refac

backend/open_webui/config.py

@@ -853,6 +853,17 @@ ONEDRIVE_SHAREPOINT_TENANT_ID = os.getenv('ONEDRIVE_SHAREPOINT_TENANT_ID', '')
 # RAG Content Extraction
 CONTENT_EXTRACTION_ENGINE = os.getenv('CONTENT_EXTRACTION_ENGINE', '').lower()
 
+content_extraction_supported_media_mime_types = os.getenv('CONTENT_EXTRACTION_SUPPORTED_MEDIA_MIME_TYPES')
+CONTENT_EXTRACTION_SUPPORTED_MEDIA_MIME_TYPES = (
+    [
+        mime_type.strip()
+        for mime_type in content_extraction_supported_media_mime_types.split(',')
+        if mime_type.strip()
+    ]
+    if content_extraction_supported_media_mime_types is not None
+    else None
+)
+
 DATALAB_MARKER_API_KEY = os.getenv('DATALAB_MARKER_API_KEY', '')
 
 DATALAB_MARKER_API_BASE_URL = os.getenv('DATALAB_MARKER_API_BASE_URL', '')
@@ -1999,6 +2010,8 @@ FOLDER_MAX_FILE_COUNT = os.getenv('FOLDER_MAX_FILE_COUNT', '')
 
 ENABLE_CHANNELS = os.getenv('ENABLE_CHANNELS', 'False').lower() == 'true'
 
+CHANNEL_MODEL_RESPONSE_MODE = os.getenv('CHANNEL_MODEL_RESPONSE_MODE', 'thread')
+
 ENABLE_CALENDAR = os.getenv('ENABLE_CALENDAR', 'True').lower() == 'true'
 
 ENABLE_AUTOMATIONS = os.getenv('ENABLE_AUTOMATIONS', 'True').lower() == 'true'
@@ -2817,6 +2830,7 @@ DEFAULT_CONFIG = {
     'onedrive.sharepoint_url': ONEDRIVE_SHAREPOINT_URL,
     'onedrive.sharepoint_tenant_id': ONEDRIVE_SHAREPOINT_TENANT_ID,
     'rag.content_extraction_engine': CONTENT_EXTRACTION_ENGINE,
+    'rag.content_extraction.supported_media_mime_types': CONTENT_EXTRACTION_SUPPORTED_MEDIA_MIME_TYPES,
     'rag.datalab_marker_api_key': DATALAB_MARKER_API_KEY,
     'rag.datalab_marker_api_base_url': DATALAB_MARKER_API_BASE_URL,
     'rag.datalab_marker_additional_config': DATALAB_MARKER_ADDITIONAL_CONFIG,
@@ -3049,6 +3063,7 @@ DEFAULT_CONFIG = {
     'folders.enable': ENABLE_FOLDERS,
     'folders.max_file_count': FOLDER_MAX_FILE_COUNT,
     'channels.enable': ENABLE_CHANNELS,
+    'channels.model_response_mode': CHANNEL_MODEL_RESPONSE_MODE,
     'calendar.enable': ENABLE_CALENDAR,
     'automations.enable': ENABLE_AUTOMATIONS,
     'subagents.enable': ENABLE_SUBAGENTS,

backend/open_webui/routers/auths.py

@@ -117,6 +117,7 @@ ADMIN_CONFIG_KEYS = {
     'AUTOMATION_MIN_INTERVAL': 'automations.min_interval',
     'ENABLE_AUTOMATIONS': 'automations.enable',
     'ENABLE_CHANNELS': 'channels.enable',
+    'CHANNEL_MODEL_RESPONSE_MODE': 'channels.model_response_mode',
     'ENABLE_CALENDAR': 'calendar.enable',
     'ENABLE_MEMORIES': 'memories.enable',
     'ENABLE_MEMORY_SYSTEM_CONTEXT': 'memories.system_context.enable',
@@ -1197,6 +1198,7 @@ class AdminConfig(BaseModel):
     AUTOMATION_MIN_INTERVAL: int | str | None = None
     ENABLE_AUTOMATIONS: bool
     ENABLE_CHANNELS: bool
+    CHANNEL_MODEL_RESPONSE_MODE: str = 'thread'
     ENABLE_CALENDAR: bool
     ENABLE_MEMORIES: bool
     ENABLE_MEMORY_SYSTEM_CONTEXT: bool
@@ -1220,6 +1222,9 @@ async def update_admin_config(request: Request, form_data: AdminConfig, user=Dep
     if form_data.DEFAULT_USER_ROLE not in ['pending', 'user', 'admin']:
         updates.pop('ui.default_user_role', None)
 
