feat: add segment_post_processor hook for external plugins

Add a minimal, non-breaking hook to WhisperLive that allows external
projects to post-process transcription segments before they are sent
to the client.

Changes:
- ServeClientBase: add segment_post_processor attribute (default None)
- ServeClientBase.send_transcription_to_client: apply post_processor
  per-segment with error handling (falls back to original segment)
- TranscriptionServer: add segment_post_processor parameter to run()
  and wire it to each client on creation

This enables downstream projects to plug in custom processing
(e.g. formatting, PII redaction, diarization tagging) without
modifying WhisperLive core code.
This commit is contained in:
Aaron Boxer
2026-05-11 19:37:10 -04:00
committed by Aaron Boxer
parent 4e31f8c61b
commit c028c4b584
2 changed files with 34 additions and 1 deletions
+21
View File
@@ -63,6 +63,13 @@ class ServeClientBase(object):
self.end_time_for_same_output = None
self.translation_queue = translation_queue
# Optional post-processing callable for segments.
# If set, called with a segment dict and must return a segment dict.
# Allows external projects to plug in custom post-processing
# (e.g. PII redaction, formatting, diarization) without modifying
# WhisperLive's core code.
self.segment_post_processor = None
# threading
self.lock = threading.Lock()
@@ -254,9 +261,23 @@ class ServeClientBase(object):
This method formats the transcription segments into a JSON object and attempts to send
this object to the client. If an error occurs during the send operation, it logs the error.
If a ``segment_post_processor`` callable is set, each segment is passed through it
before sending. The callable receives a segment dict and must return a segment dict.
Returns:
segments (list): A list of transcription segments to be sent to the client.
"""
if self.segment_post_processor is not None:
processed = []
for seg in segments:
try:
result = self.segment_post_processor(seg)
processed.append(result if result is not None else seg)
except Exception as e:
logging.error(f"[ERROR]: segment_post_processor failed: {e}")
processed.append(seg)
segments = processed
try:
self.websocket.send(
json.dumps({