admin can split a sentence into 2 sentences

sentence_api/tests/test_api.py

@@ -188,3 +188,74 @@ def test_processor_marks_ready_only_after_oss_upload(tmp_path, monkeypatch):
     assert uploads[0]["stored_filename"] == f"{VIDEO_HASH}.mp4"
     assert video["status"] == "ready"
     assert video["remote_url"] == "https://media.example.com/lesson.mp4"
+
+
+def test_admin_can_split_sentence_and_reindexes_following_sentences(tmp_path):
+    settings = replace(Settings.from_env(), data_dir=tmp_path / "data")
+    settings.ensure_directories()
+    repository = VideoRepository(settings.database_path)
+    repository.upsert_upload(
+        video_hash=VIDEO_HASH,
+        title="Manual split lesson",
+        filename="lesson.mp4",
+        stored_filename=f"{VIDEO_HASH}.mp4",
+        content_type="video/mp4",
+        size_bytes=5,
+        language="en",
+    )
+    repository.save_processing_result(
+        SentenceBoundaryDocument(
+            video_hash=VIDEO_HASH,
+            duration_ms=9000,
+            algorithm_version="test",
+            sentences=[
+                SentenceBoundary(
+                    index=0,
+                    start_ms=0,
+                    end_ms=4000,
+                    text="First part second part",
+                    language="en",
+                    reference_speech_duration_ms=3600,
+                ),
+                SentenceBoundary(
+                    index=1,
+                    start_ms=4000,
+                    end_ms=9000,
+                    text="Next sentence.",
+                    language="en",
+                    reference_speech_duration_ms=4600,
+                ),
+            ],
+        ),
+        None,
+    )
+
+    class UnavailableTranscriber:
+        available = False
+
+    client = TestClient(
+        create_app(
+            BoundaryStore(tmp_path / "boundaries.json"),
+            settings=settings,
+            repository=repository,
+            transcriber=UnavailableTranscriber(),
+        )
+    )
+    response = client.put(
+        f"/api/v1/admin/videos/{VIDEO_HASH}/sentences/0/split",
+        json={"text": "First part\nsecond part", "language": "en"},
+    )
+
+    assert response.status_code == 200
+    replacements = response.json()
+    assert [item["text"] for item in replacements] == ["First part", "second part"]
+    assert replacements[0]["start_ms"] == 0
+    assert replacements[0]["end_ms"] == replacements[1]["start_ms"]
+    assert replacements[1]["end_ms"] == 4000
+    document = repository.get_document(VIDEO_HASH)
+    assert [item.text for item in document.sentences] == [
+        "First part",
+        "second part",
+        "Next sentence.",
+    ]
+    assert [item.index for item in document.sentences] == [0, 1, 2]