Add endpoints for quantity kinds

tests/integration/concepts_integration.py

@@ -1,6 +1,6 @@
 import pytest
 
-from units.concepts import get_all_data_for_qk_iri, get_qk_for_iri
+from units.concepts import get_all_data_for_qk_iri, get_qk_for_iri, get_quantity_kinds
 
 
 def test_qk():
@@ -33,3 +33,17 @@ def test_gadfqi_namespaces():
     assert "https://vocab.sentier.dev/qudt/unit/M-SEC" in result
     subset = result["https://vocab.sentier.dev/qudt/unit/M-SEC"]
     assert ("http://www.w3.org/2004/02/skos/core#prefLabel", "Metre second", "en-gb") in subset
+
+
+def test_get_quantity_kinds():
+    result = get_quantity_kinds(
+        remove_namespaces=True,
+    )
+    assert "https://vocab.sentier.dev/qudt/quantity-kind/Acceleration" in result
+    subset = result["https://vocab.sentier.dev/qudt/quantity-kind/Acceleration"]
+    assert subset["informativeReference"] == "http://en.wikipedia.org/wiki/Acceleration"
+    assert ("accélération", "fr") in subset["prefLabel"]
+    assert len(subset["prefLabel"]) > 1
+
+    subset = result["https://vocab.sentier.dev/qudt/quantity-kind/Enthalpy"]
+    assert len(subset["prefLabel"]) > 1

units/concepts.py

@@ -1,3 +1,4 @@
+from collections import defaultdict
 from functools import partial
 from itertools import groupby
 
@@ -74,6 +75,63 @@ def get_qk_for_iri(iri: str) -> str | None:
     return response[0]["qk"]
 
 
+def get_quantity_kinds(
+    lang: str | None = None,
+    remove_namespaces: bool = True,
+) -> dict:
+    """Get the all QUDT quantity kinds."""
+    logger.debug("Using sparql endpoint url %s", settings.SPARQL_URL)
+
+    QUERY = f"""
+PREFIX qudt: <http://qudt.org/schema/qudt/>
+PREFIX skos: <http://www.w3.org/2004/02/skos/core#>
+
+SELECT ?s ?p ?o
+FROM <{settings.VOCAB_PREFIX}qudt/>
+where {{
+    ?s ?p ?o
+    FILTER (
+      contains(STR(?s), "https://vocab.sentier.dev/qudt/quantity-kind/")
+    )
+    FILTER (
+        ?p IN (skos:prefLabel, skos:altLabel, skos:exactMatch, skos:related, skos:definition, qudt:informativeReference)
+    )
+}}
+    """
+
+    logger.debug("Executing query %s", QUERY)
+    response = httpx.post(settings.SPARQL_URL, data={"query": QUERY})
+    response.raise_for_status()
+
+    lang_checker = partial(language_filter, lang=lang.lower() if lang else None)
+    data = [
+        (
+            obj["s"]["value"],
+            reformat_predicate_object(obj, remove_namespaces=remove_namespaces),
+        )
+        for obj in response.json()["results"]["bindings"]
+        if lang_checker(obj["o"])
+    ]
+    logger.info(f"Retrieved {len(data)} quantity kinds")
+
+    results = {}
+
+    # Sorry this is shit but I lost 30 mins fighting with groupby
+    # and am too burned out to find a better way...
+    for qk, elem in data:
+        if qk not in results:
+            results[qk] = defaultdict(list)
+        if len(elem) == 2:
+            results[qk][elem[0]].append(elem[1])
+        else:
+            results[qk][elem[0]].append(elem[1:])
+
+    return {
+        key: {a: (b[0] if len(b) == 1 else b) for a, b in value.items()}
+        for key, value in results.items()
+    }
+
+
 def get_all_data_for_qk_iri(
     iri: str,
     lang: str | None = None,

units/routes.py

@@ -3,7 +3,7 @@
 from fastapi_versioning import version
 
 from units.schema import VersionResponse
-from units.concepts import get_qk_for_iri, get_all_data_for_qk_iri
+from units.concepts import get_qk_for_iri, get_all_data_for_qk_iri, get_quantity_kinds
 
 router = APIRouter()
 
@@ -35,10 +35,9 @@ async def get_concept_data(
 ) -> str:
     """
     Returns all concepts in the same QUDT quantity kind class as the unit `iri`. Data is formatted
-    as a JSON `Map`, with keys of unit IRIs, and values of lists of `[predicate, object]`.
-
-    We use lists because a given unit can share the same predicate relation with more than one
-    object. For example, a unit could have multiple preferred labels in different languages.
+    as a JSON `Map`, with keys of unit IRIs, and values of values of maps of `{key: value}`. Because
+    the same key can be used more than once (e.g. a `prefLabel` can have values in different
+    languages), the `value` can be either a single object or an array of objects.
 
     Pass `remove_namespaces` to control the verbosity of the response. By default, some common
     namespace prefixes of the predicates and objects are removed:
@@ -65,7 +64,36 @@ async def get_concept_data(
     return JSONResponse(content=result)
 
 
-# https://vocab.sentier.dev/qudt/unit/M-SEC
+@router.get("/quantity_kinds")
+@version(0, 1)
+async def get_quantity_kinds_data(
+    remove_namespaces: bool = True,
+    lang: str | None = None,
+) -> str:
+    """
+    Returns all quantity kinds concepts. Data is formatted as a JSON `Map`, with keys of unit IRIs,
+    and values of maps of `{key: value}`. Because the same key can be used more than once (e.g. a
+    `prefLabel` can have values in different languages), the `value` can be either a single object
+    or an array of objects.
+
+    Pass `remove_namespaces` to control the verbosity of the response. By default, some common
+    namespace prefixes of the predicates and objects are removed:
+
+    * http://qudt.org/schema/qudt/
+    * http://www.w3.org/1999/02/22-rdf-syntax-ns#
+    * http://www.w3.org/2004/02/skos/core#
+
+    Use `lang` to control what language codes are available in the response. Response data can
+    include RDF literals with many languages, and the default is not to do any filtering. If you
+    pass `lang`, then only RDF literals who explicitly provide a language code which starts the same
+    as `lang` will be returned. In other words `lang='en'` will return object literals without a
+    language code, with a `en` language code, with a `en_GB` language code, but not a `jp` code.
+
+    """
+    result = get_quantity_kinds(
+        lang=lang, remove_namespaces=remove_namespaces
+    )
+    return JSONResponse(content=result)
 
 
 @router.get("/version")