self-review round

server/reflector/hatchet/utils.py

@@ -1,17 +0,0 @@
-"""
-Hatchet workflow utilities.
-
-Shared helpers for Hatchet task implementations.
-"""
-
-
-def to_dict(output) -> dict:
-    """Convert task output to dict, handling both dict and Pydantic model returns.
-
-    Hatchet SDK can return task outputs as either raw dicts or Pydantic models
-    depending on serialization context. This normalizes the output for consistent
-    downstream processing.
-    """
-    if isinstance(output, dict):
-        return output
-    return output.model_dump()

server/reflector/hatchet/workflows/diarization_pipeline.py

@@ -27,7 +27,6 @@ from reflector.hatchet.broadcast import (
     set_status_and_broadcast,
 )
 from reflector.hatchet.client import HatchetClientManager
-from reflector.hatchet.utils import to_dict
 from reflector.hatchet.workflows.models import (
     ConsentResult,
     FinalizeResult,
@@ -69,7 +68,7 @@ from reflector.utils.daily import (
     filter_cam_audio_tracks,
     parse_daily_recording_filename,
 )
-from reflector.utils.string import NonEmptyString
+from reflector.utils.string import NonEmptyString, assert_non_none_and_non_empty
 from reflector.zulip import post_transcript_notification
 
 
@@ -212,8 +211,8 @@ async def get_participants(input: PipelineInput, ctx: Context) -> ParticipantsRe
     """Fetch participant list from Daily.co API and update transcript in database."""
     ctx.log(f"get_participants: transcript_id={input.transcript_id}")
 
-    recording_data = to_dict(ctx.task_output(get_recording))
+    recording = ctx.task_output(get_recording)
-    mtg_session_id = recording_data.get("mtg_session_id")
+    mtg_session_id = recording.mtg_session_id
 
     async with fresh_db_connection():
         from reflector.db.transcripts import (  # noqa: PLC0415
@@ -233,15 +232,14 @@ async def get_participants(input: PipelineInput, ctx: Context) -> ParticipantsRe
                 },
             )
 
-        if not mtg_session_id or not settings.DAILY_API_KEY:
+        mtg_session_id = assert_non_none_and_non_empty(
-            return ParticipantsResult(
+            mtg_session_id, "mtg_session_id is required"
-                participants=[],
+        )
-                num_tracks=len(input.tracks),
+        daily_api_key = assert_non_none_and_non_empty(
-                source_language=transcript.source_language if transcript else "en",
+            settings.DAILY_API_KEY, "DAILY_API_KEY is required"
-                target_language=transcript.target_language if transcript else "en",
         )
 
-        async with DailyApiClient(api_key=settings.DAILY_API_KEY) as client:
+        async with DailyApiClient(api_key=daily_api_key) as client:
             participants = await client.get_meeting_participants(mtg_session_id)
 
         id_to_name = {}
@@ -302,8 +300,8 @@ async def process_tracks(input: PipelineInput, ctx: Context) -> ProcessTracksRes
     """Spawn child workflows for each track (dynamic fan-out)."""
     ctx.log(f"process_tracks: spawning {len(input.tracks)} track workflows")
 
-    participants_data = to_dict(ctx.task_output(get_participants))
+    participants_result = ctx.task_output(get_participants)
-    source_language = participants_data.get("source_language", "en")
+    source_language = participants_result.source_language
 
     child_coroutines = [
         track_workflow.aio_run(
@@ -320,7 +318,7 @@ async def process_tracks(input: PipelineInput, ctx: Context) -> ProcessTracksRes
 
     results = await asyncio.gather(*child_coroutines)
 
-    target_language = participants_data.get("target_language", "en")
+    target_language = participants_result.target_language
 
     # Collect results from each track (don't mutate lists while iterating)
     track_words = []
@@ -371,31 +369,23 @@ async def mixdown_tracks(input: PipelineInput, ctx: Context) -> MixdownResult:
     """Mix all padded tracks into single audio file using PyAV (same as Celery)."""
     ctx.log("mixdown_tracks: mixing padded tracks into single audio file")
 
-    track_data = to_dict(ctx.task_output(process_tracks))
+    track_result = ctx.task_output(process_tracks)
-    padded_tracks_data = track_data.get("padded_tracks", [])
+    padded_tracks = track_result.padded_tracks
 
-    if not padded_tracks_data:
+    if not padded_tracks:
         raise ValueError("No padded tracks to mixdown")
 
     storage = _spawn_storage()
 
     # Presign URLs on demand (avoids stale URLs on workflow replay)
     padded_urls = []
-    for track_info in padded_tracks_data:
+    for track_info in padded_tracks:
-        # Handle both dict (from to_dict) and PaddedTrackInfo
+        if track_info.key:
-        if isinstance(track_info, dict):
-            key = track_info.get("key")
-            bucket = track_info.get("bucket_name")
-        else:
-            key = track_info.key
-            bucket = track_info.bucket_name
-
-        if key:
             url = await storage.get_file_url(
-                key,
+                track_info.key,
                 operation="get_object",
                 expires_in=PRESIGNED_URL_EXPIRATION_SECONDS,
-                bucket=bucket,
+                bucket=track_info.bucket_name,
             )
             padded_urls.append(url)
 
@@ -465,8 +455,8 @@ async def generate_waveform(input: PipelineInput, ctx: Context) -> WaveformResul
     )
 
