From 513f9e803e36e88b92b64fe6684a58c30189c718 Mon Sep 17 00:00:00 2001
From: Paul Selzner <78799570+pselzner@users.noreply.github.com>
Date: Fri, 5 Dec 2025 15:00:30 +0100
Subject: [PATCH] set lemmatization as default

---
 main.py | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/main.py b/main.py
index d48e02e..230d15f 100644
--- a/main.py
+++ b/main.py
@@ -41,7 +41,7 @@ class BIBFileInput(FileSettings, InputSettings):
 class PreprocessTXT(EnvSettings):
     LANGUAGE: str = "en"
     FILTER_STOPWORDS: bool = True
-    UNIGRAM_NORMALIZER: str = "porter"
+    UNIGRAM_NORMALIZER: str = "lemma"
     USE_NGRAMS: bool = True
     NGRAM_MIN: int = 2
     NGRAM_MAX: int = 3
@@ -54,7 +54,7 @@ class PreprocessTXT(EnvSettings):
 class PreprocessBIB(EnvSettings):
     LANGUAGE: str = "en"
     FILTER_STOPWORDS: bool = True
-    UNIGRAM_NORMALIZER: str = "porter"
+    UNIGRAM_NORMALIZER: str = "lemma"
     USE_NGRAMS: bool = True
     NGRAM_MIN: int = 2
     NGRAM_MAX: int = 3