n", + " \"ci_upper\": np.quantile(x, 0.975),\n", + " }\n", + " )\n", + " )\n", + " .unstack()\n", + ")\n", + "blb_sdf" + ] + }, + { + "cell_type": "code", + "execution_count": 50, + "id": "6ecbf7d7", + "metadata": {}, + "outputs": [ + { + "data": { + "image/png": "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", + "text/plain": [ + "
" + ] + }, + "metadata": {}, + "output_type": "display_data" + } + ], + "source": [ + "plt.hlines([TRUE_MEAN], xmin=0, xmax=blb_sdf.index.max(), label=\"True mean\", color=\"black\", ls=\"--\")\n", + "plt.hlines([mean], xmin=0, xmax=blb_sdf.index.max(), label=\"Data mean\", color=\"magenta\", ls=\"-.\")\n", + "plt.plot(\n", + " blb_sdf.index,\n", + " blb_sdf[\"mean\"].cumsum() / (blb_sdf.index.values + 1),\n", + " color=\"steelblue\",\n", + " label=\"BLB mean\",\n", + ")\n", + "plt.xlabel(\"Subset\")\n", + "plt.legend()\n", + "plt.show()" + ] + }, + { + "cell_type": "code", + "execution_count": 51, + "id": "ca8a9542", + "metadata": {}, + "outputs": [ + { + "data": { + "image/png": "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", + "text/plain": [ + "
" + ] + }, + "metadata": {}, + "output_type": "display_data" + } + ], + "source": [ + "plt.hlines(\n", + " [1 / np.sqrt(N)], xmin=0, xmax=blb_sdf.index.max(), label=\"True stderr\", color=\"black\", ls=\"--\"\n", + ")\n", + "plt.hlines([ste], xmin=0, xmax=blb_sdf.index.max(), label=\"Data stderr\", color=\"magenta\", ls=\"-.\")\n", + "plt.hlines(\n", + " [boot_res.standard_error],\n", + " xmin=0,\n", + " xmax=blb_sdf.index.max(),\n", + " label=\"Bootstrap stderr\",\n", + " color=\"green\",\n", + " ls=\":\",\n", + ")\n", + "plt.plot(\n", + " blb_sdf.index,\n", + " blb_sdf[\"stderr\"].cumsum() / (blb_sdf.index.values + 1),\n", + " color=\"steelblue\",\n", + " label=\"BLB stderr\",\n", + ")\n", + "plt.xlabel(\"Subset\")\n", + "plt.legend()\n", + "plt.show()" + ] + }, + { + "cell_type": "code", + "execution_count": 52, + "id": "fa251a70", + "metadata": {}, + "outputs": [ + { + "data": { + "image/png": "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", + "text/plain": [ + "
" + ] + }, + "metadata": {}, + "output_type": "display_data" + } + ], + "source": [ + "plt.hlines([TRUE_MEAN], xmin=0, xmax=blb_sdf.index.max(), label=\"True mean\", color=\"black\")\n", + "plt.hlines(\n", + " [TRUE_MEAN - 1.96 / np.sqrt(N), TRUE_MEAN + 1.96 / np.sqrt(N)],\n", + " xmin=0,\n", + " xmax=blb_sdf.index.max(),\n", + " label=\"True CI\",\n", + " color=\"black\",\n", + " ls=\"--\",\n", + ")\n", + "plt.hlines(\n", + " [mean - 1.96 * ste, mean + 1.96 * ste],\n", + " xmin=0,\n", + " xmax=blb_sdf.index.max(),\n", + " label=\"Data CI\",\n", + " color=\"magenta\",\n", + " ls=\"-.\",\n", + ")\n", + "plt.hlines(\n", + " [boot_res.confidence_interval.low, boot_res.confidence_interval.high],\n", + " xmin=0,\n", + " xmax=blb_sdf.index.max(),\n", + " label=\"Bootstrap CI\",\n", + " color=\"green\",\n", + " ls=\":\",\n", + ")\n", + "plt.plot(\n", + " blb_sdf.index,\n", + " blb_sdf[\"ci_lower\"].cumsum() / (blb_sdf.index.values + 1),\n", + " color=\"steelblue\",\n", + " label=\"BLB CI\",\n", + ")\n", + "plt.plot(\n", + " blb_sdf.index,\n", + " blb_sdf[\"ci_upper\"].cumsum() / (blb_sdf.index.values + 1),\n", + " color=\"steelblue\",\n", + ")\n", + "plt.xlabel(\"Subset\")\n", + "plt.legend()\n", + "plt.show()" + ] + }, { "cell_type": "markdown", "id": "8ed0e077", diff --git a/tests/stats/test_blb.py b/tests/stats/test_blb.py index ad4cc9300..90f0cb0bf 100644 --- a/tests/stats/test_blb.py +++ b/tests/stats/test_blb.py @@ -31,18 +31,21 @@ def test_blb_single_array(rng: np.random.Generator): assert isinstance(summary, dict) assert summary["estimate"] == approx(mean) assert summary["rep_mean"] == approx(mean, rel=0.05) + # assert summary["rep_var"] == approx(ste * ste, rel=0.05) - assert summary["ci_lower"] == approx(mean - 1.96 * ste, rel=0.05) - assert summary["ci_upper"] == approx(mean + 1.96 * ste, rel=0.05) + assert summary["ci_lower"] == approx(mean - 1.96 * ste, rel=0.01) + assert summary["ci_upper"] == approx(mean + 1.96 * ste, rel=0.01) @mark.slow -@mark.parametrize("size", [1000, 10000, 100000]) +@mark.parametrize("size", [1000, 10000]) @mark.filterwarnings(r"error:.*ignoring \d+ nonfinite values") def test_blb_array_normal(rng: np.random.Generator, size: int): "Test BLB with arrays of normals." TRUE_MEAN = 1.0 + TRUE_SD = 1.0 + # TRUE_SVAR = TRUE_SD * TRUE_SD / size results = [] # Test: for 1000 runs, do approx. 95% of confidence intervals contain the @@ -50,10 +53,10 @@ def test_blb_array_normal(rng: np.random.Generator, size: int): NTRIALS = 200 for i in range(NTRIALS): - xs = rng.normal(TRUE_MEAN, 1.0, size) + xs = rng.normal(TRUE_MEAN, TRUE_SD, size) mean = np.mean(xs) - summary = blb_summary(xs, "mean", rng=rng, rel_tol=0.02) + summary = blb_summary(xs, "mean", rng=rng, rel_tol=0.01) assert isinstance(summary, dict) assert summary["estimate"] == approx(mean) From 1829b5140f33e5d286e62a6158bead426edbee8f Mon Sep 17 00:00:00 2001 From: Michael Ekstrand Date: Fri, 6 Jun 2025 17:41:15 -0400 Subject: [PATCH 25/59] refactor config, add but don't use BCA --- src/lenskit/stats/_blb.py | 148 +++++++++++++++++++++++++++++--------- tests/stats/test_blb.py | 2 +- 2 files changed, 115 insertions(+), 35 deletions(-) diff --git a/src/lenskit/stats/_blb.py b/src/lenskit/stats/_blb.py index 5cfee9778..d3dcc3423 100644 --- a/src/lenskit/stats/_blb.py +++ b/src/lenskit/stats/_blb.py @@ -15,6 +15,7 @@ import numpy as np import pandas as pd +import scipy.stats from numpy.typing import NDArray from lenskit.diagnostics import DataWarning @@ -28,6 +29,7 @@ SummaryStat: TypeAlias = Literal["mean"] _log = get_logger(__name__) +STD_NORM = scipy.stats.norm() # dummy assignment to typecheck that we have correctly typed weighted average __dummy_avg: WeightedStatistic = np.average @@ -44,6 +46,7 @@ def __call__( /, *, weights: NDArray[np.floating[Any] | np.integer[Any]] | None = None, + axis: int | None = None, ) -> np.floating[Any]: ... @@ -55,7 +58,7 @@ def blb_summary( b_factor: float = 0.7, rel_tol: float = 0.05, s_window: int = 3, - r_window: int = 20, + r_window: int = 100, rng: RNGInput = None, ) -> dict[str, float]: r""" @@ -103,7 +106,15 @@ def blb_summary( est = np.average(xs).item() rng = random_generator(rng) - bootstrapper = _BLBootstrapper(np.average, ci_width, rel_tol, s_window, r_window, b_factor, rng) + config = _BLBConfig( + statistic=np.average, + ci_width=ci_width, + rel_tol=rel_tol, + s_window=s_window, + r_window=r_window, + b_factor=b_factor, + ) + bootstrapper = _BLBootstrapper(config, rng) result = bootstrapper.run_bootstraps(xs) @@ -120,11 +131,29 @@ def blb_summary( @dataclass class _BootResult: + estimate: float + "Statistic computed on original data." + rep_mean: float + "Mean of the replicates." rep_var: float + "Variance of the replicates." ci_lower: float + "CI lower bound." ci_upper: float - samples: pd.DataFrame + "CI upper bound." + samples: pd.DataFrame | None = None + "Raw sample data." + + +@dataclass +class _BLBConfig: + statistic: WeightedStatistic + ci_width: float + rel_tol: float + s_window: int + r_window: int + b_factor: float class _BLBootstrapper: @@ -133,48 +162,31 @@ class _BLBootstrapper: """ _tracer: Tracer - statistic: WeightedStatistic - ci_width: float + config: _BLBConfig _ci_qmin: float _ci_qmax: float - tolerance: float - s_window: int - r_window: int - b_factor: float rng: np.random.Generator _rep_generator: ReplicateGenerator - def __init__( - self, - stat: WeightedStatistic, - ci_width: float, - tol: float, - s_w: int, - r_w: int, - b_factor: float, - rng: np.random.Generator, - ): - self.statistic = stat - self.ci_width = ci_width - self.tolerance = tol - self.s_window = s_w - self.r_window = r_w - self.b_factor = b_factor + def __init__(self, config, rng: np.random.Generator): + self.config = config self.rng = rng self.ss_stats = {} - self._ci_qmin, self._ci_qmax = ci_quantiles(ci_width) - self._tracer = get_tracer(_log, stat=stat.__name__) # type: ignore + self._ci_qmin, self._ci_qmax = ci_quantiles(config.ci_width) + self._tracer = get_tracer(_log, stat=config.statistic.__name__) # type: ignore def run_bootstraps(self, xs: NDArray[F]) -> _BootResult: n = len(xs) - b = int(n**self.b_factor) + b = int(n**self.config.b_factor) self._tracer.add_bindings(n=n, b=b) _log.debug("starting bootstrap", stat=self.statistic.__name__, n=len(xs)) # type: ignore ss_frames = {} + estimate = float(self.config.statistic(xs)) + means = StatAccum(np.mean) vars = StatAccum(np.mean) lbs = StatAccum(np.mean) @@ -187,18 +199,19 @@ def run_bootstraps(self, xs: NDArray[F]) -> _BootResult: for i, ss in enumerate(self.blb_subsets(n, b)): self._tracer.add_bindings(subset=i) self._tracer.trace("starting subset") - res = self.measure_subset(xs, ss) + res = self.measure_subset(xs, ss, estimate) ss_frames[i] = res.samples means.record(res.rep_mean) vars.record(res.rep_var) lbs.record(res.ci_lower) ubs.record(res.ci_upper) if self._check_convergence( - means, vars, lbs, ubs, tol=self.tolerance, w=self.s_window + means, vars, lbs, ubs, tol=self.config.rel_tol, w=self.config.s_window ): break return _BootResult( + estimate, means.statistic, vars.statistic, lbs.statistic, @@ -210,7 +223,7 @@ def blb_subsets(self, n: int, b: int): while True: yield self.rng.choice(n, b, replace=False) - def measure_subset(self, xs: NDArray[F], ss: NDArray[np.int64]) -> _BootResult: + def measure_subset(self, xs: NDArray[F], ss: NDArray[np.int64], estimate: float) -> _BootResult: b = len(ss) n = len(xs) xss = xs[ss] @@ -226,19 +239,23 @@ def measure_subset(self, xs: NDArray[F], ss: NDArray[np.int64]) -> _BootResult: self._tracer.trace("starting replicate") assert weights.shape == (b,) assert np.sum(weights) == n - stat = self.statistic(xss, weights=weights) + stat = self.config.statistic(xss, weights=weights) means.record(stat) vars.record(stat) lbs.record(stat) ubs.record(stat) - if self._check_convergence(means, vars, lbs, ubs, tol=self.tolerance, w=self.r_window): + if self._check_convergence( + means, vars, lbs, ubs, tol=self.config.rel_tol, w=self.config.r_window + ): loop.close() df = pd.DataFrame({"statistic": means.values}) df.index.name = "iter" self._tracer.remove_bindings("rep") - return _BootResult(means.statistic, vars.statistic, lbs.statistic, ubs.statistic, df) + return _BootResult( + estimate, means.statistic, vars.statistic, lbs.statistic, ubs.statistic, df + ) def miniboot_weights(self, n: int, b: int): flat = np.full(b, 1.0 / b) @@ -378,3 +395,66 @@ def _expand_if_needed(self): def __len__(self): return self._len + + +def _bca_range( + estimate: float, replicates: NDArray[np.floating[Any]], margin: float, accel: float +) -> tuple[float, float]: + """ + Estimate the BCa quantiles for a bootstrap. + + This follows Slide 34 of `http://users.stat.umn.edu/~helwig/notes/bootci-Notes.pdf`_. + """ + bias = _bca_bias_corrector(estimate, replicates) + + z1 = bias + STD_NORM.ppf(margin) + icd1 = z1 / (1 - accel * z1) + + z2 = bias + STD_NORM.ppf(1 - margin) + icd2 = z2 / (1 - accel * z2) + + return STD_NORM.cdf(icd1), STD_NORM.cdf(icd2) + + +def _bca_bias_corrector(statistic: float, replicates: NDArray[np.floating[Any]]) -> float: + frac = np.sum(replicates < statistic) / len(replicates) + return STD_NORM.ppf(frac) + + +def _bca_accel_term(xs: NDArray[np.floating[Any]], statistic: WeightedStatistic) -> float: + """ + Compute the BCa acceleration term. + + Follows slide 36 of + `http://users.stat.umn.edu/~helwig/notes/bootci-Notes.pdf`_, referring also + to the SciPy `scipy/stats/_resampling.py` for implementation ideas. + """ + N = len(xs) + BSIZE = 5000 + jk_vals = np.empty(N) + # batch the jackknife, because our data might be huge + # TODO: can we sample the jackknife? + for start in range(0, N, BSIZE): + end = min(start + BSIZE, N) + B = end - start + # this trick is from scipy — set up a mask + mask = np.ones((B, N), dtype=np.bool_) + np.fill_diagonal(mask[:, start:end], False) + # and reshape — again, borrwed from scipy + i = np.broadcast_to(np.arange(N), (B, N)) + i = i[mask].reshape((B, N - 1)) + + # prepare B x N batched sample and compute statistics + sample = xs[i] + stats = statistic(sample, axis=-1) + assert stats.shape == (B,) + jk_vals[start:end] = stats + + jk_est = np.mean(jk_vals) + jk_dev = jk_est - jk_vals + + # sum of cubes + accel_num = np.sum(np.power(jk_dev, 3)) + # weird term + accel_denom = 6 * np.power(np.sum(np.square(jk_dev)), 1.5) + return accel_num / accel_denom diff --git a/tests/stats/test_blb.py b/tests/stats/test_blb.py index 90f0cb0bf..540f31924 100644 --- a/tests/stats/test_blb.py +++ b/tests/stats/test_blb.py @@ -56,7 +56,7 @@ def test_blb_array_normal(rng: np.random.Generator, size: int): xs = rng.normal(TRUE_MEAN, TRUE_SD, size) mean = np.mean(xs) - summary = blb_summary(xs, "mean", rng=rng, rel_tol=0.01) + summary = blb_summary(xs, "mean", rng=rng) assert isinstance(summary, dict) assert summary["estimate"] == approx(mean) From f48afed8bed604eecd1529ef12e67b38ca8ea16c Mon Sep 17 00:00:00 2001 From: Michael Ekstrand Date: Fri, 6 Jun 2025 17:46:04 -0400 Subject: [PATCH 26/59] fix remaining config run --- src/lenskit/stats/_blb.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/lenskit/stats/_blb.py b/src/lenskit/stats/_blb.py index d3dcc3423..73514dc56 100644 --- a/src/lenskit/stats/_blb.py +++ b/src/lenskit/stats/_blb.py @@ -182,7 +182,7 @@ def run_bootstraps(self, xs: NDArray[F]) -> _BootResult: b = int(n**self.config.b_factor) self._tracer.add_bindings(n=n, b=b) - _log.debug("starting bootstrap", stat=self.statistic.