+    if form_data.CHANNEL_MODEL_RESPONSE_MODE not in ['thread', 'channel']:
+        updates.pop('channels.model_response_mode', None)
+
     pattern = r'^(-1|0|(-?\d+(\.\d+)?)(ms|s|m|h|d|w))$'
 
     # Check if the input string matches the pattern

backend/open_webui/routers/channels.py

@@ -987,13 +987,18 @@ async def model_response_handler(request, channel, message, user, db=None):
                         db=db,
                     )
                 )[::-1]
+                response_parent_id = (
+                    message.parent_id
+                    if message.parent_id
+                    else (message.id if await Config.get('channels.model_response_mode', 'thread') == 'thread' else None)
+                )
 
                 response_message, channel = await new_message_handler(
                     request,
                     channel.id,
                     MessageForm(
                         **{
-                            'parent_id': (message.parent_id if message.parent_id else message.id),
+                            'parent_id': response_parent_id,
                             'content': f'',
                             'data': {},
                             'meta': {

backend/open_webui/routers/files.py

@@ -108,6 +108,23 @@ def _cleanup_local_cache(file_path: str) -> None:
         log.warning(f'Failed to clean up local cache for {file_path}: {e}')
 
 
+def _matches_configured_mime_type(supported: list[str] | str, content_type: str) -> bool:
+    if isinstance(supported, str):
+        supported = supported.split(',')
+    supported = [item.strip() for item in (supported or []) if item.strip()]
+    if not supported:
+        return False
+    return bool(strict_match_mime_type(supported, content_type))
+
+
+def _media_supported_for_extraction(
+    content_extraction_engine: str | None, supported: list[str] | str | None, content_type: str
+) -> bool:
+    if supported is None:
+        return content_extraction_engine == 'external'
+    return bool(content_extraction_engine and _matches_configured_mime_type(supported, content_type))
+
+
 async def process_uploaded_file(
     request,
     file,
@@ -127,6 +144,10 @@ async def process_uploaded_file(
                     content_type = 'text/plain'
 
             stt_supported = await Config.get('audio.stt.supported_content_types', [])
+            content_extraction_engine = await Config.get('rag.content_extraction_engine')
+            content_extraction_supported_media_mime_types = await Config.get(
+                'rag.content_extraction.supported_media_mime_types'
+            )
 
             if content_type and strict_match_mime_type(stt_supported, content_type):
                 # Audio / STT-supported files → transcribe then index
@@ -147,9 +168,10 @@ async def process_uploaded_file(
             elif (
                 content_type
                 and content_type.startswith(('image/', 'video/'))
-                and await Config.get('rag.content_extraction_engine') != 'external'
+                and not _media_supported_for_extraction(
+                    content_extraction_engine, content_extraction_supported_media_mime_types, content_type
+                )
             ):
-                # Media files without an external extraction engine
                 if content_type.startswith('video/'):
                     # Videos are stored as-is for downstream multimodal
                     # processing (Tools, vision models). Attempting text
@@ -165,7 +187,8 @@ async def process_uploaded_file(
                     raise Exception(f'File type {content_type} is not supported for processing')
 
             else:
-                # Documents, or any file when an external engine is configured
+                # Documents, or media files explicitly enabled for the
+                # configured content extraction engine.
                 if not content_type:
                     log.info(f'File type {file.content_type} is not provided, but trying to process anyway')
                 await process_file(

