openrag 0.5.0.dev34__tar.gz → 0.5.0.dev36__tar.gz

{openrag-0.5.0.dev34/src/openrag.egg-info → openrag-0.5.0.dev36}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: openrag
-Version: 0.5.0.dev34
+Version: 0.5.0.dev36
 Summary: OpenRAG is a comprehensive Retrieval-Augmented Generation platform that enables intelligent document search and AI-powered conversations.
 Classifier: Development Status :: 4 - Beta
 Classifier: Environment :: Console

{openrag-0.5.0.dev34 → openrag-0.5.0.dev36}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "openrag"
-version = "0.5.0.dev34"
+version = "0.5.0.dev36"
 description = "OpenRAG is a comprehensive Retrieval-Augmented Generation platform that enables intelligent document search and AI-powered conversations."
 readme = "README.md"
 requires-python = ">=3.13"

{openrag-0.5.0.dev34 → openrag-0.5.0.dev36}/src/api/langflow_files.py

@@ -1,43 +1,43 @@
 import json
 import os
 import tempfile
-from typing import List, Optional
 
 from fastapi import Depends, File, Form, UploadFile
-from pydantic import BaseModel
 from fastapi.responses import JSONResponse
+from pydantic import BaseModel
 
 from dependencies import (
+    get_current_user,
     get_langflow_file_service,
+    get_optional_user,
     get_session_manager,
     get_task_service,
-    get_current_user,
-    get_optional_user,
 )
 from session_manager import User
+from utils.file_utils import langflow_safe_filename_and_mimetype
 from utils.logging_config import get_logger
 
 logger = get_logger(__name__)
 
 
 class RunIngestionBody(BaseModel):
-    file_paths: List[str] = []
-    file_ids: Optional[List[str]] = None
-    file_metadata: Optional[List[dict]] = None
-    session_id: Optional[str] = None
-    tweaks: Optional[dict] = None
-    settings: Optional[dict] = None
+    file_paths: list[str] = []
+    file_ids: list[str] | None = None
+    file_metadata: list[dict] | None = None
+    session_id: str | None = None
+    tweaks: dict | None = None
+    settings: dict | None = None
 
 
 class DeleteFilesBody(BaseModel):
-    file_ids: List[str]
+    file_ids: list[str]
 
 
 async def upload_user_file(
     file: UploadFile = File(...),
     langflow_file_service=Depends(get_langflow_file_service),
     session_manager=Depends(get_session_manager),
-    user: Optional[User] = Depends(get_optional_user),
+    user: User | None = Depends(get_optional_user),
 ):
     """Upload a file to Langflow's Files API"""
     try:
@@ -45,11 +45,11 @@ async def upload_user_file(
         logger.debug("Processing file", filename=file.filename, size=file.size)
 
         content = await file.read()
-        file_tuple = (
-            file.filename,
-            content,
-            file.content_type or "application/octet-stream",
+        # Langflow's docling chokes on text/plain — rename .txt -> .md.
+        upload_filename, upload_mimetype = langflow_safe_filename_and_mimetype(
+            file.filename, file.content_type
         )
+        file_tuple = (upload_filename, content, upload_mimetype)
 
         jwt_token = user.jwt_token if user else session_manager.get_effective_jwt_token(None, None)
 
@@ -60,6 +60,7 @@ async def upload_user_file(
     except Exception as e:
         logger.error("upload_user_file endpoint failed", error_type=type(e).__name__, error=str(e))
         import traceback
+
         logger.error("Full traceback", traceback=traceback.format_exc())
         return JSONResponse({"error": str(e)}, status_code=500)
 
@@ -111,6 +112,7 @@ async def run_ingestion(
 
     if jwt_token:
         from auth_context import set_auth_context
+
         set_auth_context(user.user_id, jwt_token)
 
     try:
@@ -132,9 +134,9 @@ async def run_ingestion(
 
 async def upload_and_ingest_user_file(
     file: UploadFile = File(...),
-    session_id: Optional[str] = Form(None),
-    settings_json: Optional[str] = Form(None, alias="settings"),
-    tweaks_json: Optional[str] = Form(None, alias="tweaks"),
+    session_id: str | None = Form(None),
+    settings_json: str | None = Form(None, alias="settings"),
+    tweaks_json: str | None = Form(None, alias="tweaks"),
     delete_after_ingest: str = Form("true"),
     langflow_file_service=Depends(get_langflow_file_service),
     session_manager=Depends(get_session_manager),
@@ -193,6 +195,7 @@ async def upload_and_ingest_user_file(
             )
         except Exception:
             from utils.file_utils import safe_unlink
+
             safe_unlink(temp_path)
             raise
 