__name__, n=len(xs)) # type: ignore + _log.debug("starting bootstrap", stat=self.config.statistic.__name__, n=len(xs)) # type: ignore ss_frames = {} estimate = float(self.config.statistic(xs)) From d7b3708a4ffa4e736b3b08e08cc8df5d52895ed4 Mon Sep 17 00:00:00 2001 From: Michael Ekstrand Date: Fri, 6 Jun 2025 18:04:24 -0400 Subject: [PATCH 27/59] initial pass on bias-corrected bootstrap --- src/lenskit/stats/_blb.py | 45 ++++++++++++++++++++++++++++++--------- 1 file changed, 35 insertions(+), 10 deletions(-) diff --git a/src/lenskit/stats/_blb.py b/src/lenskit/stats/_blb.py index 73514dc56..45ba04989 100644 --- a/src/lenskit/stats/_blb.py +++ b/src/lenskit/stats/_blb.py @@ -155,6 +155,10 @@ class _BLBConfig: r_window: int b_factor: float + @property + def ci_margin(self) -> float: + return 0.5 * (1 - self.ci_width) + class _BLBootstrapper: """ @@ -192,6 +196,9 @@ def run_bootstraps(self, xs: NDArray[F]) -> _BootResult: lbs = StatAccum(np.mean) ubs = StatAccum(np.mean) + self._tracer.trace("estimating acceleration term") + accel = _bca_accel_term(xs, self.config.statistic) + self._rep_generator = ReplicateGenerator(n, b, self.rng) self._tracer.trace("let's go!") @@ -199,7 +206,7 @@ def run_bootstraps(self, xs: NDArray[F]) -> _BootResult: for i, ss in enumerate(self.blb_subsets(n, b)): self._tracer.add_bindings(subset=i) self._tracer.trace("starting subset") - res = self.measure_subset(xs, ss, estimate) + res = self.measure_subset(xs, ss, estimate, accel) ss_frames[i] = res.samples means.record(res.rep_mean) vars.record(res.rep_var) @@ -223,15 +230,17 @@ def blb_subsets(self, n: int, b: int): while True: yield self.rng.choice(n, b, replace=False) - def measure_subset(self, xs: NDArray[F], ss: NDArray[np.int64], estimate: float) -> _BootResult: + def measure_subset( + self, xs: NDArray[F], ss: NDArray[np.int64], estimate: float, accel: float + ) -> _BootResult: b = len(ss) n = len(xs) xss = xs[ss] means = StatAccum(np.mean) vars = StatAccum(np.var) - lbs = StatAccum(lambda a: np.quantile(a, self._ci_qmin)) - ubs = StatAccum(lambda a: np.quantile(a, self._ci_qmax)) + lbs = StatAccum(None) + ubs = StatAccum(None) loop = self._rep_generator.subsets() for i, weights in enumerate(loop): @@ -242,8 +251,14 @@ def measure_subset(self, xs: NDArray[F], ss: NDArray[np.int64], estimate: float) stat = self.config.statistic(xss, weights=weights) means.record(stat) vars.record(stat) - lbs.record(stat) - ubs.record(stat) + + stats = means.values + ql, qh = _bca_range(estimate, stats, self.config.ci_margin, accel) + self._tracer.trace("bias-corrected quantiles: [%.4f, %.4f]", ql, qh, accel=accel) + lb, ub = np.quantile(stats, [ql, qh]) + lbs.record(stat, lb) + ubs.record(stat, ub) + del stats if self._check_convergence( means, vars, lbs, ubs, tol=self.config.rel_tol, w=self.config.r_window @@ -377,7 +392,9 @@ def statistic(self) -> float: def stat_history(self) -> NDArray[np.float64]: return self._cum_stat[: self._len] - def record(self, x: float | np.floating[Any]) -> None: + def record( + self, x: float | np.floating[Any], stat: float | np.floating[Any] | None = None + ) -> None: "Record a new value in the accumulator." self._expand_if_needed() i = self._len @@ -385,7 +402,9 @@ def record(self, x: float | np.floating[Any]) -> None: # record and update the cumulative mean self._values[i] = x - self._cum_stat[i] = self._stat_func(self.values) + if stat is None: + stat = self._stat_func(self.values) + self._cum_stat[i] = stat def _expand_if_needed(self): cap = len(self._values) @@ -406,6 +425,7 @@ def _bca_range( This follows Slide 34 of `http://users.stat.umn.edu/~helwig/notes/bootci-Notes.pdf`_. """ bias = _bca_bias_corrector(estimate, replicates) + trace(_log, "B=%d, estimate=%f, bias=%f", len(replicates), estimate, bias) z1 = bias + STD_NORM.ppf(margin) icd1 = z1 / (1 - accel * z1) @@ -417,8 +437,13 @@ def _bca_range( def _bca_bias_corrector(statistic: float, replicates: NDArray[np.floating[Any]]) -> float: - frac = np.sum(replicates < statistic) / len(replicates) - return STD_NORM.ppf(frac) + B = len(replicates) + nlow = np.sum(replicates < statistic) + if nlow == 0 or nlow == B: + # extremely biased, but goes OOB. Should only happen early in the bootstrap. + return 0 + else: + return STD_NORM.ppf(nlow / B) def _bca_accel_term(xs: NDArray[np.floating[Any]], statistic: WeightedStatistic) -> float: From f4f8d9c84841e0bf62c3b1b8053718b002264065 Mon Sep 17 00:00:00 2001 From: Michael Ekstrand Date: Fri, 6 Jun 2025 18:07:40 -0400 Subject: [PATCH 28/59] BCA seems to work now --- src/lenskit/stats/_blb.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/lenskit/stats/_blb.py b/src/lenskit/stats/_blb.py index 45ba04989..b4ef8b982 100644 --- a/src/lenskit/stats/_blb.py +++ b/src/lenskit/stats/_blb.py @@ -56,7 +56,7 @@ def blb_summary( *, ci_width: float = 0.95, b_factor: float = 0.7, - rel_tol: float = 0.05, + rel_tol: float = 0.02, s_window: int = 3, r_window: int = 100, rng: RNGInput = None, From 1d93c943fc759793a92601885716b099641ac4cc Mon Sep 17 00:00:00 2001 From: Michael Ekstrand Date: Fri, 6 Jun 2025 18:30:48 -0400 Subject: [PATCH 29/59] do some more re-testing on BLB --- notebooks/BLB.ipynb | 919 +++++++++++++++++++++----------------------- 1 file changed, 441 insertions(+), 478 deletions(-) diff --git a/notebooks/BLB.ipynb b/notebooks/BLB.ipynb index 763c17f72..4a28ec0b8 100644 --- a/notebooks/BLB.ipynb +++ b/notebooks/BLB.ipynb @@ -22,7 +22,7 @@ }, { "cell_type": "code", - "execution_count": 1, + "execution_count": null, "id": "115b4f9e", "metadata": {}, "outputs": [], @@ -41,7 +41,7 @@ "metadata": {}, "outputs": [], "source": [ - "from lenskit.stats._blb import _BLBootstrapper, blb_summary" + "from lenskit.stats._blb import _BLBConfig, _BLBootstrapper, blb_summary" ] }, { @@ -64,7 +64,7 @@ }, { "cell_type": "code", - "execution_count": 38, + "execution_count": 4, "id": "85677bc9", "metadata": {}, "outputs": [ @@ -72,14 +72,15 @@ "name": "stdout", "output_type": "stream", "text": [ - "0.0952 (0.0756, 0.1148)\n" + "0.2508 (0.2470, 0.2546)\n" ] } ], "source": [ "N = 10_000\n", - "TRUE_MEAN = 0.1\n", - "data = rng.normal(TRUE_MEAN, 1.0, N)\n", + "TRUE_MEAN = 0.25\n", + "TRUE_SD = np.sqrt(3 / ((1 + 3) ** 2 * (1 + 3 + 1)))\n", + "data = rng.beta(1, 3, N)\n", "mean = np.mean(data)\n", "std = np.std(data)\n", "ste = std / np.sqrt(N)\n", @@ -88,17 +89,17 @@ }, { "cell_type": "code", - "execution_count": 39, + "execution_count": 5, "id": "4bb810c8", "metadata": {}, "outputs": [ { "data": { "text/plain": [ - "ConfidenceInterval(low=np.float64(0.07528506138908903), high=np.float64(0.11462418644027217))" + "ConfidenceInterval(low=np.float64(0.24723537123505493), high=np.float64(0.2546752145819843))" ] }, - "execution_count": 39, + "execution_count": 5, "metadata": {}, "output_type": "execute_result" } @@ -110,12 +111,13 @@ }, { "cell_type": "code", - "execution_count": 47, + "execution_count": 6, "id": "d136fce6", "metadata": {}, "outputs": [], "source": [ - "blb = _BLBootstrapper(np.average, 0.95, 0.01, 3, 20, 0.7, rng)\n", + "config = _BLBConfig(np.average, 0.95, 0.01, 3, 200, 0.7)\n", + "blb = _BLBootstrapper(config, rng)\n", "blb_df = blb.run_bootstraps(data).samples\n", "_gstat = blb_df.groupby([\"subset\"])[\"statistic\"]\n", "blb_df[\"cum_mean\"] = _gstat.cumsum() / (_gstat.cumcount() + 1)\n", @@ -124,13 +126,13 @@ }, { "cell_type": "code", - "execution_count": 48, + "execution_count": 7, "id": "2e37fc8d", "metadata": {}, "outputs": [ { "data": { - "image/png": "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", + "image/png": "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", "text/plain": [ "
" ] @@ -158,7 +160,7 @@ }, { "cell_type": "code", - "execution_count": 49, + "execution_count": 8, "id": "8bda69d7", "metadata": {}, "outputs": [ @@ -199,192 +201,45 @@ " \n", " \n", " 0\n", - " 0.058831\n", - " 0.009507\n", - " 0.040531\n", - " 0.078314\n", + " 0.246118\n", + " 0.001976\n", + " 0.242340\n", + " 0.250043\n", " \n", " \n", " 1\n", - " 0.022663\n", - " 0.010140\n", - " 0.004403\n", - " 0.042291\n", + " 0.248454\n", + " 0.002031\n", + " 0.244547\n", + " 0.252199\n", " \n", " \n", " 2\n", - " 0.072463\n", - " 0.009355\n", - " 0.054276\n", - " 0.088416\n", + " 0.253811\n", + " 0.001847\n", + " 0.250333\n", + " 0.257304\n", " \n", " \n", " 3\n", - " 0.086848\n", - " 0.008889\n", - " 0.068695\n", - " 0.106284\n", + " 0.239053\n", + " 0.001960\n", + " 0.235347\n", + " 0.242675\n", " \n", " \n", " 4\n", - " 0.102174\n", - " 0.009807\n", - " 0.083923\n", - " 0.121398\n", + " 0.244940\n", + " 0.001908\n", + " 0.241283\n", + " 0.248641\n", " \n", " \n", " 5\n", - " 0.098283\n", - " 0.010168\n", - " 0.078821\n", - " 0.115761\n", - " \n", - " \n", - " 6\n", - " 0.073001\n", - " 0.009974\n", - " 0.056203\n", - " 0.092642\n", - " \n", - " \n", - " 7\n", - " 0.100349\n", - " 0.009426\n", - " 0.082732\n", - " 0.117557\n", - " \n", - " \n", - " 8\n", - " 0.190546\n", - " 0.009800\n", - " 0.174566\n", - " 0.211778\n", - " \n", - " \n", - " 9\n", - " 0.121083\n", - " 0.009870\n", - " 0.102033\n", - " 0.138631\n", - " \n", - " \n", - " 10\n", - " 0.129687\n", - " 0.009779\n", - " 0.110567\n", - " 0.147924\n", - " \n", - " \n", - " 11\n", - " 0.113160\n", - " 0.009704\n", - " 0.092846\n", - " 0.131451\n", - " \n", - " \n", - " 12\n", - " 0.137220\n", - " 0.009411\n", - " 0.119333\n", - " 0.154639\n", - " \n", - " \n", - " 13\n", - " -0.001184\n", - " 0.009929\n", - " -0.020354\n", - " 0.018983\n", - " \n", - " \n", - " 14\n", - " 0.110325\n", - " 0.009312\n", - " 0.094095\n", - " 0.127731\n", - " \n", - " \n", - " 15\n", - " 0.123717\n", - " 0.010577\n", - " 0.101289\n", - " 0.144286\n", - " \n", - " \n", - " 16\n", - " 0.028525\n", - " 0.010309\n", - " 0.007317\n", - " 0.048091\n", - " \n", - " \n", - " 17\n", - " 0.099634\n", - " 0.009037\n", - " 0.083595\n", - " 0.114753\n", - " \n", - " \n", - " 18\n", - " 0.120931\n", - " 0.009823\n", - " 0.099400\n", - " 0.137266\n", - " \n", - " \n", - " 19\n", - " 0.126081\n", - " 0.010183\n", - " 0.106516\n", - " 0.147253\n", - " \n", - " \n", - " 20\n", - " 0.115704\n", - " 0.010111\n", - " 0.097187\n", - " 0.134352\n", - " \n", - " \n", - " 21\n", - " 0.139377\n", - " 0.009369\n", - " 0.120133\n", - " 0.158314\n", - " \n", - " \n", - " 22\n", - " 0.092480\n", - " 0.009205\n", - " 0.073383\n", - " 0.109897\n", - " \n", - " \n", - " 23\n", - " 0.177344\n", - " 0.009408\n", - " 0.163170\n", - " 0.196364\n", - " \n", - " \n", - " 24\n", - " 0.115144\n", - " 0.010687\n", - " 0.095312\n", - " 0.134840\n", - " \n", - " \n", - " 25\n", - " 0.071106\n", - " 0.008966\n", - " 0.055531\n", - " 0.089163\n", - " \n", - " \n", - " 26\n", - " 0.109064\n", - " 0.009396\n", - " 0.093441\n", - " 0.128746\n", + " 0.259034\n", + " 0.001963\n", + " 0.255048\n", + " 0.262777\n", " \n", " \n", "\n", @@ -393,36 +248,15 @@ "text/plain": [ " mean stderr ci_lower ci_upper\n", "subset \n", - "0 0.058831 0.009507 0.040531 0.078314\n", - "1 0.022663 0.010140 0.004403 0.042291\n", - "2 0.072463 0.009355 0.054276 0.088416\n", - "3 0.086848 0.008889 0.068695 0.106284\n", - "4 0.102174 0.009807 0.083923 0.121398\n", - "5 0.098283 0.010168 0.078821 0.115761\n", - "6 0.073001 0.009974 0.056203 0.092642\n", - "7 0.100349 0.009426 0.082732 0.117557\n", - "8 0.190546 0.009800 0.174566 0.211778\n", - "9 0.121083 0.009870 0.102033 0.138631\n", - "10 0.129687 0.009779 0.110567 0.147924\n", - "11 0.113160 0.009704 0.092846 0.131451\n", - "12 0.137220 0.009411 0.119333 0.154639\n", - "13 -0.001184 0.009929 -0.020354 0.018983\n", - "14 0.110325 0.009312 0.094095 0.127731\n", - "15 0.123717 0.010577 0.101289 0.144286\n", - "16 0.028525 0.010309 0.007317 0.048091\n", - "17 0.099634 0.009037 0.083595 0.114753\n", - "18 0.120931 0.009823 0.099400 0.137266\n", - "19 0.126081 0.010183 0.106516 0.147253\n", - "20 0.115704 0.010111 0.097187 0.134352\n", - "21 0.139377 0.009369 0.120133 0.158314\n", - "22 0.092480 0.009205 0.073383 0.109897\n", - "23 0.177344 0.009408 0.163170 0.196364\n", - "24 0.115144 0.010687 0.095312 0.134840\n", - "25 0.071106 0.008966 0.055531 0.089163\n", - "26 0.109064 0.009396 0.093441 0.128746" + "0 0.246118 0.001976 0.242340 0.250043\n", + "1 0.248454 0.002031 0.244547 0.252199\n", + "2 0.253811 0.001847 0.250333 0.257304\n", + "3 0.239053 0.001960 0.235347 0.242675\n", + "4 0.244940 0.001908 0.241283 0.248641\n", + "5 0.259034 0.001963 0.255048 0.262777" ] }, - "execution_count": 49, + "execution_count": 8, "metadata": {}, "output_type": "execute_result" } @@ -447,13 +281,13 @@ }, { "cell_type": "code", - "execution_count": 50, + "execution_count": 9, "id": "6ecbf7d7", "metadata": {}, "outputs": [ { "data": { - "image/png": "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", + "image/png": "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", "text/plain": [ "
" ] @@ -478,13 +312,13 @@ }, { "cell_type": "code", - "execution_count": 51, + "execution_count": 10, "id": "ca8a9542", "metadata": {}, "outputs": [ { "data": { - "image/png": "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", + "image/png": "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", "text/plain": [ "
" ] @@ -495,7 +329,12 @@ ], "source": [ "plt.hlines(\n", - " [1 / np.sqrt(N)], xmin=0, xmax=blb_sdf.index.max(), label=\"True stderr\", color=\"black\", ls=\"--\"\n", + " [TRUE_SD / np.sqrt(N)],\n", + " xmin=0,\n", + " xmax=blb_sdf.index.max(),\n", + " label=\"True stderr\",\n", + " color=\"black\",\n", + " ls=\"--\",\n", ")\n", "plt.hlines([ste], xmin=0, xmax=blb_sdf.index.max(), label=\"Data stderr\", color=\"magenta\", ls=\"-.\")\n", "plt.hlines(\n", @@ -519,13 +358,13 @@ }, { "cell_type": "code", - "execution_count": 52, + "execution_count": 11, "id": "fa251a70", "metadata": {}, "outputs": [ { "data": { - "image/png": "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", + "image/png": "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", "text/plain": [ "
