self-review

server/reflector/hatchet/workflows/diarization_pipeline.py

@@ -81,6 +81,7 @@ from reflector.utils.daily import (
     parse_daily_recording_filename,
 )
 from reflector.utils.string import NonEmptyString, assert_non_none_and_non_empty
+from reflector.utils.transcript_constants import TOPIC_CHUNK_WORD_COUNT
 from reflector.zulip import post_transcript_notification
 
 
@@ -558,9 +559,6 @@ async def detect_topics(input: PipelineInput, ctx: Context) -> TopicsResult:
         TranscriptTopic,
         transcripts_controller,
     )
-    from reflector.utils.transcript_constants import (  # noqa: PLC0415
-        TOPIC_CHUNK_WORD_COUNT,
-    )
 
     chunk_size = TOPIC_CHUNK_WORD_COUNT
     chunks = []
@@ -606,13 +604,11 @@ async def detect_topics(input: PipelineInput, ctx: Context) -> TopicsResult:
 
     results = await topic_chunk_workflow.aio_run_many(bulk_runs)
 
-    topic_results = []
-    for result in results:
-        task_result = result.get("detect_chunk_topic", {})
-        if task_result:
-            topic_results.append(task_result)
-
-    topic_results.sort(key=lambda t: t.get("timestamp", 0))
+    topic_results = [
+        result.get("detect_chunk_topic", {})
+        for result in results
+        if result.get("detect_chunk_topic")
+    ]
 
     async with fresh_db_connection():
         transcript = await transcripts_controller.get_by_id(input.transcript_id)
@@ -821,10 +817,9 @@ async def process_subjects(input: PipelineInput, ctx: Context) -> ProcessSubject
 
     results = await subject_workflow.aio_run_many(bulk_runs)
 
-    subject_summaries = []
-    for result in results:
-        task_result = result.get("generate_detailed_summary", {})
-        subject_summaries.append(task_result)
+    subject_summaries = [
+        result.get("generate_detailed_summary", {}) for result in results
+    ]
 
     ctx.log(f"process_subjects complete: {len(subject_summaries)} summaries")