Implement support for segments with an optional categories key

tests/test_new_transcript_schema.py

@@ -1,4 +1,6 @@
-from audita.core.schemas import parse_source_transcript_json, parse_transcript_json
+import json
+
+from audita.core.schemas import parse_source_transcript_json, parse_transcript_json, transcript_to_json
 
 
 SERIATIM_TRANSCRIPT = """
@@ -60,6 +62,8 @@ def test_parse_source_transcript_json_accepts_seriatim_transcript_object():
     assert segments[0].end == 3.5
     assert segments[0].text == "Hello there."
     assert segments[1].text == "Resolved word run"
+    assert segments[0].categories is None
+    assert segments[1].categories == ["backchannel"]
 
 
 def test_parse_transcript_json_accepts_seriatim_transcript_object_and_ignores_unused_fields():
@@ -67,3 +71,14 @@ def test_parse_transcript_json_accepts_seriatim_transcript_object_and_ignores_un
 
     assert [segment.id for segment in segments] == [1, 2]
     assert [segment.text for segment in segments] == ["Hello there.", "Resolved word run"]
+    assert segments[0].categories is None
+    assert segments[1].categories == ["backchannel"]
+
+
+def test_transcript_to_json_emits_categories_only_when_present():
+    segments = parse_transcript_json(SERIATIM_TRANSCRIPT)
+
+    payload = json.loads(transcript_to_json(segments))
+
+    assert "categories" not in payload[0]
+    assert payload[1]["categories"] == ["backchannel"]