" ] @@ -537,7 +376,7 @@ "source": [ "plt.hlines([TRUE_MEAN], xmin=0, xmax=blb_sdf.index.max(), label=\"True mean\", color=\"black\")\n", "plt.hlines(\n", - " [TRUE_MEAN - 1.96 / np.sqrt(N), TRUE_MEAN + 1.96 / np.sqrt(N)],\n", + " [TRUE_MEAN - 1.96 * TRUE_SD / np.sqrt(N), TRUE_MEAN + 1.96 * TRUE_SD / np.sqrt(N)],\n", " xmin=0,\n", " xmax=blb_sdf.index.max(),\n", " label=\"True CI\",\n", @@ -588,25 +427,25 @@ }, { "cell_type": "code", - "execution_count": null, + "execution_count": 25, "id": "509893b4", "metadata": {}, "outputs": [ { "data": { "text/plain": [ - "(200, 100000)" + "(100, 10000)" ] }, - "execution_count": 51, + "execution_count": 25, "metadata": {}, "output_type": "execute_result" } ], "source": [ - "M = 200\n", - "N = 100_000\n", - "means = rng.normal(0, 10, size=M)\n", + "M = 100\n", + "N = 10_000\n", + "means = rng.beta(1, 3, size=M)\n", "stds = rng.standard_exponential(size=M) + 0.1\n", "\n", "data = rng.normal(\n", @@ -617,7 +456,7 @@ }, { "cell_type": "code", - "execution_count": null, + "execution_count": 26, "id": "1a38c5ce", "metadata": {}, "outputs": [], @@ -626,7 +465,8 @@ "data_stds = np.std(data, axis=1)\n", "param_stats = pd.DataFrame(\n", " {\n", - " \"mean\": data_means,\n", + " \"rep_mean\": data_means,\n", + " \"rep_var\": (data_stds * data_stds) / N,\n", " \"ci_lower\": data_means - 1.96 * (data_stds / np.sqrt(N)),\n", " \"ci_upper\": data_means - 1.96 * (data_stds / np.sqrt(N)),\n", " }\n", @@ -635,7 +475,7 @@ }, { "cell_type": "code", - "execution_count": null, + "execution_count": 27, "id": "3a64a3c6", "metadata": {}, "outputs": [], @@ -654,35 +494,163 @@ }, { "cell_type": "code", - "execution_count": null, + "execution_count": 28, "id": "c59e3e12", "metadata": {}, "outputs": [ { - "ename": "KeyboardInterrupt", - "evalue": "", - "output_type": "error", - "traceback": [ - "\u001b[31m---------------------------------------------------------------------------\u001b[39m", - "\u001b[31mKeyboardInterrupt\u001b[39m Traceback (most recent call last)", - "\u001b[36mCell\u001b[39m\u001b[36m \u001b[39m\u001b[32mIn[54]\u001b[39m\u001b[32m, line 1\u001b[39m\n\u001b[32m----> \u001b[39m\u001b[32m1\u001b[39m blbs = [\u001b[43mblb_summary\u001b[49m\u001b[43m(\u001b[49m\u001b[43mdata\u001b[49m\u001b[43m[\u001b[49m\u001b[43mi\u001b[49m\u001b[43m,\u001b[49m\u001b[43m:\u001b[49m\u001b[43m]\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[33;43m'\u001b[39;49m\u001b[33;43mmean\u001b[39;49m\u001b[33;43m'\u001b[39;49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mb_factor\u001b[49m\u001b[43m=\u001b[49m\u001b[32;43m0.8\u001b[39;49m\u001b[43m)\u001b[49m \u001b[38;5;28;01mfor\u001b[39;00m i \u001b[38;5;129;01min\u001b[39;00m \u001b[38;5;28mrange\u001b[39m(M)]\n\u001b[32m 2\u001b[39m blb_stats = pd.DataFrame.from_records(blbs)\n\u001b[32m 3\u001b[39m blb_stats\n", - "\u001b[36mFile \u001b[39m\u001b[32m~/LensKit/lkpy/src/lenskit/stats/_blb.py:69\u001b[39m, in \u001b[36mblb_summary\u001b[39m\u001b[34m(xs, stat, ci_width, b_factor, rel_tol, s_window, r_window, rng)\u001b[39m\n\u001b[32m 0\u001b[39m \n", - "\u001b[36mFile \u001b[39m\u001b[32m~/LensKit/lkpy/src/lenskit/stats/_blb.py:139\u001b[39m, in \u001b[36mrun_bootstraps\u001b[39m\u001b[34m(self, xs)\u001b[39m\n\u001b[32m 136\u001b[39m lbs = StatAccum(np.mean)\n\u001b[32m 137\u001b[39m ubs = StatAccum(np.mean)\n\u001b[32m--> \u001b[39m\u001b[32m139\u001b[39m \u001b[38;5;28;01mfor\u001b[39;00m i, ss \u001b[38;5;129;01min\u001b[39;00m \u001b[38;5;28menumerate\u001b[39m(\u001b[38;5;28mself\u001b[39m.blb_subsets(xs)):\n\u001b[32m 140\u001b[39m res = \u001b[38;5;28mself\u001b[39m.measure_subset(xs, ss)\n\u001b[32m 141\u001b[39m ss_frames[i] = res.samples\n", - "\u001b[36mFile \u001b[39m\u001b[32m~/LensKit/lkpy/src/lenskit/stats/_blb.py:168\u001b[39m, in \u001b[36mmeasure_subset\u001b[39m\u001b[34m(self, xs, ss)\u001b[39m\n\u001b[32m 165\u001b[39m xss = xs[ss]\n\u001b[32m 167\u001b[39m values = []\n\u001b[32m--> \u001b[39m\u001b[32m168\u001b[39m means = StatAccum(np.mean)\n\u001b[32m 169\u001b[39m svs = StatAccum(np.var)\n\u001b[32m 170\u001b[39m lbs = StatAccum(\u001b[38;5;28;01mlambda\u001b[39;00m a: np.quantile(a, \u001b[38;5;28mself\u001b[39m._ci_qmin))\n", - "\u001b[36mFile \u001b[39m\u001b[32m~/LensKit/lkpy/src/lenskit/stats/_blb.py:189\u001b[39m, in \u001b[36mminiboot_weights\u001b[39m\u001b[34m(self, n, b)\u001b[39m\n\u001b[32m 186\u001b[39m df.index.name = \u001b[33m\"\u001b[39m\u001b[33miter\u001b[39m\u001b[33m\"\u001b[39m\n\u001b[32m 187\u001b[39m \u001b[38;5;28;01mreturn\u001b[39;00m _BootResult(means.statistic, svs.statistic, lbs.statistic, ubs.statistic, df)\n\u001b[32m--> \u001b[39m\u001b[32m189\u001b[39m \u001b[38;5;28;01mdef\u001b[39;00m\u001b[38;5;250m \u001b[39m\u001b[34mminiboot_weights\u001b[39m(\u001b[38;5;28mself\u001b[39m, n: \u001b[38;5;28mint\u001b[39m, b: \u001b[38;5;28mint\u001b[39m):\n\u001b[32m 190\u001b[39m flat = np.full(b, \u001b[32m1.0\u001b[39m / b)\n\u001b[32m 192\u001b[39m \u001b[38;5;28;01mwhile\u001b[39;00m \u001b[38;5;28;01mTrue\u001b[39;00m:\n", - "\u001b[31mKeyboardInterrupt\u001b[39m: " - ] + "data": { + "text/html": [ + "
\n", + "\n", + "\n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + "
estimaterep_meanrep_varci_lowerci_upper
00.1555560.1602160.0000010.1576330.162040
10.1037240.0979680.0005610.0710800.148033
20.3435960.3260110.0003130.2902500.354611
30.0751310.0697430.0000100.0624660.073803
40.0726330.0731200.0000020.0706710.074739
..................
950.0646430.0784660.0000340.0685940.090591
960.4491190.4361310.0000410.4282120.448650
970.1568580.1597170.0003380.1268300.194622
980.1272990.1210390.0000470.1129420.136177
990.5343620.5319470.0000070.5284200.537567
\n", + "

100 rows × 5 columns

\n", + "
" + ], + "text/plain": [ + " estimate rep_mean rep_var ci_lower ci_upper\n", + "0 0.155556 0.160216 0.000001 0.157633 0.162040\n", + "1 0.103724 0.097968 0.000561 0.071080 0.148033\n", + "2 0.343596 0.326011 0.000313 0.290250 0.354611\n", + "3 0.075131 0.069743 0.000010 0.062466 0.073803\n", + "4 0.072633 0.073120 0.000002 0.070671 0.074739\n", + ".. ... ... ... ... ...\n", + "95 0.064643 0.078466 0.000034 0.068594 0.090591\n", + "96 0.449119 0.436131 0.000041 0.428212 0.448650\n", + "97 0.156858 0.159717 0.000338 0.126830 0.194622\n", + "98 0.127299 0.121039 0.000047 0.112942 0.136177\n", + "99 0.534362 0.531947 0.000007 0.528420 0.537567\n", + "\n", + "[100 rows x 5 columns]" + ] + }, + "execution_count": 28, + "metadata": {}, + "output_type": "execute_result" } ], "source": [ - "blbs = [blb_summary(data[i, :], \"mean\", b_factor=0.8) for i in range(M)]\n", + "blbs = [blb_summary(data[i, :], \"mean\", rel_tol=0.05) for i in range(M)]\n", "blb_stats = pd.DataFrame.from_records(blbs)\n", "blb_stats" ] }, { "cell_type": "code", - "execution_count": null, + "execution_count": 29, "id": "b229ea22", "metadata": {}, "outputs": [ @@ -709,7 +677,6 @@ " \n", " \n", " Parametric\n", - " Bootstrap\n", " BLB\n", " Error\n", " RelError\n", @@ -721,50 +688,44 @@ " \n", " \n", " \n", - " \n", " \n", " \n", " \n", " \n", - " ci_lower\n", + " rep_mean\n", " 0\n", - " 2.597184\n", - " -4.423093\n", - " 2.593609\n", - " -0.003575\n", - " 0.001377\n", + " 0.155556\n", + " 0.160216\n", + " 0.004660\n", + " 0.029956\n", " \n", " \n", " 1\n", - " -12.386757\n", - " -4.425957\n", - " -12.386300\n", - " 0.000458\n", - " 0.000037\n", + " 0.103724\n", + " 0.097968\n", + " -0.005757\n", + " 0.055498\n", " \n", " \n", " 2\n", - " 13.514136\n", - " -8.080191\n", - " 13.510683\n", - " -0.003453\n", - " 0.000256\n", + " 0.343596\n", + " 0.326011\n", + " -0.017584\n", + " 0.051178\n", " \n", " \n", " 3\n", - " 9.921847\n", - " -25.329503\n", - " 9.921969\n", - " 0.000122\n", - " 0.000012\n", + " 0.075131\n", + " 0.069743\n", + " -0.005388\n", + " 0.071712\n", " \n", " \n", " 4\n", - " -0.510517\n", - " -2.748500\n", - " -0.509992\n", - " 0.000526\n", - " 0.001029\n", + " 0.072633\n", + " 0.073120\n", + " 0.000487\n", + " 0.006699\n", " \n", " \n", " ...\n", @@ -773,73 +734,67 @@ " ...\n", " ...\n", " ...\n", - " ...\n", " \n", " \n", - " mean\n", - " 195\n", - " 1.883268\n", - " NaN\n", - " 1.884418\n", - " 0.001150\n", - " 0.000611\n", + " ci_upper\n", + " 95\n", + " 0.052551\n", + " 0.090591\n", + " 0.038041\n", + " 0.723889\n", " \n", " \n", - " 196\n", - " -12.321563\n", - " NaN\n", - " -12.321220\n", - " 0.000343\n", - " 0.000028\n", + " 96\n", + " 0.435598\n", + " 0.448650\n", + " 0.013052\n", + " 0.029963\n", " \n", " \n", - " 197\n", - " 19.500813\n", - " NaN\n", - " 19.500918\n", - " 0.000105\n", - " 0.000005\n", + " 97\n", + " 0.119923\n", + " 0.194622\n", + " 0.074699\n", + " 0.622887\n", " \n", " \n", - " 198\n", - " -7.208713\n", - " NaN\n", - " -7.208516\n", - " 0.000196\n", - " 0.000027\n", + " 98\n", + " 0.113820\n", + " 0.136177\n", + " 0.022357\n", + " 0.196422\n", " \n", " \n", - " 199\n", - " 9.404558\n", - " NaN\n", - " 9.404817\n", - " 0.000259\n", - " 0.000028\n", + " 99\n", + " 0.529435\n", + " 0.537567\n", + " 0.008131\n", + " 0.015359\n", " \n", " \n", "\n", - "

600 rows × 5 columns

\n", + "

400 rows × 4 columns

\n", "" ], "text/plain": [ - " Parametric Bootstrap BLB Error RelError\n", - "quantity samp \n", - "ci_lower 0 2.597184 -4.423093 2.593609 -0.003575 0.001377\n", - " 1 -12.386757 -4.425957 -12.386300 0.000458 0.000037\n", - " 2 13.514136 -8.080191 13.510683 -0.003453 0.000256\n", - " 3 9.921847 -25.329503 9.921969 0.000122 0.000012\n", - " 4 -0.510517 -2.748500 -0.509992 0.000526 0.001029\n", - "... ... ... ... ... ...\n", - "mean 195 1.883268 NaN 1.884418 0.001150 0.000611\n", - " 196 -12.321563 NaN -12.321220 0.000343 0.000028\n", - " 197 19.500813 NaN 19.500918 0.000105 0.000005\n", - " 198 -7.208713 NaN -7.208516 0.000196 0.000027\n", - " 199 9.404558 NaN 9.404817 0.000259 0.000028\n", + " Parametric BLB Error RelError\n", + "quantity samp \n", + "rep_mean 0 0.155556 0.160216 0.004660 0.029956\n", + " 1 0.103724 0.097968 -0.005757 0.055498\n", + " 2 0.343596 0.326011 -0.017584 0.051178\n", + " 3 0.075131 0.069743 -0.005388 0.071712\n", + " 4 0.072633 0.073120 0.000487 0.006699\n", + "... ... ... ... ...\n", + "ci_upper 95 0.052551 0.090591 0.038041 0.723889\n", + " 96 0.435598 0.448650 0.013052 0.029963\n", + " 97 0.119923 0.194622 0.074699 0.622887\n", + " 98 0.113820 0.136177 0.022357 0.196422\n", + " 99 0.529435 0.537567 0.008131 0.015359\n", "\n", - "[600 rows x 5 columns]" + "[400 rows x 4 columns]" ] }, - "execution_count": 37, + "execution_count": 29, "metadata": {}, "output_type": "execute_result" } @@ -849,7 +804,7 @@ " {\n", " \"Parametric\": param_stats.unstack(),\n", " # \"Bootstrap\": boot_stats.unstack(),\n", - " \"BLB\": blb_stats.drop(columns=[\"value\"]).unstack(),\n", + " \"BLB\": blb_stats.drop(columns=[\"estimate\"]).unstack(),\n", " }\n", ")\n", "comb_stats.index.rename([\"quantity\", \"samp\"], inplace=True)\n", @@ -862,7 +817,7 @@ }, { "cell_type": "code", - "execution_count": null, + "execution_count": 30, "id": "41e29fb8", "metadata": {}, "outputs": [ @@ -890,7 +845,6 @@ " quantity\n", " samp\n", " Parametric\n", - " Bootstrap\n", " BLB\n", " Error\n", " RelError\n", @@ -902,68 +856,63 @@ " \n", " \n", " 0\n", - " ci_lower\n", + " rep_mean\n", " 0\n", - " 2.597184\n", - " -4.423093\n", - " 2.593609\n", - " -0.003575\n", - " 0.001377\n", - " 2.601141\n", - " 1.592588\n", - " 2.601141\n", + " 0.155556\n", + " 0.160216\n", + " 0.004660\n", + " 0.029956\n", + " 0.154939\n", + " 0.123640\n", + " 0.154939\n", " \n", " \n", " 1\n", - " ci_lower\n", + " rep_mean\n", " 1\n", - " -12.386757\n", - " -4.425957\n", - " -12.386300\n", - " 0.000458\n", - " 0.000037\n", - " -12.385795\n", - " 0.547635\n", - " 12.385795\n", + " 0.103724\n", + " 0.097968\n", + " -0.005757\n", + " 0.055498\n", + " 0.083691\n", + " 2.412370\n", + " 0.083691\n", " \n", " \n", " 2\n", - " ci_lower\n", + " rep_mean\n", " 2\n", - " 13.514136\n", - " -8.080191\n", - " 13.510683\n", - " -0.003453\n", - " 0.000256\n", - " 13.514852\n", - " 1.421265\n", - " 13.514852\n", + " 0.343596\n", + " 0.326011\n", + " -0.017584\n", + " 0.051178\n", + " 0.381152\n", + " 1.841601\n", + " 0.381152\n", " \n", " \n", " 3\n", - " ci_lower\n", + " rep_mean\n", " 3\n", - " 9.921847\n", - " -25.329503\n", - " 9.921969\n", - " 0.000122\n", - " 0.000012\n", - " 9.922137\n", - " 0.140994\n", - " 9.922137\n", + " 0.075131\n", + " 0.069743\n", + " -0.005388\n", + " 0.071712\n", + " 0.081031\n", + " 0.337456\n", + " 0.081031\n", " \n", " \n", " 4\n", - " ci_lower\n", + " rep_mean\n", " 4\n", - " -0.510517\n", - " -2.748500\n", - " -0.509992\n", - " 0.000526\n", - " 0.001029\n", - " -0.509412\n", - " 0.362041\n", - " 0.509412\n", + " 0.072633\n", + " 0.073120\n", + " 0.000487\n", + " 0.006699\n", + " 0.072287\n", + " 0.121747\n", + " 0.072287\n", " \n", " \n", " ...\n", @@ -976,109 +925,103 @@ " ...\n", " ...\n", " ...\n", - " ...\n", " \n", " \n", - " 595\n", - " mean\n", - " 195\n", - " 1.883268\n", - " NaN\n", - " 1.884418\n", - " 0.001150\n", - " 0.000611\n", - " 1.884801\n", - " 2.149275\n", - " 1.884801\n", + " 395\n", + " ci_upper\n", + " 95\n", + " 0.052551\n", + " 0.090591\n", + " 0.038041\n", + " 0.723889\n", + " 0.071423\n", + " 0.622047\n", + " 0.071423\n", " \n", " \n", - " 596\n", - " mean\n", - " 196\n", - " -12.321563\n", - " NaN\n", - " -12.321220\n", - " 0.000343\n", - " 0.000028\n", - " -12.319299\n", - " 2.509345\n", - " 12.319299\n", + " 396\n", + " ci_upper\n", + " 96\n", + " 0.435598\n", + " 0.448650\n", + " 0.013052\n", + " 0.029963\n", + " 0.454315\n", + " 0.693925\n", + " 0.454315\n", " \n", " \n", - " 597\n", - " mean\n", - " 197\n", - " 19.500813\n", - " NaN\n", - " 19.500918\n", - " 0.000105\n", - " 0.000005\n", - " 19.501153\n", - " 0.612372\n", - " 19.501153\n", + " 397\n", + " ci_upper\n", + " 97\n", + " 0.119923\n", + " 0.194622\n", + " 0.074699\n", + " 0.622887\n", + " 0.171009\n", + " 1.900373\n", + " 0.171009\n", " \n", " \n", - " 598\n", - " mean\n", - " 198\n", - " -7.208713\n", - " NaN\n", - " -7.208516\n", - " 0.000196\n", - " 0.000027\n", - " -7.209443\n", - " 0.213760\n", - " 7.209443\n", + " 398\n", + " ci_upper\n", + " 98\n", + " 0.113820\n", + " 0.136177\n", + " 0.022357\n", + " 0.196422\n", + " 0.128295\n", + " 0.687364\n", + " 0.128295\n", " \n", " \n", - " 599\n", - " mean\n", - " 199\n", - " 9.404558\n", - " NaN\n", - " 9.404817\n", - " 0.000259\n", - " 0.000028\n", - " 9.404399\n", - " 1.037094\n", - " 9.404399\n", + " 399\n", + " ci_upper\n", + " 99\n", + " 0.529435\n", + " 0.537567\n", + " 0.008131\n", + " 0.015359\n", + " 0.533373\n", + " 0.253168\n", + " 0.533373\n", " \n", " \n", "\n", - "