backend/open_webui/routers/retrieval.py

@@ -265,6 +265,7 @@ RETRIEVAL_CONFIG_KEYS = {
     'CHUNK_MIN_SIZE_TARGET': 'rag.chunk_min_size_target',
     'CHUNK_OVERLAP': 'rag.chunk_overlap',
     'CHUNK_SIZE': 'rag.chunk_size',
+    'CONTENT_EXTRACTION_SUPPORTED_MEDIA_MIME_TYPES': 'rag.content_extraction.supported_media_mime_types',
     'CONTENT_EXTRACTION_ENGINE': 'rag.content_extraction_engine',
     'DATALAB_MARKER_ADDITIONAL_CONFIG': 'rag.datalab_marker_additional_config',
     'DATALAB_MARKER_API_BASE_URL': 'rag.datalab_marker_api_base_url',
@@ -630,6 +631,7 @@ async def get_rag_config(request: Request, user=Depends(get_admin_user)):
         'HYBRID_BM25_WEIGHT': config.HYBRID_BM25_WEIGHT,
         # Content extraction settings
         'CONTENT_EXTRACTION_ENGINE': config.CONTENT_EXTRACTION_ENGINE,
+        'CONTENT_EXTRACTION_SUPPORTED_MEDIA_MIME_TYPES': config.CONTENT_EXTRACTION_SUPPORTED_MEDIA_MIME_TYPES,
         'PDF_EXTRACT_IMAGES': config.PDF_EXTRACT_IMAGES,
         'PDF_LOADER_MODE': config.PDF_LOADER_MODE,
         'DATALAB_MARKER_API_KEY': config.DATALAB_MARKER_API_KEY,
@@ -861,6 +863,7 @@ class ConfigForm(BaseModel):
 
     # Content extraction settings
     CONTENT_EXTRACTION_ENGINE: str | None = None
+    CONTENT_EXTRACTION_SUPPORTED_MEDIA_MIME_TYPES: list[str] | None = None
     PDF_EXTRACT_IMAGES: bool | None = None
     PDF_LOADER_MODE: str | None = None
 
@@ -973,6 +976,11 @@ async def update_rag_config(request: Request, form_data: ConfigForm, user=Depend
         if form_data.CONTENT_EXTRACTION_ENGINE is not None
         else config.CONTENT_EXTRACTION_ENGINE
     )
+    config.CONTENT_EXTRACTION_SUPPORTED_MEDIA_MIME_TYPES = (
+        form_data.CONTENT_EXTRACTION_SUPPORTED_MEDIA_MIME_TYPES
+        if form_data.CONTENT_EXTRACTION_SUPPORTED_MEDIA_MIME_TYPES is not None
+        else config.CONTENT_EXTRACTION_SUPPORTED_MEDIA_MIME_TYPES
+    )
     config.PDF_EXTRACT_IMAGES = (
         form_data.PDF_EXTRACT_IMAGES if form_data.PDF_EXTRACT_IMAGES is not None else config.PDF_EXTRACT_IMAGES
     )
@@ -1333,6 +1341,7 @@ async def update_rag_config(request: Request, form_data: ConfigForm, user=Depend
         'HYBRID_BM25_WEIGHT': config.HYBRID_BM25_WEIGHT,
         # Content extraction settings
         'CONTENT_EXTRACTION_ENGINE': config.CONTENT_EXTRACTION_ENGINE,
+        'CONTENT_EXTRACTION_SUPPORTED_MEDIA_MIME_TYPES': config.CONTENT_EXTRACTION_SUPPORTED_MEDIA_MIME_TYPES,
         'PDF_EXTRACT_IMAGES': config.PDF_EXTRACT_IMAGES,
         'PDF_LOADER_MODE': config.PDF_LOADER_MODE,
         'DATALAB_MARKER_API_KEY': config.DATALAB_MARKER_API_KEY,