fix: ignore empty trailing pieces in sentence count

CHANGELOG.md

@@ -18,3 +18,4 @@ All notable changes to this project will be documented in this file.
 ### Removed
 
 ### Fixed
+- Ignore empty pieces in `verify_sentence_constraint` after a trailing terminator. (https://github.com/allenai/open-instruct/pull/1780).

open_instruct/if_functions.py

@@ -230,8 +230,8 @@ def verify_sentence_constraint(text: str, N: int, quantifier: str) -> bool:
     Returns:
         bool: True if the text contains the expected number of sentences, False otherwise
     """
-    # Split the text into sentences
-    sentences = re.split(r"(?<!\w\.\w.)(?<![A-Z][a-z]\.)(?<=\.|\?|!)\s", text)
+    # Split the text into sentences; drop empty trailing pieces from a final terminator.
+    sentences = [s for s in re.split(r"(?<!\w\.\w.)(?<![A-Z][a-z]\.)(?<=\.|\?|!)\s", text) if s.strip()]
 
     # Count the number of sentences
     actual_count = len(sentences)

open_instruct/test_if_functions.py

@@ -39,5 +39,19 @@ def test_validate_frequency_capital_words(self, _name, text, n, quantifier, expe
         self.assertEqual(if_functions.validate_frequency_capital_words(text, n, quantifier), expected)
 
 
+
+class TestVerifySentenceConstraint(unittest.TestCase):
+    @parameterized.expand(
+        [
+            ("two_sentences", "One. Two.", 2, "at least", True),
+            ("trailing_space", "One. Two. ", 2, "at most", True),
+            ("trailing_was_three", "One. Two. ", 3, "at least", False),
+            ("empty", "", 1, "at least", False),
+        ]
+    )
+    def test_verify_sentence_constraint(self, _name, text, n, quantifier, expected):
+        self.assertEqual(if_functions.verify_sentence_constraint(text, n, quantifier), expected)
+
+
 if __name__ == "__main__":
     unittest.main()