600 rows × 10 columns

\n", + "

400 rows × 9 columns

\n", "" ], "text/plain": [ - " quantity samp Parametric Bootstrap BLB Error RelError \\\n", - "0 ci_lower 0 2.597184 -4.423093 2.593609 -0.003575 0.001377 \n", - "1 ci_lower 1 -12.386757 -4.425957 -12.386300 0.000458 0.000037 \n", - "2 ci_lower 2 13.514136 -8.080191 13.510683 -0.003453 0.000256 \n", - "3 ci_lower 3 9.921847 -25.329503 9.921969 0.000122 0.000012 \n", - "4 ci_lower 4 -0.510517 -2.748500 -0.509992 0.000526 0.001029 \n", - ".. ... ... ... ... ... ... ... \n", - "595 mean 195 1.883268 NaN 1.884418 0.001150 0.000611 \n", - "596 mean 196 -12.321563 NaN -12.321220 0.000343 0.000028 \n", - "597 mean 197 19.500813 NaN 19.500918 0.000105 0.000005 \n", - "598 mean 198 -7.208713 NaN -7.208516 0.000196 0.000027 \n", - "599 mean 199 9.404558 NaN 9.404817 0.000259 0.000028 \n", + " quantity samp Parametric BLB Error RelError RealMean \\\n", + "0 rep_mean 0 0.155556 0.160216 0.004660 0.029956 0.154939 \n", + "1 rep_mean 1 0.103724 0.097968 -0.005757 0.055498 0.083691 \n", + "2 rep_mean 2 0.343596 0.326011 -0.017584 0.051178 0.381152 \n", + "3 rep_mean 3 0.075131 0.069743 -0.005388 0.071712 0.081031 \n", + "4 rep_mean 4 0.072633 0.073120 0.000487 0.006699 0.072287 \n", + ".. ... ... ... ... ... ... ... \n", + "395 ci_upper 95 0.052551 0.090591 0.038041 0.723889 0.071423 \n", + "396 ci_upper 96 0.435598 0.448650 0.013052 0.029963 0.454315 \n", + "397 ci_upper 97 0.119923 0.194622 0.074699 0.622887 0.171009 \n", + "398 ci_upper 98 0.113820 0.136177 0.022357 0.196422 0.128295 \n", + "399 ci_upper 99 0.529435 0.537567 0.008131 0.015359 0.533373 \n", "\n", - " RealMean RealSTD AbsMean \n", - "0 2.601141 1.592588 2.601141 \n", - "1 -12.385795 0.547635 12.385795 \n", - "2 13.514852 1.421265 13.514852 \n", - "3 9.922137 0.140994 9.922137 \n", - "4 -0.509412 0.362041 0.509412 \n", - ".. ... ... ... \n", - "595 1.884801 2.149275 1.884801 \n", - "596 -12.319299 2.509345 12.319299 \n", - "597 19.501153 0.612372 19.501153 \n", - "598 -7.209443 0.213760 7.209443 \n", - "599 9.404399 1.037094 9.404399 \n", + " RealSTD AbsMean \n", + "0 0.123640 0.154939 \n", + "1 2.412370 0.083691 \n", + "2 1.841601 0.381152 \n", + "3 0.337456 0.081031 \n", + "4 0.121747 0.072287 \n", + ".. ... ... \n", + "395 0.622047 0.071423 \n", + "396 0.693925 0.454315 \n", + "397 1.900373 0.171009 \n", + "398 0.687364 0.128295 \n", + "399 0.253168 0.533373 \n", "\n", - "[600 rows x 10 columns]" + "[400 rows x 9 columns]" ] }, - "execution_count": 38, + "execution_count": 30, "metadata": {}, "output_type": "execute_result" } @@ -1097,15 +1040,15 @@ }, { "cell_type": "code", - "execution_count": null, + "execution_count": 31, "id": "34c4fd67", "metadata": {}, "outputs": [ { "data": { - "image/png": "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", + "image/png": "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", "text/plain": [ - "
" + "
" ] }, "metadata": {}, @@ -1119,15 +1062,36 @@ }, { "cell_type": "code", - "execution_count": null, + "execution_count": 19, "id": "4ce6bdd3", "metadata": {}, + "outputs": [], + "source": [ + "# sns.relplot(comb_stats, x=\"Bootstrap\", y=\"BLB\", col=\"quantity\", kind=\"scatter\")\n", + "# plt.show()" + ] + }, + { + "cell_type": "code", + "execution_count": 33, + "id": "d22ac527", + "metadata": {}, "outputs": [ { "data": { - "image/png": "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", "text/plain": [ - "
" + "" + ] + }, + "execution_count": 33, + "metadata": {}, + "output_type": "execute_result" + }, + { + "data": { + "image/png": "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", + "text/plain": [ + "
" ] }, "metadata": {}, @@ -1135,31 +1099,30 @@ } ], "source": [ - "sns.relplot(comb_stats, x=\"Bootstrap\", y=\"BLB\", col=\"quantity\", kind=\"scatter\")\n", - "plt.show()" + "sns.displot(comb_stats, x=\"Error\", col=\"quantity\")" ] }, { "cell_type": "code", - "execution_count": null, - "id": "d22ac527", + "execution_count": 34, + "id": "d9c7b259", "metadata": {}, "outputs": [ { "data": { "text/plain": [ - "" + "" ] }, - "execution_count": 41, + "execution_count": 34, "metadata": {}, "output_type": "execute_result" }, { "data": { - "image/png": "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", + "image/png": "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", "text/plain": [ - "
" + "
" ] }, "metadata": {}, @@ -1167,30 +1130,30 @@ } ], "source": [ - "sns.displot(comb_stats, x=\"Error\", col=\"quantity\")" + "sns.displot(comb_stats, x=\"RelError\", col=\"quantity\")" ] }, { "cell_type": "code", - "execution_count": null, + "execution_count": 35, "id": "9cc95f1d", "metadata": {}, "outputs": [ { "data": { "text/plain": [ - "" + "" ] }, - "execution_count": 42, + "execution_count": 35, "metadata": {}, "output_type": "execute_result" }, { "data": { - "image/png": "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", + "image/png": "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", "text/plain": [ - "
" + "
" ] }, "metadata": {}, @@ -1203,25 +1166,25 @@ }, { "cell_type": "code", - "execution_count": null, + "execution_count": 36, "id": "82008957", "metadata": {}, "outputs": [ { "data": { "text/plain": [ - "" + "" ] }, - "execution_count": 43, + "execution_count": 36, "metadata": {}, "output_type": "execute_result" }, { "data": { - "image/png": "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", + "image/png": "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", "text/plain": [ - "
" + "
" ] }, "metadata": {}, From 06d45f0f8a1d7739f69b23c97c1f0dba225fcb9a Mon Sep 17 00:00:00 2001 From: Michael Ekstrand Date: Fri, 6 Jun 2025 18:38:23 -0400 Subject: [PATCH 30/59] increase default r_window --- src/lenskit/stats/_blb.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/lenskit/stats/_blb.py b/src/lenskit/stats/_blb.py index b4ef8b982..2fa248695 100644 --- a/src/lenskit/stats/_blb.py +++ b/src/lenskit/stats/_blb.py @@ -58,7 +58,7 @@ def blb_summary( b_factor: float = 0.7, rel_tol: float = 0.02, s_window: int = 3, - r_window: int = 100, + r_window: int = 200, rng: RNGInput = None, ) -> dict[str, float]: r""" From 6a0ad57aba9d456edb56ad948489cc3b7f1a158a Mon Sep 17 00:00:00 2001 From: Michael Ekstrand Date: Fri, 6 Jun 2025 18:41:20 -0400 Subject: [PATCH 31/59] CIs misbehave --- tests/stats/test_blb.py | 1 + 1 file changed, 1 insertion(+) diff --git a/tests/stats/test_blb.py b/tests/stats/test_blb.py index 540f31924..67fdbd7be 100644 --- a/tests/stats/test_blb.py +++ b/tests/stats/test_blb.py @@ -38,6 +38,7 @@ def test_blb_single_array(rng: np.random.Generator): @mark.slow +@mark.skip("CIs are not yet behaving correctly") @mark.parametrize("size", [1000, 10000]) @mark.filterwarnings(r"error:.*ignoring \d+ nonfinite values") def test_blb_array_normal(rng: np.random.Generator, size: int): From 67ef1a767dcaef779b14874eaf028baec4470c18 Mon Sep 17 00:00:00 2001 From: Michael Ekstrand Date: Tue, 10 Jun 2025 18:02:08 -0400 Subject: [PATCH 32/59] try better BLB tests and defaults --- src/lenskit/stats/_blb.py | 15 ++++++------ tests/stats/test_blb.py | 49 +++++++++++++++++++++++++++++++-------- 2 files changed, 47 insertions(+), 17 deletions(-) diff --git a/src/lenskit/stats/_blb.py b/src/lenskit/stats/_blb.py index 2fa248695..fdbfa5248 100644 --- a/src/lenskit/stats/_blb.py +++ b/src/lenskit/stats/_blb.py @@ -57,8 +57,8 @@ def blb_summary( ci_width: float = 0.95, b_factor: float = 0.7, rel_tol: float = 0.02, - s_window: int = 3, - r_window: int = 200, + s_window: int = 5, + r_window: int = 50, rng: RNGInput = None, ) -> dict[str, float]: r""" @@ -196,8 +196,8 @@ def run_bootstraps(self, xs: NDArray[F]) -> _BootResult: lbs = StatAccum(np.mean) ubs = StatAccum(np.mean) - self._tracer.trace("estimating acceleration term") - accel = _bca_accel_term(xs, self.config.statistic) + # self._tracer.trace("estimating acceleration term") + # accel = _bca_accel_term(xs, self.config.statistic) self._rep_generator = ReplicateGenerator(n, b, self.rng) self._tracer.trace("let's go!") @@ -206,7 +206,7 @@ def run_bootstraps(self, xs: NDArray[F]) -> _BootResult: for i, ss in enumerate(self.blb_subsets(n, b)): self._tracer.add_bindings(subset=i) self._tracer.trace("starting subset") - res = self.measure_subset(xs, ss, estimate, accel) + res = self.measure_subset(xs, ss, estimate, 0) ss_frames[i] = res.samples means.record(res.rep_mean) vars.record(res.rep_var) @@ -253,8 +253,9 @@ def measure_subset( vars.record(stat) stats = means.values - ql, qh = _bca_range(estimate, stats, self.config.ci_margin, accel) - self._tracer.trace("bias-corrected quantiles: [%.4f, %.4f]", ql, qh, accel=accel) + # ql, qh = _bca_range(estimate, stats, self.config.ci_margin, accel) + # self._tracer.trace("bias-corrected quantiles: [%.4f, %.4f]", ql, qh, accel=accel) + ql, qh = ci_quantiles(self.config.ci_width) lb, ub = np.quantile(stats, [ql, qh]) lbs.record(stat, lb) ubs.record(stat, ub) diff --git a/tests/stats/test_blb.py b/tests/stats/test_blb.py index 67fdbd7be..612f640b5 100644 --- a/tests/stats/test_blb.py +++ b/tests/stats/test_blb.py @@ -16,6 +16,7 @@ from lenskit.data.types import NPVector from lenskit.diagnostics import DataWarning +from lenskit.parallel.ray import ensure_cluster, ray_available from lenskit.random import random_generator from lenskit.stats import blb_summary @@ -38,12 +39,14 @@ def test_blb_single_array(rng: np.random.Generator): @mark.slow -@mark.skip("CIs are not yet behaving correctly") -@mark.parametrize("size", [1000, 10000]) +@mark.skipif(not ray_available(), reason="bulk BLB test requires Ray") +@mark.parametrize("size", [1000, 10_000, 100_000]) @mark.filterwarnings(r"error:.*ignoring \d+ nonfinite values") def test_blb_array_normal(rng: np.random.Generator, size: int): "Test BLB with arrays of normals." + import ray + ensure_cluster() TRUE_MEAN = 1.0 TRUE_SD = 1.0 # TRUE_SVAR = TRUE_SD * TRUE_SD / size @@ -52,26 +55,52 @@ def test_blb_array_normal(rng: np.random.Generator, size: int): # Test: for 1000 runs, do approx. 95% of confidence intervals contain the # true mean? - NTRIALS = 200 - for i in range(NTRIALS): - xs = rng.normal(TRUE_MEAN, TRUE_SD, size) - mean = np.mean(xs) + worker = ray.remote(num_cpus=1)(_blb_worker) - summary = blb_summary(xs, "mean", rng=rng) - assert isinstance(summary, dict) - assert summary["estimate"] == approx(mean) + NBATCHES = 20 + PERBATCH = 50 + NTRIALS = NBATCHES * PERBATCH + rngs = rng.spawn(NBATCHES) + tasks = [worker.remote(PERBATCH, TRUE_MEAN, TRUE_SD, size, t) for t in rngs] + for task in tasks: + bres = ray.get(task) + for mean, summary in bres: + assert isinstance(summary, dict) + assert summary["estimate"] == approx(mean) - results.append(summary) + results.append(summary) n_lb_good = len([r for r in results if r["ci_lower"] <= TRUE_MEAN]) pct_lb_good = (n_lb_good / NTRIALS) * 100 n_ub_good = len([r for r in results if TRUE_MEAN <= r["ci_upper"]]) pct_ub_good = (n_ub_good / NTRIALS) * 100 + n_good = len([r for r in results if r["ci_lower"] <= TRUE_MEAN <= r["ci_upper"]]) + pct_good = (n_good / NTRIALS) * 100 + print( + "{:.1f}% CIs good ({:1f}% LB fail, {:.1f}% UB fail)".format( + pct_good, 100 - pct_lb_good, 100 - pct_ub_good + ) + ) # leave a little wiggle room assert 90 <= pct_lb_good <= 99 assert 90 <= pct_ub_good <= 99 +def _blb_worker( + nreps: int, true_mean: float, true_sd: float, size: int, rng: np.random.Generator +) -> list[tuple[float, dict[str, float]]]: + results = [] + bf = 0.7 if size > 50_000 else 0.8 + + for _i in range(nreps): + xs = rng.normal(true_mean, true_sd, size) + mean = np.mean(xs).item() + + results.append((mean, blb_summary(xs, "mean", rng=rng, b_factor=bf))) + + return results + + @mark.skip("need to find better parameters") @given( nph.arrays(shape=st.integers(10000, 1_000_000), dtype=nph.floating_dtypes(endianness="=")), From ca3648743a17069ca436cb4cb492cdeef8fe3259 Mon Sep 17 00:00:00 2001 From: Michael Ekstrand Date: Tue, 10 Jun 2025 18:02:39 -0400 Subject: [PATCH 33/59] update BLB test --- tests/stats/test_blb.py | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/tests/stats/test_blb.py b/tests/stats/test_blb.py index 612f640b5..97ffae3a7 100644 --- a/tests/stats/test_blb.py +++ b/tests/stats/test_blb.py @@ -82,8 +82,7 @@ def test_blb_array_normal(rng: np.random.Generator, size: int): ) ) # leave a little wiggle room - assert 90 <= pct_lb_good <= 99 - assert 90 <= pct_ub_good <= 99 + assert 90 <= pct_good <= 98 def _blb_worker( From 12f868eb25dd0e67fc7da4007b43ce3b9611c484 Mon Sep 17 00:00:00 2001 From: Michael Ekstrand Date: Thu, 12 Jun 2025 12:25:28 -0400 Subject: [PATCH 34/59] adjust defaults and test --- src/lenskit/stats/_blb.py | 2 +- tests/stats/test_blb.py | 4 ++-- 2 files changed, 3 insertions(+), 3 deletions(-) diff --git a/src/lenskit/stats/_blb.py b/src/lenskit/stats/_blb.py index fdbfa5248..f14093538 100644 --- a/src/lenskit/stats/_blb.py +++ b/src/lenskit/stats/_blb.py @@ -57,7 +57,7 @@ def blb_summary( ci_width: float = 0.95, b_factor: float = 0.7, rel_tol: float = 0.02, - s_window: int = 5, + s_window: int = 10, r_window: int = 50, rng: RNGInput = None, ) -> dict[str, float]: diff --git a/tests/stats/test_blb.py b/tests/stats/test_blb.py index 97ffae3a7..7a108e0b0 100644 --- a/tests/stats/test_blb.py +++ b/tests/stats/test_blb.py @@ -89,13 +89,13 @@ def _blb_worker( nreps: int, true_mean: float, true_sd: float, size: int, rng: np.random.Generator ) -> list[tuple[float, dict[str, float]]]: results = [] - bf = 0.7 if size > 50_000 else 0.8 + # bf = 0.7 if size > 50_000 else 0.8 for _i in range(nreps): xs = rng.normal(true_mean, true_sd, size) mean = np.mean(xs).item() - results.append((mean, blb_summary(xs, "mean", rng=rng, b_factor=bf))) + results.append((mean, blb_summary(xs, "mean", rng=rng))) return results From 38e846adc2ebc16b61ac0082a2a7614495be8ecb Mon Sep 17 00:00:00 2001 From: Michael Ekstrand Date: Tue, 17 Jun 2025 14:27:22 -0400 Subject: [PATCH 35/59] fix parallel logging --- src/lenskit/logging/multiprocess/_worker.py | 2 +- src/lenskit/parallel/ray.py | 6 ++++-- 2 files changed, 5 insertions(+), 3 deletions(-) diff --git a/src/lenskit/logging/multiprocess/_worker.py b/src/lenskit/logging/multiprocess/_worker.py index 083e00551..6869c84bd 100644 --- a/src/lenskit/logging/multiprocess/_worker.py +++ b/src/lenskit/logging/multiprocess/_worker.py @@ -80,7 +80,7 @@ def current(cls, *, from_monitor: bool = True): if mon.log_address is None: raise RuntimeError("monitor has no log address") cfg = active_logging_config() - level = cfg.effective_level if cfg is not None else logging.INFO + level = cfg.effective_level if cfg is not None else logging.DEBUG return cls( address=mon.log_address, level=level, authkey=bytes(mp.current_process().authkey) ) diff --git a/src/lenskit/parallel/ray.py b/src/lenskit/parallel/ray.py index 60c5b0429..799560f45 100644 --- a/src/lenskit/parallel/ray.py +++ b/src/lenskit/parallel/ray.py @@ -71,7 +71,7 @@ def init_cluster( proc_slots: int | None = None, resources: dict[str, float] | None = None, worker_parallel: ParallelConfig | None = None, - global_logging: bool = False, + global_logging: bool = True, **kwargs, ): """ @@ -130,7 +130,7 @@ def init_cluster( setup = _worker_setup if global_logging else None runtime = ray.runtime_env.RuntimeEnv(env_vars=env, worker_process_setup_hook=setup) - _log.info("starting Ray cluster") + _log.info("starting Ray cluster", logging=global_logging) ray.init(num_cpus=num_cpus, resources=resources, runtime_env=runtime, **kwargs) @@ -422,5 +422,7 @@ def init_worker(*, autostart: bool = True) -> WorkerContext: if autostart: context.start() + _log.debug("worker context initialized") + ensure_parallel_init() return context From ac4f9531b561cd5a92c25dd42a69b37887a173ac Mon Sep 17 00:00:00 2001 From: Michael Ekstrand Date: Tue, 17 Jun 2025 14:27:28 -0400 Subject: [PATCH 36/59] tweak BLB --- src/lenskit/stats/_blb.py | 5 +++++ 1 file changed, 5 insertions(+) diff --git a/src/lenskit/stats/_blb.py b/src/lenskit/stats/_blb.py index f14093538..053139d0f 100644 --- a/src/lenskit/stats/_blb.py +++ b/src/lenskit/stats/_blb.py @@ -257,6 +257,11 @@ def measure_subset( # self._tracer.trace("bias-corrected quantiles: [%.4f, %.4f]", ql, qh, accel=accel) ql, qh = ci_quantiles(self.config.ci_width) lb, ub = np.quantile(stats, [ql, qh]) + self._tracer.trace("initial bounds: %f < s < %f", lb, ub) + # recenter bounds around estimate + lb = estimate - (stat - lb) + ub = estimate + (ub - stat) + self._tracer.trace("adjusted bounds: %f < s < %f", lb, ub) lbs.record(stat, lb) ubs.record(stat, ub) del stats From 0145985b961b0244b856fbff3f51446795862bea Mon Sep 17 00:00:00 2001 From: Michael Ekstrand Date: Wed, 13 Aug 2025 13:35:51 -0400 Subject: [PATCH 37/59] dial back resource requirement --- tests/stats/test_blb.py | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/tests/stats/test_blb.py b/tests/stats/test_blb.py index 7a108e0b0..218b04f4a 100644 --- a/tests/stats/test_blb.py +++ b/tests/stats/test_blb.py @@ -55,7 +55,7 @@ def test_blb_array_normal(rng: np.random.Generator, size: int): # Test: for 1000 runs, do approx. 95% of confidence intervals contain the # true mean? - worker = ray.remote(num_cpus=1)(_blb_worker) + worker = ray.remote(num_cpus=2)(_blb_worker) NBATCHES = 20 PERBATCH = 50 @@ -95,7 +95,7 @@ def _blb_worker( xs = rng.normal(true_mean, true_sd, size) mean = np.mean(xs).item() - results.append((mean, blb_summary(xs, "mean", rng=rng))) + results.append((mean, blb_summary(xs, "mean", rng=rng, b_factor=0.6))) return results From a15a344f741170d74dce8b1fecb30c7ce90ee64e Mon Sep 17 00:00:00 2001 From: Michael Ekstrand Date: Wed, 13 Aug 2025 13:38:12 -0400 Subject: [PATCH 38/59] dial back resource more --- tests/stats/test_blb.py | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/tests/stats/test_blb.py b/tests/stats/test_blb.py index 218b04f4a..8dbf7d37f 100644 --- a/tests/stats/test_blb.py +++ b/tests/stats/test_blb.py @@ -95,7 +95,9 @@ def _blb_worker( xs = rng.normal(true_mean, true_sd, size) mean = np.mean(xs).item() - results.append((mean, blb_summary(xs, "mean", rng=rng, b_factor=0.6))) + results.append( + (mean, blb_summary(xs, "mean", rng=rng, b_factor=0.6, s_window=5, r_window=20)) + ) return results From 2a78cd410b5ea11194002a7cf203c5d7868ee083 Mon Sep 17 00:00:00 2001 From: Michael Ekstrand Date: Wed, 13 Aug 2025 17:00:50 -0400 Subject: [PATCH 39/59] fix BLB typo + quantile correction bug --- src/lenskit/stats/_blb.py | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-) diff --git a/src/lenskit/stats/_blb.py b/src/lenskit/stats/_blb.py index 053139d0f..908b91824 100644 --- a/src/lenskit/stats/_blb.py +++ b/src/lenskit/stats/_blb.py @@ -77,7 +77,7 @@ def blb_summary( statistics to support weighted computation (this is what allows it to speed up the bootstrap procedure). ci_width: - The width of the confidence interval to estimat.e + The width of the confidence interval to estimate. b_factor: The shrinking factor :math:`\gamma` to use to derive subsample sizes. Each subsample has size :math:`N^{\gamma}`. @@ -259,8 +259,9 @@ def measure_subset( lb, ub = np.quantile(stats, [ql, qh]) self._tracer.trace("initial bounds: %f < s < %f", lb, ub) # recenter bounds around estimate - lb = estimate - (stat - lb) - ub = estimate + (ub - stat) + ec = means.statistic + lb = estimate - (ec - lb) + ub = estimate + (ub - ec) self._tracer.trace("adjusted bounds: %f < s < %f", lb, ub) lbs.record(stat, lb) ubs.record(stat, ub) From 39ce2ad2634d3d1e381bd1a21de37a47f26b4534 Mon Sep 17 00:00:00 2001 From: Michael Ekstrand Date: Wed, 13 Aug 2025 17:01:24 -0400 Subject: [PATCH 40/59] tweak BLB test parameters + re-randomize --- tests/stats/test_blb.py | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) diff --git a/tests/stats/test_blb.py b/tests/stats/test_blb.py index 8dbf7d37f..7a1b3fac3 100644 --- a/tests/stats/test_blb.py +++ b/tests/stats/test_blb.py @@ -42,10 +42,12 @@ def test_blb_single_array(rng: np.random.Generator): @mark.skipif(not ray_available(), reason="bulk BLB test requires Ray") @mark.parametrize("size", [1000, 10_000, 100_000]) @mark.filterwarnings(r"error:.*ignoring \d+ nonfinite values") -def test_blb_array_normal(rng: np.random.Generator, size: int): +def test_blb_array_normal(size: int): "Test BLB with arrays of normals." import ray + rng = np.random.default_rng() + ensure_cluster() TRUE_MEAN = 1.0 TRUE_SD = 1.0 @@ -96,7 +98,7 @@ def _blb_worker( mean = np.mean(xs).item() results.append( - (mean, blb_summary(xs, "mean", rng=rng, b_factor=0.6, s_window=5, r_window=20)) + (mean, blb_summary(xs, "mean", rng=rng, b_factor=0.7, s_window=20, r_window=40)) ) return results From dfbb152ea69f3e45a9f645e2bb140a76db1c4c6c Mon Sep 17 00:00:00 2001 From: Michael Ekstrand Date: Wed, 13 Aug 2025 17:45:15 -0400 Subject: [PATCH 41/59] clean up some testing --- tests/stats/test_blb.py | 30 +++++++++++++++++------------- 1 file changed, 17 insertions(+), 13 deletions(-) diff --git a/tests/stats/test_blb.py b/tests/stats/test_blb.py index 7a1b3fac3..dd915b31e 100644 --- a/tests/stats/test_blb.py +++ b/tests/stats/test_blb.py @@ -8,6 +8,7 @@ import numpy as np from numpy.typing import NDArray +from scipy.stats import binomtest import hypothesis.extra.numpy as nph import hypothesis.strategies as st @@ -16,10 +17,13 @@ from lenskit.data.types import NPVector from lenskit.diagnostics import DataWarning +from lenskit.logging import get_logger from lenskit.parallel.ray import ensure_cluster, ray_available from lenskit.random import random_generator from lenskit.stats import blb_summary +_log = get_logger(__name__) + def test_blb_single_array(rng: np.random.Generator): "Quick one-array test to fail fast" @@ -42,12 +46,10 @@ def test_blb_single_array(rng: np.random.Generator): @mark.skipif(not ray_available(), reason="bulk BLB test requires Ray") @mark.parametrize("size", [1000, 10_000, 100_000]) @mark.filterwarnings(r"error:.*ignoring \d+ nonfinite values") -def test_blb_array_normal(size: int): +def test_blb_array_normal(rng: np.random.Generator, size: int): "Test BLB with arrays of normals." import ray - rng = np.random.default_rng() - ensure_cluster() TRUE_MEAN = 1.0 TRUE_SD = 1.0 @@ -73,18 +75,20 @@ def test_blb_array_normal(size: int): results.append(summary) n_lb_good = len([r for r in results if r["ci_lower"] <= TRUE_MEAN]) - pct_lb_good = (n_lb_good / NTRIALS) * 100 + f_lb_good = n_lb_good / NTRIALS n_ub_good = len([r for r in results if TRUE_MEAN <= r["ci_upper"]]) - pct_ub_good = (n_ub_good / NTRIALS) * 100 + f_ub_good = n_ub_good / NTRIALS n_good = len([r for r in results if r["ci_lower"] <= TRUE_MEAN <= r["ci_upper"]]) - pct_good = (n_good / NTRIALS) * 100 - print( - "{:.1f}% CIs good ({:1f}% LB fail, {:.1f}% UB fail)".format( - pct_good, 100 - pct_lb_good, 100 - pct_ub_good - ) + f_good = n_good / NTRIALS + bt = binomtest(n_good, NTRIALS, 0.95) + _log.info( + "{:.1%} CIs good ({:1%} LB fail, {:.1%} UB fail), p={:.4f}".format( + f_good, 1 - f_lb_good, 1 - f_ub_good, bt.pvalue + ), + test=bt, ) - # leave a little wiggle room - assert 90 <= pct_good <= 98 + # leave some wiggle room + assert bt.pvalue >= 0.05 def _blb_worker( @@ -98,7 +102,7 @@ def _blb_worker( mean = np.mean(xs).item() results.append( - (mean, blb_summary(xs, "mean", rng=rng, b_factor=0.7, s_window=20, r_window=40)) + (mean, blb_summary(xs, "mean", rng=rng, b_factor=0.7, s_window=10, r_window=20)) ) return results From a0ce38dcc1f9df1d785c50e72bcab2ed555f523d Mon Sep 17 00:00:00 2001 From: Michael Ekstrand Date: Wed, 13 Aug 2025 18:28:08 -0400 Subject: [PATCH 42/59] use the expanded percentile --- src/lenskit/stats/_blb.py | 16 ++++++++-------- src/lenskit/stats/_distributions.py | 20 +++++++++++++++++++- tests/stats/test_ci_utils.py | 8 ++++++++ 3 files changed, 35 insertions(+), 9 deletions(-) diff --git a/src/lenskit/stats/_blb.py b/src/lenskit/stats/_blb.py index 908b91824..fb544e947 100644 --- a/src/lenskit/stats/_blb.py +++ b/src/lenskit/stats/_blb.py @@ -178,11 +178,11 @@ def __init__(self, config, rng: np.random.Generator): self.rng = rng self.ss_stats = {} - self._ci_qmin, self._ci_qmax = ci_quantiles(config.ci_width) self._tracer = get_tracer(_log, stat=config.statistic.__name__) # type: ignore def run_bootstraps(self, xs: NDArray[F]) -> _BootResult: n = len(xs) + self._ci_qmin, self._ci_qmax = ci_quantiles(self.config.ci_width, expand=n) b = int(n**self.config.b_factor) self._tracer.add_bindings(n=n, b=b) @@ -255,14 +255,14 @@ def measure_subset( stats = means.values # ql, qh = _bca_range(estimate, stats, self.config.ci_margin, accel) # self._tracer.trace("bias-corrected quantiles: [%.4f, %.4f]", ql, qh, accel=accel) - ql, qh = ci_quantiles(self.config.ci_width) - lb, ub = np.quantile(stats, [ql, qh]) - self._tracer.trace("initial bounds: %f < s < %f", lb, ub) + lb, ub = np.quantile(stats, [self._ci_qmin, self._ci_qmax]) + self._tracer.trace("expanded bounds: %f < s < %f", lb, ub) # recenter bounds around estimate - ec = means.statistic - lb = estimate - (ec - lb) - ub = estimate + (ub - ec) - self._tracer.trace("adjusted bounds: %f < s < %f", lb, ub) + # this is the reverse-bootstrap percentile interval + # see: https://arxiv.org/pdf/1411.5279 + # ec = means.statistic + # lb = estimate - (ec - lb) + # ub = estimate + (ub - ec) lbs.record(stat, lb) ubs.record(stat, ub) del stats diff --git a/src/lenskit/stats/_distributions.py b/src/lenskit/stats/_distributions.py index ed2e36df0..f9dab3453 100644 --- a/src/lenskit/stats/_distributions.py +++ b/src/lenskit/stats/_distributions.py @@ -10,15 +10,33 @@ from typing import Annotated +import numpy as np from annotated_types import Gt, Lt from pydantic import validate_call +from scipy import stats @validate_call -def ci_quantiles(width: Annotated[float, Gt(0), Lt(1)]) -> tuple[float, float]: +def ci_quantiles( + width: Annotated[float, Gt(0), Lt(1)], *, expand: Annotated[int, Gt(1)] | None = None +) -> tuple[float, float]: r""" Convert a confidence interval width to CI quantile bounds. + + Args: + width: + The CI interval width. + expand: + If not ``None``, a sample size :math:`n` to use to + expand the CI as in the expanded percentile bootstrap. """ margin = 0.5 * (1 - width) + if expand: + factor = np.sqrt(expand / (expand - 1)) + # get t_(alpha/2),n-1 + t = stats.t.ppf(margin, expand - 1) + # get standard normal CDF + margin = stats.norm.cdf(factor * t) + return margin, 1 - margin diff --git a/tests/stats/test_ci_utils.py b/tests/stats/test_ci_utils.py index 3318d135c..420c86107 100644 --- a/tests/stats/test_ci_utils.py +++ b/tests/stats/test_ci_utils.py @@ -20,3 +20,11 @@ def test_ci_bounds(width: float): qlo, qhi = ci_quantiles(width) assert qhi - qlo == approx(width) assert 1 - qhi == approx(qlo) + + +@given(st.floats(0.1, 0.9)) +def test_ci_bounds_expanded(width: float): + oql, oqh = ci_quantiles(width) + qlo, qhi = ci_quantiles(width, expand=500) + assert qlo < oql + assert qhi > oqh From 0bc70a3478485b31686a383145bc4853b608bde0 Mon Sep 17 00:00:00 2001 From: Michael Ekstrand Date: Wed, 13 Aug 2025 18:28:20 -0400 Subject: [PATCH 43/59] test tweaks --- tests/stats/test_blb.