     # Cleanup temporary padded S3 files (deferred until after mixdown)
-    track_data = to_dict(ctx.task_output(process_tracks))
+    track_result = ctx.task_output(process_tracks)
-    created_padded_files = track_data.get("created_padded_files", [])
+    created_padded_files = track_result.created_padded_files
     if created_padded_files:
         ctx.log(f"Cleaning up {len(created_padded_files)} temporary S3 files")
         storage = _spawn_storage()
@@ -483,8 +473,8 @@ async def generate_waveform(input: PipelineInput, ctx: Context) -> WaveformResul
                     error=str(result),
                 )
 
-    mixdown_data = to_dict(ctx.task_output(mixdown_tracks))
+    mixdown_result = ctx.task_output(mixdown_tracks)
-    audio_key = mixdown_data.get("audio_key")
+    audio_key = mixdown_result.audio_key
 
     storage = _spawn_storage()
     audio_url = await storage.get_file_url(
@@ -532,9 +522,9 @@ async def detect_topics(input: PipelineInput, ctx: Context) -> TopicsResult:
     """Detect topics using LLM and save to database (matches Celery on_topic callback)."""
     ctx.log("detect_topics: analyzing transcript for topics")
 
-    track_data = to_dict(ctx.task_output(process_tracks))
+    track_result = ctx.task_output(process_tracks)
-    words = track_data.get("all_words", [])
+    words = track_result.all_words
-    target_language = track_data.get("target_language", "en")
+    target_language = track_result.target_language
 
     from reflector.db.transcripts import (  # noqa: PLC0415
         TranscriptTopic,
@@ -589,8 +579,8 @@ async def generate_title(input: PipelineInput, ctx: Context) -> TitleResult:
     """Generate meeting title using LLM and save to database (matches Celery on_title callback)."""
     ctx.log("generate_title: generating title from topics")
 
-    topics_data = to_dict(ctx.task_output(detect_topics))
+    topics_result = ctx.task_output(detect_topics)
-    topics = topics_data.get("topics", [])
+    topics = topics_result.topics
 
     from reflector.db.transcripts import (  # noqa: PLC0415
         TranscriptFinalTitle,
@@ -638,8 +628,8 @@ async def generate_summary(input: PipelineInput, ctx: Context) -> SummaryResult:
     """Generate meeting summary using LLM and save to database (matches Celery callbacks)."""
     ctx.log("generate_summary: generating long and short summaries")
 
-    topics_data = to_dict(ctx.task_output(detect_topics))
+    topics_result = ctx.task_output(detect_topics)
-    topics = topics_data.get("topics", [])
+    topics = topics_result.topics
 
     from reflector.db.transcripts import (  # noqa: PLC0415
         TranscriptFinalLongSummary,
@@ -718,11 +708,11 @@ async def finalize(input: PipelineInput, ctx: Context) -> FinalizeResult:
     """
     ctx.log("finalize: saving transcript and setting status to 'ended'")
 
-    mixdown_data = to_dict(ctx.task_output(mixdown_tracks))
+    mixdown_result = ctx.task_output(mixdown_tracks)
-    track_data = to_dict(ctx.task_output(process_tracks))
+    track_result = ctx.task_output(process_tracks)
 
-    duration = mixdown_data.get("duration", 0)
+    duration = mixdown_result.duration
-    all_words = track_data.get("all_words", [])
+    all_words = track_result.all_words
 
     async with fresh_db_connection():
         from reflector.db.transcripts import (  # noqa: PLC0415

server/reflector/hatchet/workflows/track_processing.py

@@ -23,7 +23,6 @@ from hatchet_sdk import Context
 from pydantic import BaseModel
 
 from reflector.hatchet.client import HatchetClientManager
-from reflector.hatchet.utils import to_dict
 from reflector.hatchet.workflows.models import PadTrackResult, TranscribeTrackResult
 from reflector.logger import logger
 from reflector.utils.audio_constants import PRESIGNED_URL_EXPIRATION_SECONDS
@@ -166,9 +165,9 @@ async def transcribe_track(input: TrackInput, ctx: Context) -> TranscribeTrackRe
     )
 
     try:
-        pad_result = to_dict(ctx.task_output(pad_track))
+        pad_result = ctx.task_output(pad_track)
-        padded_key = pad_result.get("padded_key")
+        padded_key = pad_result.padded_key
-        bucket_name = pad_result.get("bucket_name")
+        bucket_name = pad_result.bucket_name
 
         if not padded_key:
             raise ValueError("Missing padded_key from pad_track")

server/reflector/utils/string.py

@@ -2,6 +2,8 @@ from typing import Annotated, TypeVar
 
 from pydantic import Field, TypeAdapter, constr
 
+from reflector.utils.common import assert_not_none
+
 NonEmptyStringBase = constr(min_length=1, strip_whitespace=False)
 NonEmptyString = Annotated[
     NonEmptyStringBase,
@@ -30,3 +32,11 @@ def assert_equal[T](s1: T, s2: T) -> T:
     if s1 != s2:
         raise ValueError(f"assert_equal: {s1} != {s2}")
     return s1
+
+
+def assert_non_none_and_non_empty(
+    value: str | None, error: str | None = None
+) -> NonEmptyString:
+    return parse_non_empty_string(
+        assert_not_none(value, error or "Value is None"), error
+    )