This document outlines the post-processing workflow for memories in the Omi application.
routers/postprocessing.py
/v1/memories/{memory_id}/post-processing
postprocess_memory
function in routers/postprocessing.py
processes the requestdatabase/memories.py
database/users.py
)memories_recordings_bucket
in Google Cloud Storagepostprocessing_audio_bucket
in Google Cloud Storageutils/other/storage.py
fal_whisperx
function in utils/stt/pre_recorded.py
sends audio to FAL.aifal_postprocessing
function in utils/stt/pre_recorded.py
:
TranscriptSegment
objectsget_speech_profile_matching_predictions
in utils/stt/speech_profile.py
:
is_user
and person_id
flagsprocess_memory
in utils/memories/process_memory.py
re-processes memoryprocess_user_emotion
function called asynchronously