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tests/stats/test_blb.py b/tests/stats/test_blb.py index dd915b31e..822793fdc 100644 --- a/tests/stats/test_blb.py +++ b/tests/stats/test_blb.py @@ -82,7 +82,7 @@ def test_blb_array_normal(rng: np.random.Generator, size: int): f_good = n_good / NTRIALS bt = binomtest(n_good, NTRIALS, 0.95) _log.info( - "{:.1%} CIs good ({:1%} LB fail, {:.1%} UB fail), p={:.4f}".format( + "{:.1%} CIs good ({:1%} LB fail, {:.1%} UB fail), p={:.3g}".format( f_good, 1 - f_lb_good, 1 - f_ub_good, bt.pvalue ), test=bt, From a71c6c7f62e80779049e67397b229e3677f577fb Mon Sep 17 00:00:00 2001 From: Michael Ekstrand Date: Wed, 13 Aug 2025 18:44:03 -0400 Subject: [PATCH 44/59] more test tweaking --- tests/stats/test_blb.py | 11 ++++++++--- 1 file changed, 8 insertions(+), 3 deletions(-) diff --git a/tests/stats/test_blb.py b/tests/stats/test_blb.py index 822793fdc..03db3c284 100644 --- a/tests/stats/test_blb.py +++ b/tests/stats/test_blb.py @@ -56,8 +56,8 @@ def test_blb_array_normal(rng: np.random.Generator, size: int): # TRUE_SVAR = TRUE_SD * TRUE_SD / size results = [] - # Test: for 1000 runs, do approx. 95% of confidence intervals contain the - # true mean? + # Test: for NBATCHES * PERBATCH runs, do approx. 95% of confidence intervals + # contain the true mean? worker = ray.remote(num_cpus=2)(_blb_worker) @@ -102,7 +102,12 @@ def _blb_worker( mean = np.mean(xs).item() results.append( - (mean, blb_summary(xs, "mean", rng=rng, b_factor=0.7, s_window=10, r_window=20)) + ( + mean, + blb_summary( + xs, "mean", rng=rng, b_factor=0.7, s_window=10, r_window=20, rel_tol=0.01 + ), + ) ) return results From 15234dc5eeeb175948458b563ad993bfbb817547 Mon Sep 17 00:00:00 2001 From: Michael Ekstrand Date: Thu, 14 Aug 2025 11:24:20 -0400 Subject: [PATCH 45/59] instrumentation and configurability --- tests/stats/test_blb.py | 24 ++++++++++++------------ 1 file changed, 12 insertions(+), 12 deletions(-) diff --git a/tests/stats/test_blb.py b/tests/stats/test_blb.py index 03db3c284..1ecb3b0e4 100644 --- a/tests/stats/test_blb.py +++ b/tests/stats/test_blb.py @@ -4,6 +4,7 @@ # Licensed under the MIT license, see LICENSE.md for details. # SPDX-License-Identifier: MIT +import os from math import sqrt import numpy as np @@ -17,7 +18,7 @@ from lenskit.data.types import NPVector from lenskit.diagnostics import DataWarning -from lenskit.logging import get_logger +from lenskit.logging import Stopwatch, get_logger from lenskit.parallel.ray import ensure_cluster, ray_available from lenskit.random import random_generator from lenskit.stats import blb_summary @@ -55,6 +56,7 @@ def test_blb_array_normal(rng: np.random.Generator, size: int): TRUE_SD = 1.0 # TRUE_SVAR = TRUE_SD * TRUE_SD / size results = [] + times = [] # Test: for NBATCHES * PERBATCH runs, do approx. 95% of confidence intervals # contain the true mean? @@ -62,18 +64,20 @@ def test_blb_array_normal(rng: np.random.Generator, size: int): worker = ray.remote(num_cpus=2)(_blb_worker) NBATCHES = 20 - PERBATCH = 50 + PERBATCH = int(os.environ.get("BLB_TRIALS_PER_BATCH", 50)) NTRIALS = NBATCHES * PERBATCH rngs = rng.spawn(NBATCHES) tasks = [worker.remote(PERBATCH, TRUE_MEAN, TRUE_SD, size, t) for t in rngs] for task in tasks: bres = ray.get(task) - for mean, summary in bres: + for mean, summary, time in bres: assert isinstance(summary, dict) assert summary["estimate"] == approx(mean) results.append(summary) + times.append(time) + _log.info("completed %d trials (avg %.2fms / trial)", len(results), np.mean(times) * 1000) n_lb_good = len([r for r in results if r["ci_lower"] <= TRUE_MEAN]) f_lb_good = n_lb_good / NTRIALS n_ub_good = len([r for r in results if TRUE_MEAN <= r["ci_upper"]]) @@ -93,7 +97,7 @@ def test_blb_array_normal(rng: np.random.Generator, size: int): def _blb_worker( nreps: int, true_mean: float, true_sd: float, size: int, rng: np.random.Generator -) -> list[tuple[float, dict[str, float]]]: +) -> list[tuple[float, dict[str, float], float]]: results = [] # bf = 0.7 if size > 50_000 else 0.8 @@ -101,14 +105,10 @@ def _blb_worker( xs = rng.normal(true_mean, true_sd, size) mean = np.mean(xs).item() - results.append( - ( - mean, - blb_summary( - xs, "mean", rng=rng, b_factor=0.7, s_window=10, r_window=20, rel_tol=0.01 - ), - ) - ) + timer = Stopwatch() + s = blb_summary(xs, "mean", rng=rng, b_factor=0.75, s_window=20, r_window=50, rel_tol=0.01) + + results.append((mean, s, timer.elapsed())) return results From 887e1bbd7681c23c6b5026ec500b2a8aa2ecd75d Mon Sep 17 00:00:00 2001 From: Michael Ekstrand Date: Thu, 14 Aug 2025 11:32:29 -0400 Subject: [PATCH 46/59] log CI centers and widths --- tests/stats/test_blb.py | 14 ++++++++++++-- 1 file changed, 12 insertions(+), 2 deletions(-) diff --git a/tests/stats/test_blb.py b/tests/stats/test_blb.py index 1ecb3b0e4..d3b097fa5 100644 --- a/tests/stats/test_blb.py +++ b/tests/stats/test_blb.py @@ -9,7 +9,7 @@ import numpy as np from numpy.typing import NDArray -from scipy.stats import binomtest +from scipy.stats import binomtest, ttest_1samp import hypothesis.extra.numpy as nph import hypothesis.strategies as st @@ -55,6 +55,7 @@ def test_blb_array_normal(rng: np.random.Generator, size: int): TRUE_MEAN = 1.0 TRUE_SD = 1.0 # TRUE_SVAR = TRUE_SD * TRUE_SD / size + THEORETICAL_SE = TRUE_SD / np.sqrt(size) results = [] times = [] @@ -85,11 +86,20 @@ def test_blb_array_normal(rng: np.random.Generator, size: int): n_good = len([r for r in results if r["ci_lower"] <= TRUE_MEAN <= r["ci_upper"]]) f_good = n_good / NTRIALS bt = binomtest(n_good, NTRIALS, 0.95) + _log.info("binomal test for CI hit rate: stat=%.3f, p=%.3g", bt.statistic, bt.pvalue, test=bt) + + rmeans = np.array([r["rep_mean"] for r in results]) + rmt = ttest_1samp(rmeans, TRUE_MEAN) + _log.info("t-test for CI centers: stat=%.5f, p=%.3g", rmt.statistic, rmt.pvalue, test=rmt) + + widths = np.array([r["ci_upper"] - r["ci_lower"] for r in results]) + wt = ttest_1samp(widths, 2 * 1.96 * THEORETICAL_SE) + _log.info("t-test for CI width: stat=%.5f, p=%.3g", wt.statistic, wt.pvalue, test=wt) + _log.info( "{:.1%} CIs good ({:1%} LB fail, {:.1%} UB fail), p={:.3g}".format( f_good, 1 - f_lb_good, 1 - f_ub_good, bt.pvalue ), - test=bt, ) # leave some wiggle room assert bt.pvalue >= 0.05 From 4004459b042b1b2deb81e04d294f60d64fc4c137 Mon Sep 17 00:00:00 2001 From: Michael Ekstrand Date: Thu, 14 Aug 2025 12:28:52 -0400 Subject: [PATCH 47/59] readability cleanup --- src/lenskit/stats/_blb.py | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) diff --git a/src/lenskit/stats/_blb.py b/src/lenskit/stats/_blb.py index fb544e947..4be501992 100644 --- a/src/lenskit/stats/_blb.py +++ b/src/lenskit/stats/_blb.py @@ -98,9 +98,11 @@ def blb_summary( if stat != "mean": raise ValueError(f"unsupported statistic {stat}") + n = len(xs) mask = np.isfinite(xs) - if ninf := int(np.sum(~mask)): - warnings.warn(f"ignoring {ninf} nonfinite values", DataWarning, stacklevel=2) + nfinite = np.sum(mask) + if nfinite < n: + warnings.warn(f"ignoring {n - nfinite} nonfinite values", DataWarning, stacklevel=2) xs = xs[mask] est = np.average(xs).item() From d34ea14381a7eff95d0f73fb0581dd5cb0e18036 Mon Sep 17 00:00:00 2001 From: Michael Ekstrand Date: Thu, 14 Aug 2025 15:01:08 -0400 Subject: [PATCH 48/59] fix up logging etc. --- tests/stats/test_blb.py | 44 ++++++++++++++++++++++++++++++++--------- 1 file changed, 35 insertions(+), 9 deletions(-) diff --git a/tests/stats/test_blb.py b/tests/stats/test_blb.py index d3b097fa5..df6f8f098 100644 --- a/tests/stats/test_blb.py +++ b/tests/stats/test_blb.py @@ -9,7 +9,7 @@ import numpy as np from numpy.typing import NDArray -from scipy.stats import binomtest, ttest_1samp +from scipy.stats import binomtest, describe, ttest_1samp, ttest_rel import hypothesis.extra.numpy as nph import hypothesis.strategies as st @@ -56,6 +56,7 @@ def test_blb_array_normal(rng: np.random.Generator, size: int): TRUE_SD = 1.0 # TRUE_SVAR = TRUE_SD * TRUE_SD / size THEORETICAL_SE = TRUE_SD / np.sqrt(size) + THEORETICAL_WIDTH = 2 * 1.96 * THEORETICAL_SE results = [] times = [] @@ -88,19 +89,44 @@ def test_blb_array_normal(rng: np.random.Generator, size: int): bt = binomtest(n_good, NTRIALS, 0.95) _log.info("binomal test for CI hit rate: stat=%.3f, p=%.3g", bt.statistic, bt.pvalue, test=bt) + smeans = np.array([r["estimate"] for r in results]) + smt = ttest_1samp(smeans, TRUE_MEAN) + _log.info("sample means: %s", describe(smeans)) + if smt.pvalue >= 0.05: + _log.info("t-test for sample means: stat=%.5f, p=%.3g", smt.statistic, smt.pvalue, test=smt) + else: + _log.warn("t-test for sample means: stat=%.5f, p=%.3g", smt.statistic, smt.pvalue, test=smt) rmeans = np.array([r["rep_mean"] for r in results]) - rmt = ttest_1samp(rmeans, TRUE_MEAN) - _log.info("t-test for CI centers: stat=%.5f, p=%.3g", rmt.statistic, rmt.pvalue, test=rmt) + rmt = ttest_rel(rmeans, smeans) + _log.info("bootstrap means: %s", describe(rmeans)) + if rmt.pvalue >= 0.05: + _log.info("t-test for CI centers: stat=%.5f, p=%.3g", rmt.statistic, rmt.pvalue, test=rmt) + else: + _log.warn("t-test for CI centers: stat=%.5f, p=%.3g", rmt.statistic, rmt.pvalue, test=rmt) widths = np.array([r["ci_upper"] - r["ci_lower"] for r in results]) - wt = ttest_1samp(widths, 2 * 1.96 * THEORETICAL_SE) - _log.info("t-test for CI width: stat=%.5f, p=%.3g", wt.statistic, wt.pvalue, test=wt) - _log.info( - "{:.1%} CIs good ({:1%} LB fail, {:.1%} UB fail), p={:.3g}".format( - f_good, 1 - f_lb_good, 1 - f_ub_good, bt.pvalue - ), + "bootstrap CI widths (expected: {:.4f}): {}".format(THEORETICAL_WIDTH, describe(widths)) ) + wt = ttest_1samp(widths, THEORETICAL_WIDTH) + if wt.pvalue >= 0.05: + _log.info("t-test for CI width: stat=%.5f, p=%.3g", wt.statistic, wt.pvalue, test=wt) + else: + _log.warn("t-test for CI width: stat=%.5f, p=%.3g", wt.statistic, wt.pvalue, test=wt) + + if bt.pvalue >= 0.05: + _log.info( + "{:.1%} CIs good ({:1%} LB fail, {:.1%} UB fail), p={:.3g}".format( + f_good, 1 - f_lb_good, 1 - f_ub_good, bt.pvalue + ), + ) + else: + _log.error( + "{:.1%} CIs good ({:1%} LB fail, {:.1%} UB fail), p={:.3g}".format( + f_good, 1 - f_lb_good, 1 - f_ub_good, bt.pvalue + ), + ) + # leave some wiggle room assert bt.pvalue >= 0.05 From 4272277bfd58b13d7d1f89ab7132c446a8245827 Mon Sep 17 00:00:00 2001 From: Michael Ekstrand Date: Thu, 14 Aug 2025 15:01:14 -0400 Subject: [PATCH 49/59] fix result docs --- src/lenskit/stats/_blb.py | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/src/lenskit/stats/_blb.py b/src/lenskit/stats/_blb.py index 4be501992..ea681f6d2 100644 --- a/src/lenskit/stats/_blb.py +++ b/src/lenskit/stats/_blb.py @@ -137,9 +137,9 @@ class _BootResult: "Statistic computed on original data." rep_mean: float - "Mean of the replicates." + "Mean of the statistic computed on the replicates." rep_var: float - "Variance of the replicates." + "Variance of the statistic computed on the replicates." ci_lower: float "CI lower bound." ci_upper: float From fb769325972030b9c7f71f933d22b62cd17ed05f Mon Sep 17 00:00:00 2001 From: Michael Ekstrand Date: Thu, 14 Aug 2025 15:32:50 -0400 Subject: [PATCH 50/59] some blb update --- src/lenskit/stats/_blb.py | 15 +++++---------- 1 file changed, 5 insertions(+), 10 deletions(-) diff --git a/src/lenskit/stats/_blb.py b/src/lenskit/stats/_blb.py index ea681f6d2..7dc3f798c 100644 --- a/src/lenskit/stats/_blb.py +++ b/src/lenskit/stats/_blb.py @@ -198,8 +198,8 @@ def run_bootstraps(self, xs: NDArray[F]) -> _BootResult: lbs = StatAccum(np.mean) ubs = StatAccum(np.mean) - # self._tracer.trace("estimating acceleration term") - # accel = _bca_accel_term(xs, self.config.statistic) + self._tracer.trace("estimating acceleration term") + accel = _bca_accel_term(xs, self.config.statistic) self._rep_generator = ReplicateGenerator(n, b, self.rng) self._tracer.trace("let's go!") @@ -208,7 +208,7 @@ def run_bootstraps(self, xs: NDArray[F]) -> _BootResult: for i, ss in enumerate(self.blb_subsets(n, b)): self._tracer.add_bindings(subset=i) self._tracer.trace("starting subset") - res = self.measure_subset(xs, ss, estimate, 0) + res = self.measure_subset(xs, ss, estimate, accel) ss_frames[i] = res.samples means.record(res.rep_mean) vars.record(res.rep_var) @@ -258,13 +258,8 @@ def measure_subset( # ql, qh = _bca_range(estimate, stats, self.config.ci_margin, accel) # self._tracer.trace("bias-corrected quantiles: [%.4f, %.4f]", ql, qh, accel=accel) lb, ub = np.quantile(stats, [self._ci_qmin, self._ci_qmax]) - self._tracer.trace("expanded bounds: %f < s < %f", lb, ub) - # recenter bounds around estimate - # this is the reverse-bootstrap percentile interval - # see: https://arxiv.org/pdf/1411.5279 - # ec = means.statistic - # lb = estimate - (ec - lb) - # ub = estimate + (ub - ec) + # lb, ub = np.quantile(stats, [ql, qh]) + self._tracer.trace("CI bounds: %f < s < %f", lb, ub) lbs.record(stat, lb) ubs.record(stat, ub) del stats From 5cf1880cad5aab085ba22378562941cbcf56bdba Mon Sep 17 00:00:00 2001 From: Michael Ekstrand Date: Thu, 14 Aug 2025 16:01:14 -0400 Subject: [PATCH 51/59] test tweaks --- tests/stats/test_blb.