Merge branch 'main' of github.com:theeluwin/textrankr into main

theeluwin · Feb 11, 2021 · 94d6cdb · 94d6cdb
2 parents ac50936 + 3c5fa36
commit 94d6cdb
Showing 1 changed file with 1 addition and 1 deletion.
diff --git a/textrankr/utils.py b/textrankr/utils.py
@@ -32,7 +32,7 @@ def parse_text_into_sentences(text: str, tokenizer: Callable[[str], List[str]])
     sentences: List[Sentence] = []
 
     # parse text
-    candidates: List[str] = split(r'(?:(?<=[^0-9])\.|\n)', text)
+    candidates: List[str] = split(r'(?:(?<=[^0-9])\.|\n|!|\?)', text)
     for candidate in candidates:
 
         # cleanse the candidate