continue fixing

sentence_api/tests/test_transcription.py

@@ -351,6 +351,83 @@ def test_split_sentences_handles_ellipsis_and_dots_only():
     ) == []
 
 
+def test_split_sentences_ignores_decimal_points():
+    segment = TranscriptionSegment(
+        start_seconds=0.0,
+        end_seconds=6.0,
+        text="The rate rose by 0.4%. That is a big jump.",
+    )
+    sentences = split_sentences_at_punctuation([segment])
+    assert [s.text for s in sentences] == [
+        "The rate rose by 0.4%.",
+        "That is a big jump.",
+    ]
+
+
+def test_split_sentences_ignores_am_pm_abbreviations():
+    segment = TranscriptionSegment(
+        start_seconds=0.0,
+        end_seconds=8.0,
+        text=(
+            "It was at 9 o'clock p.m. Eastern, President Trump was speaking. "
+            "The room was quiet."
+        ),
+    )
+    sentences = split_sentences_at_punctuation([segment])
+    assert [s.text for s in sentences] == [
+        "It was at 9 o'clock p.m. Eastern, President Trump was speaking.",
+        "The room was quiet.",
+    ]
+
+    morning = TranscriptionSegment(
+        start_seconds=0.0,
+        end_seconds=4.0,
+        text="I woke up at 7 a.m. and had coffee. Then I left.",
+    )
+    sentences = split_sentences_at_punctuation([morning])
+    assert [s.text for s in sentences] == [
+        "I woke up at 7 a.m. and had coffee.",
+        "Then I left.",
+    ]
+
+
+def test_split_sentences_merges_segment_ending_with_abbreviation():
+    segments = [
+        TranscriptionSegment(0.0, 5.0, "The meeting ends at 5 p.m."),
+        TranscriptionSegment(5.0, 9.0, "Eastern time. Then we go home."),
+    ]
+    sentences = split_sentences_at_punctuation(segments)
+    assert [s.text for s in sentences] == [
+        "The meeting ends at 5 p.m. Eastern time.",
+        "Then we go home.",
+    ]
+
+
+def test_split_sentences_distinguishes_no_reply_from_number():
+    reply = TranscriptionSegment(0.0, 4.0, "No. I don't think so.")
+    assert [s.text for s in split_sentences_at_punctuation([reply])] == [
+        "No.",
+        "I don't think so.",
+    ]
+    numbered = TranscriptionSegment(0.0, 4.0, "See no. 5 on the list.")
+    assert [s.text for s in split_sentences_at_punctuation([numbered])] == [
+        "See no. 5 on the list."
+    ]
+
+
+def test_split_sentences_ignores_common_abbreviations():
+    segment = TranscriptionSegment(
+        start_seconds=0.0,
+        end_seconds=4.0,
+        text="Dr. Smith said the U.S. economy is growing. He was right.",
+    )
+    sentences = split_sentences_at_punctuation([segment])
+    assert [s.text for s in sentences] == [
+        "Dr. Smith said the U.S. economy is growing.",
+        "He was right.",
+    ]
+
+
 def test_post_audio_requests_word_timestamps_and_falls_back(monkeypatch):
     calls = []