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tests/stats/test_blb.py b/tests/stats/test_blb.py index df6f8f098..658482853 100644 --- a/tests/stats/test_blb.py +++ b/tests/stats/test_blb.py @@ -142,7 +142,7 @@ def _blb_worker( mean = np.mean(xs).item() timer = Stopwatch() - s = blb_summary(xs, "mean", rng=rng, b_factor=0.75, s_window=20, r_window=50, rel_tol=0.01) + s = blb_summary(xs, "mean", rng=rng, b_factor=0.8, s_window=20, r_window=50, rel_tol=0.01) results.append((mean, s, timer.elapsed())) From 53a371400fe626da9c4b2487da6d5a61787cf104 Mon Sep 17 00:00:00 2001 From: Michael Ekstrand Date: Thu, 14 Aug 2025 16:15:07 -0400 Subject: [PATCH 52/59] refactor test to make more tests easier --- tests/stats/test_blb.py | 219 ++++++++++++++++++---------------------- 1 file changed, 99 insertions(+), 120 deletions(-) diff --git a/tests/stats/test_blb.py b/tests/stats/test_blb.py index 658482853..46b259d6c 100644 --- a/tests/stats/test_blb.py +++ b/tests/stats/test_blb.py @@ -6,6 +6,7 @@ import os from math import sqrt +from typing import ClassVar import numpy as np from numpy.typing import NDArray @@ -43,102 +44,109 @@ def test_blb_single_array(rng: np.random.Generator): assert summary["ci_upper"] == approx(mean + 1.96 * ste, rel=0.01) -@mark.slow -@mark.skipif(not ray_available(), reason="bulk BLB test requires Ray") -@mark.parametrize("size", [1000, 10_000, 100_000]) -@mark.filterwarnings(r"error:.*ignoring \d+ nonfinite values") -def test_blb_array_normal(rng: np.random.Generator, size: int): - "Test BLB with arrays of normals." - import ray - - ensure_cluster() - TRUE_MEAN = 1.0 - TRUE_SD = 1.0 - # TRUE_SVAR = TRUE_SD * TRUE_SD / size - THEORETICAL_SE = TRUE_SD / np.sqrt(size) - THEORETICAL_WIDTH = 2 * 1.96 * THEORETICAL_SE - results = [] - times = [] - - # Test: for NBATCHES * PERBATCH runs, do approx. 95% of confidence intervals - # contain the true mean? - - worker = ray.remote(num_cpus=2)(_blb_worker) - - NBATCHES = 20 - PERBATCH = int(os.environ.get("BLB_TRIALS_PER_BATCH", 50)) - NTRIALS = NBATCHES * PERBATCH - rngs = rng.spawn(NBATCHES) - tasks = [worker.remote(PERBATCH, TRUE_MEAN, TRUE_SD, size, t) for t in rngs] - for task in tasks: - bres = ray.get(task) - for mean, summary, time in bres: - assert isinstance(summary, dict) - assert summary["estimate"] == approx(mean) - - results.append(summary) - times.append(time) - - _log.info("completed %d trials (avg %.2fms / trial)", len(results), np.mean(times) * 1000) - n_lb_good = len([r for r in results if r["ci_lower"] <= TRUE_MEAN]) - f_lb_good = n_lb_good / NTRIALS - n_ub_good = len([r for r in results if TRUE_MEAN <= r["ci_upper"]]) - f_ub_good = n_ub_good / NTRIALS - n_good = len([r for r in results if r["ci_lower"] <= TRUE_MEAN <= r["ci_upper"]]) - f_good = n_good / NTRIALS - bt = binomtest(n_good, NTRIALS, 0.95) - _log.info("binomal test for CI hit rate: stat=%.3f, p=%.3g", bt.statistic, bt.pvalue, test=bt) - - smeans = np.array([r["estimate"] for r in results]) - smt = ttest_1samp(smeans, TRUE_MEAN) - _log.info("sample means: %s", describe(smeans)) - if smt.pvalue >= 0.05: - _log.info("t-test for sample means: stat=%.5f, p=%.3g", smt.statistic, smt.pvalue, test=smt) - else: - _log.warn("t-test for sample means: stat=%.5f, p=%.3g", smt.statistic, smt.pvalue, test=smt) - rmeans = np.array([r["rep_mean"] for r in results]) - rmt = ttest_rel(rmeans, smeans) - _log.info("bootstrap means: %s", describe(rmeans)) - if rmt.pvalue >= 0.05: - _log.info("t-test for CI centers: stat=%.5f, p=%.3g", rmt.statistic, rmt.pvalue, test=rmt) - else: - _log.warn("t-test for CI centers: stat=%.5f, p=%.3g", rmt.statistic, rmt.pvalue, test=rmt) - - widths = np.array([r["ci_upper"] - r["ci_lower"] for r in results]) - _log.info( - "bootstrap CI widths (expected: {:.4f}): {}".format(THEORETICAL_WIDTH, describe(widths)) - ) - wt = ttest_1samp(widths, THEORETICAL_WIDTH) - if wt.pvalue >= 0.05: - _log.info("t-test for CI width: stat=%.5f, p=%.3g", wt.statistic, wt.pvalue, test=wt) - else: - _log.warn("t-test for CI width: stat=%.5f, p=%.3g", wt.statistic, wt.pvalue, test=wt) - - if bt.pvalue >= 0.05: +class CITester: + NBATCHES: ClassVar[int] = 20 + PERBATCH: ClassVar[int] = int(os.environ.get("BLB_TRIALS_PER_BATCH", 50)) + + parameter: float + expected_width: float + + def generate_sample(self, rng: np.random.Generator) -> NDArray[np.float64]: ... + + @mark.filterwarnings(r"error:.*ignoring \d+ nonfinite values") + @mark.parametrize("size", [1000]) + def test_compute(self, size: int, rng: np.random.Generator): + import ray + + ensure_cluster() + + results = [] + times = [] + n_trials = self.NBATCHES * self.PERBATCH + + worker = ray.remote(num_cpus=2)(_blb_worker) + rngs = rng.spawn(self.NBATCHES) + tasks = [worker.remote(self.PERBATCH, size, self, t) for t in rngs] + for task in tasks: + bres = ray.get(task) + for mean, summary, time in bres: + assert isinstance(summary, dict) + assert summary["estimate"] == approx(mean) + + results.append(summary) + times.append(time) + + _log.info("completed %d trials (avg %.2fms / trial)", len(results), np.mean(times) * 1000) + n_lb_good = len([r for r in results if r["ci_lower"] <= self.parameter]) + f_lb_good = n_lb_good / n_trials + n_ub_good = len([r for r in results if self.parameter <= r["ci_upper"]]) + f_ub_good = n_ub_good / n_trials + n_good = len([r for r in results if r["ci_lower"] <= self.parameter <= r["ci_upper"]]) + f_good = n_good / n_trials + bt = binomtest(n_good, n_trials, 0.95) _log.info( - "{:.1%} CIs good ({:1%} LB fail, {:.1%} UB fail), p={:.3g}".format( - f_good, 1 - f_lb_good, 1 - f_ub_good, bt.pvalue - ), - ) - else: - _log.error( - "{:.1%} CIs good ({:1%} LB fail, {:.1%} UB fail), p={:.3g}".format( - f_good, 1 - f_lb_good, 1 - f_ub_good, bt.pvalue - ), + "binomal test for CI hit rate: stat=%.3f, p=%.3g", bt.statistic, bt.pvalue, test=bt ) - # leave some wiggle room - assert bt.pvalue >= 0.05 + smeans = np.array([r["estimate"] for r in results]) + smt = ttest_1samp(smeans, self.parameter) + _log.info("sample means: %s", describe(smeans)) + if smt.pvalue >= 0.05: + _log.info( + "t-test for sample means: stat=%.5f, p=%.3g", smt.statistic, smt.pvalue, test=smt + ) + else: + _log.warn( + "t-test for sample means: stat=%.5f, p=%.3g", smt.statistic, smt.pvalue, test=smt + ) + rmeans = np.array([r["rep_mean"] for r in results]) + rmt = ttest_rel(rmeans, smeans) + _log.info("bootstrap means: %s", describe(rmeans)) + if rmt.pvalue >= 0.05: + _log.info( + "t-test for CI centers: stat=%.5f, p=%.3g", rmt.statistic, rmt.pvalue, test=rmt + ) + else: + _log.warn( + "t-test for CI centers: stat=%.5f, p=%.3g", rmt.statistic, rmt.pvalue, test=rmt + ) + + widths = np.array([r["ci_upper"] - r["ci_lower"] for r in results]) + _log.info( + "bootstrap CI widths (expected: {:.4f}): {}".format( + self.expected_width, describe(widths) + ) + ) + wt = ttest_1samp(widths, self.expected_width) + if wt.pvalue >= 0.05: + _log.info("t-test for CI width: stat=%.5f, p=%.3g", wt.statistic, wt.pvalue, test=wt) + else: + _log.warn("t-test for CI width: stat=%.5f, p=%.3g", wt.statistic, wt.pvalue, test=wt) + + if bt.pvalue >= 0.05: + _log.info( + "{:.1%} CIs good ({:1%} LB fail, {:.1%} UB fail), p={:.3g}".format( + f_good, 1 - f_lb_good, 1 - f_ub_good, bt.pvalue + ), + ) + else: + _log.error( + "{:.1%} CIs good ({:1%} LB fail, {:.1%} UB fail), p={:.3g}".format( + f_good, 1 - f_lb_good, 1 - f_ub_good, bt.pvalue + ), + ) + + # leave some wiggle room + assert bt.pvalue >= 0.05 def _blb_worker( - nreps: int, true_mean: float, true_sd: float, size: int, rng: np.random.Generator + nreps: int, size: int, test: CITester, rng: np.random.Generator ) -> list[tuple[float, dict[str, float], float]]: results = [] - # bf = 0.7 if size > 50_000 else 0.8 for _i in range(nreps): - xs = rng.normal(true_mean, true_sd, size) + xs = test.generate_sample(size, rng) mean = np.mean(xs).item() timer = Stopwatch() @@ -149,38 +157,9 @@ def _blb_worker( return results -@mark.skip("need to find better parameters") -@given( - nph.arrays(shape=st.integers(10000, 1_000_000), dtype=nph.floating_dtypes(endianness="=")), - st.integers(0), -) -def test_blb_array(xs: NDArray[np.floating], seed: int): - "Test BLB with more aggressive edge-case hunting." - xsf = xs[np.isfinite(xs)] - mean = np.mean(xsf) - # ignore grotesquely out-of-bounds cases (for now) - assume(np.isfinite(mean)) - n = len(xsf) - std = np.std(xsf) - ste = std / sqrt(n) - - if np.all(np.isfinite(xs)): - summary = blb_summary(xs, "mean", rng=seed) - else: - with warns(DataWarning, match=r"ignoring \d+ nonfinite"): - summary = blb_summary(xs, "mean", rng=seed) +class TestSimpleNormal(CITester): + parameter = 1.0 + true_sd = 1.0 - assert isinstance(summary, dict) - assert summary["value"] == approx(mean, nan_ok=True) - assert summary["mean"] == approx(mean, rel=0.01, nan_ok=True) - - if n == 0: - assert np.isnan(summary["ci_min"]) - assert np.isnan(summary["ci_max"]) - elif np.allclose(xs, np.min(xs)): - # standard error is zero - assert summary["ci_min"] == approx(mean, rel=0.01) - assert summary["ci_max"] == approx(mean, rel=0.01) - else: - assert summary["ci_min"] == approx(mean - 1.96 * ste, rel=0.01) - assert summary["ci_max"] == approx(mean + 1.96 * ste, rel=0.01) + def generate_sample(self, size: int, rng): + return rng.normal(self.parameter, self.true_sd, size=size) From 8972837e4367c1194cb7f98f655aa54fb02ad26b Mon Sep 17 00:00:00 2001 From: Michael Ekstrand Date: Thu, 14 Aug 2025 16:26:55 -0400 Subject: [PATCH 53/59] refactor tests and test the t-test --- tests/stats/test_blb.py | 115 +++++++++++++++++++++++++++------------- 1 file changed, 78 insertions(+), 37 deletions(-) diff --git a/tests/stats/test_blb.py b/tests/stats/test_blb.py index 46b259d6c..3d30cae4a 100644 --- a/tests/stats/test_blb.py +++ b/tests/stats/test_blb.py @@ -49,9 +49,14 @@ class CITester: PERBATCH: ClassVar[int] = int(os.environ.get("BLB_TRIALS_PER_BATCH", 50)) parameter: float - expected_width: float - def generate_sample(self, rng: np.random.Generator) -> NDArray[np.float64]: ... + def generate_sample(self, size: int, rng: np.random.Generator) -> NDArray[np.float64]: ... + def compute_stats( + self, xs: NDArray[np.float64], rng: np.random.Generator + ) -> dict[str, float]: ... + + def expected_width(self, size: int) -> float | None: + return None @mark.filterwarnings(r"error:.*ignoring \d+ nonfinite values") @mark.parametrize("size", [1000]) @@ -99,42 +104,47 @@ def test_compute(self, size: int, rng: np.random.Generator): _log.warn( "t-test for sample means: stat=%.5f, p=%.3g", smt.statistic, smt.pvalue, test=smt ) - rmeans = np.array([r["rep_mean"] for r in results]) - rmt = ttest_rel(rmeans, smeans) - _log.info("bootstrap means: %s", describe(rmeans)) - if rmt.pvalue >= 0.05: - _log.info( - "t-test for CI centers: stat=%.5f, p=%.3g", rmt.statistic, rmt.pvalue, test=rmt - ) - else: - _log.warn( - "t-test for CI centers: stat=%.5f, p=%.3g", rmt.statistic, rmt.pvalue, test=rmt - ) + try: + rmeans = np.array([r["rep_mean"] for r in results]) + rmt = ttest_rel(rmeans, smeans) + _log.info("bootstrap means: %s", describe(rmeans)) + if rmt.pvalue >= 0.05: + _log.info( + "t-test for CI centers: stat=%.5f, p=%.3g", rmt.statistic, rmt.pvalue, test=rmt + ) + else: + _log.warn( + "t-test for CI centers: stat=%.5f, p=%.3g", rmt.statistic, rmt.pvalue, test=rmt + ) + except KeyError: + pass widths = np.array([r["ci_upper"] - r["ci_lower"] for r in results]) - _log.info( - "bootstrap CI widths (expected: {:.4f}): {}".format( - self.expected_width, describe(widths) - ) - ) - wt = ttest_1samp(widths, self.expected_width) - if wt.pvalue >= 0.05: - _log.info("t-test for CI width: stat=%.5f, p=%.3g", wt.statistic, wt.pvalue, test=wt) - else: - _log.warn("t-test for CI width: stat=%.5f, p=%.3g", wt.statistic, wt.pvalue, test=wt) - - if bt.pvalue >= 0.05: - _log.info( - "{:.1%} CIs good ({:1%} LB fail, {:.1%} UB fail), p={:.3g}".format( - f_good, 1 - f_lb_good, 1 - f_ub_good, bt.pvalue - ), - ) - else: - _log.error( - "{:.1%} CIs good ({:1%} LB fail, {:.1%} UB fail), p={:.3g}".format( - f_good, 1 - f_lb_good, 1 - f_ub_good, bt.pvalue - ), - ) + ew = self.expected_width(size) + _log.info("bootstrap CI widths (expected: {:.4f}): {}".format(ew, describe(widths))) + if ew is not None: + wt = ttest_1samp(widths, ew) + if wt.pvalue >= 0.05: + _log.info( + "t-test for CI width: stat=%.5f, p=%.3g", wt.statistic, wt.pvalue, test=wt + ) + else: + _log.warn( + "t-test for CI width: stat=%.5f, p=%.3g", wt.statistic, wt.pvalue, test=wt + ) + + if bt.pvalue >= 0.05: + _log.info( + "{:.1%} CIs good ({:1%} LB fail, {:.1%} UB fail), p={:.3g}".format( + f_good, 1 - f_lb_good, 1 - f_ub_good, bt.pvalue + ), + ) + else: + _log.error( + "{:.1%} CIs good ({:1%} LB fail, {:.1%} UB fail), p={:.3g}".format( + f_good, 1 - f_lb_good, 1 - f_ub_good, bt.pvalue + ), + ) # leave some wiggle room assert bt.pvalue >= 0.05 @@ -150,16 +160,47 @@ def _blb_worker( mean = np.mean(xs).item() timer = Stopwatch() - s = blb_summary(xs, "mean", rng=rng, b_factor=0.8, s_window=20, r_window=50, rel_tol=0.01) + s = test.compute_stats(xs, rng) results.append((mean, s, timer.elapsed())) return results +class TestParamNormal(CITester): + parameter = 1.0 + true_sd = 1.0 + + def expected_width(self, size: int): + se = self.true_sd / np.sqrt(size) + return 2 * 1.96 * se + + def generate_sample(self, size: int, rng): + return rng.normal(self.parameter, self.true_sd, size=size) + + def compute_stats(self, xs, rng: np.random.Generator): + mean = np.mean(xs) + ssd = np.std(xs, ddof=1) + sse = ssd / np.sqrt(len(xs)) + return { + "estimate": mean, + "ci_lower": mean - 1.96 * sse, + "ci_upper": mean + 1.96 * sse, + } + + class TestSimpleNormal(CITester): parameter = 1.0 true_sd = 1.0 + def expected_width(self, size: int): + se = self.true_sd / np.sqrt(size) + return 2 * 1.96 * se + def generate_sample(self, size: int, rng): return rng.normal(self.parameter, self.true_sd, size=size) + + def compute_stats(self, xs, rng: np.random.Generator): + return blb_summary( + xs, "mean", rng=rng, b_factor=0.8, s_window=20, r_window=50, rel_tol=0.01 + ) From df701c3996c500c818f4972f175f0eb69386aedd Mon Sep 17 00:00:00 2001 From: Michael Ekstrand Date: Thu, 14 Aug 2025 16:38:20 -0400 Subject: [PATCH 54/59] compute samples in-thread --- src/lenskit/stats/_blb.