@@ -221,4 +224,4 @@ async def delete_user_files(
             "errors": errors,
         },
         status_code=status,
-    )
+    )

{openrag-0.5.0.dev34 → openrag-0.5.0.dev36}/src/connectors/langflow_connector_service.py

@@ -3,11 +3,15 @@ from typing import Any, Dict, List, Optional
 # Create custom processor for connector files using Langflow
 from models.processors import LangflowConnectorFileProcessor
 from services.langflow_file_service import LangflowFileService
+from utils.file_utils import (
+    clean_connector_filename,
+    get_file_extension,
+    langflow_safe_filename_and_mimetype,
+)
 from utils.logging_config import get_logger
 
 from .base import BaseConnector, ConnectorDocument
 from .connection_manager import ConnectionManager
-from utils.file_utils import get_file_extension, clean_connector_filename
 
 logger = get_logger(__name__)
 
@@ -73,11 +77,15 @@ class LangflowConnectorService:
             
             # Clean filename and ensure we don't add a double extension
             processed_filename = clean_connector_filename(document.filename, document.mimetype)
+            # Langflow's docling chokes on text/plain — rename .txt -> .md.
+            processed_filename, processed_mimetype = langflow_safe_filename_and_mimetype(
+                processed_filename, document.mimetype
+            )
 
             file_tuple = (
                 processed_filename,
                 content,
-                document.mimetype or "application/octet-stream",
+                processed_mimetype,
             )
 
             # Step 0: Delete existing chunks for this file before re-ingesting

{openrag-0.5.0.dev34 → openrag-0.5.0.dev36}/src/models/processors.py

@@ -5,6 +5,7 @@ from utils.file_utils import (
     get_file_extension,
     clean_connector_filename,
     get_filename_aliases,
+    langflow_safe_filename_and_mimetype,
 )
 
 logger = get_logger(__name__)
@@ -868,14 +869,10 @@ class LangflowFileProcessor(TaskProcessor):
             if not content_type:
                 content_type = 'application/octet-stream'
 
-            # Rename .txt to .md for Langflow compatibility
-            # Langflow has issues processing text/plain files
-            langflow_filename = original_filename
-            if original_filename.lower().endswith('.txt'):
-                langflow_filename = original_filename[:-4] + '.md'
-                content_type = 'text/markdown'
-                logger.debug(f"Renamed {original_filename} to {langflow_filename} for Langflow")
-
+            # Langflow's docling chokes on text/plain — rename .txt -> .md.
+            langflow_filename, content_type = langflow_safe_filename_and_mimetype(
+                original_filename, content_type
+            )
             file_tuple = (langflow_filename, content, content_type)
 
             # Get JWT token using same logic as DocumentFileProcessor

{openrag-0.5.0.dev34 → openrag-0.5.0.dev36/src/openrag.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: openrag
-Version: 0.5.0.dev34
+Version: 0.5.0.dev36
 Summary: OpenRAG is a comprehensive Retrieval-Augmented Generation platform that enables intelligent document search and AI-powered conversations.
 Classifier: Development Status :: 4 - Beta
 Classifier: Environment :: Console

{openrag-0.5.0.dev34 → openrag-0.5.0.dev36}/src/utils/file_utils.py

@@ -101,6 +101,25 @@ def clean_connector_filename(filename: str, mimetype: str) -> str:
     return clean_name
 
 
+def langflow_safe_filename_and_mimetype(filename: str, mimetype: str | None) -> tuple[str, str]:
+    """Apply the .txt -> .md workaround for Langflow ingestion.
+
+    Langflow's docling component fails on text/plain (see commit f6b9fe0).
+    The workaround is to rename .txt files to .md with text/markdown before
+    handing the file to Langflow. Centralised here so every Langflow-upload
+    site applies it identically — local user uploads, connector-driven
+    ingestion, and the direct upload route.
+
+    Returns the (possibly rewritten) (filename, mimetype) pair. Pass-through
+    for all other extensions; falls back to application/octet-stream when
+    mimetype is missing.
+    """
+    safe_mimetype = mimetype or "application/octet-stream"
+    if filename and filename.lower().endswith(".txt"):
+        return filename[:-4] + ".md", "text/markdown"
+    return filename, safe_mimetype
+
+
 def get_filename_aliases(filename: str) -> list[str]:
     """Return equivalent filename variants used by ingestion/indexing.