Speakr is a personal, self-hosted web application designed for transcribing audio recordings
Python
Fix audio chunking not working after refactor (#140)
Murtaza Nasir6afc2dbfc4b0
diff --git a/src/config/app_config.py b/src/config/app_config.pyindex 8938aaeb..dcc75a37 100644--- a/src/config/app_config.py+++ b/src/config/app_config.py@@ -40,6 +40,9 @@ ENABLE_CHUNKING = os.environ.get('ENABLE_CHUNKING', 'true').lower() == 'true'CHUNK_SIZE_MB = int(os.environ.get('CHUNK_SIZE_MB', '20'))CHUNK_OVERLAP_SECONDS = int(os.environ.get('CHUNK_OVERLAP_SECONDS', '3'))+# Create chunking service at module level so it can be imported by processing.py+chunking_service = AudioChunkingService(CHUNK_SIZE_MB, CHUNK_OVERLAP_SECONDS) if ENABLE_CHUNKING else None+def initialize_config(app):"""Initialize application configuration."""@@ -59,7 +62,7 @@ def initialize_config(app):except Exception as e:app.logger.error(f"Failed to initialize LLM client: {e}")- chunking_service = AudioChunkingService(CHUNK_SIZE_MB, CHUNK_OVERLAP_SECONDS) if ENABLE_CHUNKING else None+ # Use module-level chunking_service (already created above)version = get_version()app.logger.info(f"=== Speakr {version} Starting Up ===")diff --git a/src/tasks/processing.py b/src/tasks/processing.pyindex c96c7533..7481350a 100644--- a/src/tasks/processing.py+++ b/src/tasks/processing.py@@ -18,6 +18,7 @@ import subprocessimport httpxfrom datetime import datetimefrom flask import current_app+from openai import OpenAIfrom src.database import dbfrom src.models import Recording, Tag, Event, TranscriptChunk, SystemSetting, GroupMembership, RecordingTag, InternalShare, SharedRecordingState, User@@ -25,7 +26,7 @@ from src.services.embeddings import process_recording_chunksfrom src.services.llm import is_using_openai_api, call_llm_completion, format_api_error_message, TEXT_MODEL_NAME, client, http_client_no_proxyfrom src.utils import extract_json_object, safe_json_loadsfrom src.audio_chunking import AudioChunkingService, ChunkProcessingError, ChunkingNotSupportedError-from src.config.app_config import ASR_DIARIZE, ASR_BASE_URL, transcription_api_key, transcription_base_url+from src.config.app_config import ASR_DIARIZE, ASR_BASE_URL, transcription_api_key, transcription_base_url, chunking_service, ENABLE_CHUNKINGfrom src.file_exporter import export_recording, ENABLE_AUTO_EXPORT# Configuration for internal sharing@@ -174,20 +175,13 @@ def clean_llm_response(text):USE_ASR_ENDPOINT = os.environ.get('USE_ASR_ENDPOINT', 'false').lower() == 'true'ASR_ENDPOINT = os.environ.get('ASR_ENDPOINT', '')ASR_API_KEY = os.environ.get('ASR_API_KEY', '')-OPENAI_API_KEY = os.environ.get('OPENAI_API_KEY')ENABLE_INQUIRE_MODE = os.environ.get('ENABLE_INQUIRE_MODE', 'false').lower() == 'true'-ENABLE_CHUNKING = os.environ.get('ENABLE_CHUNKING', 'true').lower() == 'true'--# Chunking service will be passed in from app-chunking_service = None-# Import OpenAI client-from openai import OpenAI+# chunking_service, ENABLE_CHUNKING, transcription_api_key, and transcription_base_url+# are imported from src.config.app_config-if USE_ASR_ENDPOINT:- asr_client = OpenAI(base_url=ASR_ENDPOINT, api_key=ASR_API_KEY)-else:- asr_client = OpenAI(api_key=OPENAI_API_KEY)+# Note: OpenAI clients are created inside each transcription function as needed,+# not at module level (matching original pre-refactor behavior)def generate_title_task(app_context, recording_id, will_auto_summarize=False):