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/lenskit/stats/_blb.py b/src/lenskit/stats/_blb.py index 7dc3f798c..8f30497a5 100644 --- a/src/lenskit/stats/_blb.py +++ b/src/lenskit/stats/_blb.py @@ -244,7 +244,7 @@ def measure_subset( lbs = StatAccum(None) ubs = StatAccum(None) - loop = self._rep_generator.subsets() + loop = self.miniboot_weights(n, b) for i, weights in enumerate(loop): self._tracer.add_bindings(rep=i) self._tracer.trace("starting replicate") From 2d6fefbd369c698e3727dbdabfa103597995a8b5 Mon Sep 17 00:00:00 2001 From: Michael Ekstrand Date: Thu, 14 Aug 2025 16:39:59 -0400 Subject: [PATCH 55/59] eliminate parallel replicate logic --- src/lenskit/stats/_blb.py | 109 ++++++-------------------------------- 1 file changed, 17 insertions(+), 92 deletions(-) diff --git a/src/lenskit/stats/_blb.py b/src/lenskit/stats/_blb.py index 8f30497a5..7cb469da7 100644 --- a/src/lenskit/stats/_blb.py +++ b/src/lenskit/stats/_blb.py @@ -7,11 +7,9 @@ from __future__ import annotations import warnings -from collections import deque -from collections.abc import Callable, Generator +from collections.abc import Callable from dataclasses import dataclass -from threading import Condition, Lock, Thread -from typing import Any, ClassVar, Deque, Literal, Protocol, TypeAlias, TypeVar +from typing import Any, ClassVar, Literal, Protocol, TypeAlias, TypeVar import numpy as np import pandas as pd @@ -173,7 +171,6 @@ class _BLBootstrapper: _ci_qmax: float rng: np.random.Generator - _rep_generator: ReplicateGenerator def __init__(self, config, rng: np.random.Generator): self.config = config @@ -201,23 +198,21 @@ def run_bootstraps(self, xs: NDArray[F]) -> _BootResult: self._tracer.trace("estimating acceleration term") accel = _bca_accel_term(xs, self.config.statistic) - self._rep_generator = ReplicateGenerator(n, b, self.rng) self._tracer.trace("let's go!") - with self._rep_generator: - for i, ss in enumerate(self.blb_subsets(n, b)): - self._tracer.add_bindings(subset=i) - self._tracer.trace("starting subset") - res = self.measure_subset(xs, ss, estimate, accel) - ss_frames[i] = res.samples - means.record(res.rep_mean) - vars.record(res.rep_var) - lbs.record(res.ci_lower) - ubs.record(res.ci_upper) - if self._check_convergence( - means, vars, lbs, ubs, tol=self.config.rel_tol, w=self.config.s_window - ): - break + for i, ss in enumerate(self.blb_subsets(n, b)): + self._tracer.add_bindings(subset=i) + self._tracer.trace("starting subset") + res = self.measure_subset(xs, ss, estimate, accel) + ss_frames[i] = res.samples + means.record(res.rep_mean) + vars.record(res.rep_var) + lbs.record(res.ci_lower) + ubs.record(res.ci_upper) + if self._check_convergence( + means, vars, lbs, ubs, tol=self.config.rel_tol, w=self.config.s_window + ): + break return _BootResult( estimate, @@ -244,8 +239,7 @@ def measure_subset( lbs = StatAccum(None) ubs = StatAccum(None) - loop = self.miniboot_weights(n, b) - for i, weights in enumerate(loop): + for i, weights in enumerate(self.miniboot_weights(n, b)): self._tracer.add_bindings(rep=i) self._tracer.trace("starting replicate") assert weights.shape == (b,) @@ -267,7 +261,7 @@ def measure_subset( if self._check_convergence( means, vars, lbs, ubs, tol=self.config.rel_tol, w=self.config.r_window ): - loop.close() + break df = pd.DataFrame({"statistic": means.values}) df.index.name = "iter" @@ -297,75 +291,6 @@ def _check_convergence(self, *arrays: StatAccum, tol: float, w: int) -> bool: return np.all(gaps < tol).item() -class ReplicateGenerator: - """ - Generate the subset samples for a bootstrap in a background thread. - """ - - n: int - b: int - - _rng: np.random.Generator - _flat: NDArray[np.float64] - _lock: Lock - _notify: Condition - _running: bool = True - _queue: Deque - _thread: Thread - - def __init__(self, n: int, b: int, rng: np.random.Generator): - self.n = n - self.b = b - self._rng = rng.spawn(1)[0] - self._queue = deque() - self._flat = np.full(b, 1.0 / b) - self._lock = Lock() - self._notify = Condition(self._lock) - - def subsets(self) -> Generator[NDArray[np.int64], None, None]: - while True: - with self._notify: - while self._thread.is_alive() and len(self._queue) == 0: - self._notify.wait() - - try: - val = self._queue.popleft() - self._notify.notify_all() - except IndexError: - break # things have shut down, loop is over - except GeneratorExit: - break # we've been asked to close - - yield val - - def _generate(self): - with self._notify: - while True: - # check if we need to wake up - while self._running and len(self._queue) >= 5: - trace(_log, "waiting for queue", len=len(self._queue)) - self._notify.wait() - - # are we done? - if not self._running: - break - - # generate a new value - val = self._rng.multinomial(self.n, self._flat) - self._queue.append(val) - self._notify.notify_all() - - def __enter__(self): - self._thread = Thread(target=self._generate) - self._thread.start() - return self - - def __exit__(self, *args: Any): - with self._notify: - self._running = False - self._notify.notify_all() - - class StatAccum: INIT_SIZE: ClassVar[int] = 100 From 2e6660dee92f234edfcbe115856dda1ae3b97546 Mon Sep 17 00:00:00 2001 From: Michael Ekstrand Date: Thu, 14 Aug 2025 16:40:29 -0400 Subject: [PATCH 56/59] eliminate BCa logic --- src/lenskit/stats/_blb.py | 80 ++------------------------------------- 1 file changed, 3 insertions(+), 77 deletions(-) diff --git a/src/lenskit/stats/_blb.py b/src/lenskit/stats/_blb.py index 7cb469da7..25072dd78 100644 --- a/src/lenskit/stats/_blb.py +++ b/src/lenskit/stats/_blb.py @@ -17,7 +17,7 @@ from numpy.typing import NDArray from lenskit.diagnostics import DataWarning -from lenskit.logging import Tracer, get_logger, get_tracer, trace +from lenskit.logging import Tracer, get_logger, get_tracer from lenskit.random import RNGInput, random_generator from ._distributions import ci_quantiles @@ -195,15 +195,12 @@ def run_bootstraps(self, xs: NDArray[F]) -> _BootResult: lbs = StatAccum(np.mean) ubs = StatAccum(np.mean) - self._tracer.trace("estimating acceleration term") - accel = _bca_accel_term(xs, self.config.statistic) - self._tracer.trace("let's go!") for i, ss in enumerate(self.blb_subsets(n, b)): self._tracer.add_bindings(subset=i) self._tracer.trace("starting subset") - res = self.measure_subset(xs, ss, estimate, accel) + res = self.measure_subset(xs, ss, estimate) ss_frames[i] = res.samples means.record(res.rep_mean) vars.record(res.rep_var) @@ -227,9 +224,7 @@ def blb_subsets(self, n: int, b: int): while True: yield self.rng.choice(n, b, replace=False) - def measure_subset( - self, xs: NDArray[F], ss: NDArray[np.int64], estimate: float, accel: float - ) -> _BootResult: + def measure_subset(self, xs: NDArray[F], ss: NDArray[np.int64], estimate: float) -> _BootResult: b = len(ss) n = len(xs) xss = xs[ss] @@ -343,72 +338,3 @@ def _expand_if_needed(self): def __len__(self): return self._len - - -def _bca_range( - estimate: float, replicates: NDArray[np.floating[Any]], margin: float, accel: float -) -> tuple[float, float]: - """ - Estimate the BCa quantiles for a bootstrap. - - This follows Slide 34 of `http://users.stat.umn.edu/~helwig/notes/bootci-Notes.pdf`_. - """ - bias = _bca_bias_corrector(estimate, replicates) - trace(_log, "B=%d, estimate=%f, bias=%f", len(replicates), estimate, bias) - - z1 = bias + STD_NORM.ppf(margin) - icd1 = z1 / (1 - accel * z1) - - z2 = bias + STD_NORM.ppf(1 - margin) - icd2 = z2 / (1 - accel * z2) - - return STD_NORM.cdf(icd1), STD_NORM.cdf(icd2) - - -def _bca_bias_corrector(statistic: float, replicates: NDArray[np.floating[Any]]) -> float: - B = len(replicates) - nlow = np.sum(replicates < statistic) - if nlow == 0 or nlow == B: - # extremely biased, but goes OOB. Should only happen early in the bootstrap. - return 0 - else: - return STD_NORM.ppf(nlow / B) - - -def _bca_accel_term(xs: NDArray[np.floating[Any]], statistic: WeightedStatistic) -> float: - """ - Compute the BCa acceleration term. - - Follows slide 36 of - `http://users.stat.umn.edu/~helwig/notes/bootci-Notes.pdf`_, referring also - to the SciPy `scipy/stats/_resampling.py` for implementation ideas. - """ - N = len(xs) - BSIZE = 5000 - jk_vals = np.empty(N) - # batch the jackknife, because our data might be huge - # TODO: can we sample the jackknife? - for start in range(0, N, BSIZE): - end = min(start + BSIZE, N) - B = end - start - # this trick is from scipy — set up a mask - mask = np.ones((B, N), dtype=np.bool_) - np.fill_diagonal(mask[:, start:end], False) - # and reshape — again, borrwed from scipy - i = np.broadcast_to(np.arange(N), (B, N)) - i = i[mask].reshape((B, N - 1)) - - # prepare B x N batched sample and compute statistics - sample = xs[i] - stats = statistic(sample, axis=-1) - assert stats.shape == (B,) - jk_vals[start:end] = stats - - jk_est = np.mean(jk_vals) - jk_dev = jk_est - jk_vals - - # sum of cubes - accel_num = np.sum(np.power(jk_dev, 3)) - # weird term - accel_denom = 6 * np.power(np.sum(np.square(jk_dev)), 1.5) - return accel_num / accel_denom From 0a5182d23e3f9f4e2ca06eea256fb9b7236b48d4 Mon Sep 17 00:00:00 2001 From: Michael Ekstrand Date: Thu, 14 Aug 2025 16:49:37 -0400 Subject: [PATCH 57/59] support minimum inner iterations --- src/lenskit/stats/_blb.py | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) diff --git a/src/lenskit/stats/_blb.py b/src/lenskit/stats/_blb.py index 25072dd78..2dee7bdf9 100644 --- a/src/lenskit/stats/_blb.py +++ b/src/lenskit/stats/_blb.py @@ -57,6 +57,7 @@ def blb_summary( rel_tol: float = 0.02, s_window: int = 10, r_window: int = 50, + r_min: int = 0, rng: RNGInput = None, ) -> dict[str, float]: r""" @@ -112,6 +113,7 @@ def blb_summary( rel_tol=rel_tol, s_window=s_window, r_window=r_window, + r_min=r_min, b_factor=b_factor, ) bootstrapper = _BLBootstrapper(config, rng) @@ -153,6 +155,7 @@ class _BLBConfig: rel_tol: float s_window: int r_window: int + r_min: int b_factor: float @property @@ -234,7 +237,7 @@ def measure_subset(self, xs: NDArray[F], ss: NDArray[np.int64], estimate: float) lbs = StatAccum(None) ubs = StatAccum(None) - for i, weights in enumerate(self.miniboot_weights(n, b)): + for i, weights in enumerate(self.miniboot_weights(n, b), start=1): self._tracer.add_bindings(rep=i) self._tracer.trace("starting replicate") assert weights.shape == (b,) @@ -253,7 +256,7 @@ def measure_subset(self, xs: NDArray[F], ss: NDArray[np.int64], estimate: float) ubs.record(stat, ub) del stats - if self._check_convergence( + if i >= self.config.r_min and self._check_convergence( means, vars, lbs, ubs, tol=self.config.rel_tol, w=self.config.r_window ): break From 82e84755824b233aae69561b0a22da1f18e0a791 Mon Sep 17 00:00:00 2001 From: Michael Ekstrand Date: Thu, 14 Aug 2025 16:49:51 -0400 Subject: [PATCH 58/59] tweak up params for BLB test --- tests/stats/test_blb.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tests/stats/test_blb.py b/tests/stats/test_blb.py index 3d30cae4a..c4f108521 100644 --- a/tests/stats/test_blb.py +++ b/tests/stats/test_blb.py @@ -202,5 +202,5 @@ def generate_sample(self, size: int, rng): def compute_stats(self, xs, rng: np.random.Generator): return blb_summary( - xs, "mean", rng=rng, b_factor=0.8, s_window=20, r_window=50, rel_tol=0.01 + xs, "mean", rng=rng, b_factor=0.6, s_window=20, r_window=25, r_min=100, rel_tol=0.01 ) From 76334fefd3ca28b3b59a484be8fd99e322215056 Mon Sep 17 00:00:00 2001 From: Michael Ekstrand Date: Thu, 14 Aug 2025 16:50:15 -0400 Subject: [PATCH 59/59] one more param tweak --- tests/stats/test_blb.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tests/stats/test_blb.py b/tests/stats/test_blb.py index c4f108521..df9620276 100644 --- a/tests/stats/test_blb.py +++ b/tests/stats/test_blb.py @@ -202,5 +202,5 @@ def generate_sample(self, size: int, rng): def compute_stats(self, xs, rng: np.random.Generator): return blb_summary( - xs, "mean", rng=rng, b_factor=0.6, s_window=20, r_window=25, r_min=100, rel_tol=0.01 + xs, "mean", rng=rng, b_factor=0.8, s_window=20, r_window=25, r_min=100, rel_tol=0.01 )