Merge pull request #18306 from palazski/main

feat: add mineru as document parser backend with support of both local and managed api
This commit is contained in:
Tim Jaeryang Baek
2025-10-16 10:49:29 -05:00
committed by GitHub
6 changed files with 730 additions and 0 deletions
+42
View File
@@ -466,6 +466,11 @@ async def get_rag_config(request: Request, user=Depends(get_admin_user)):
"DOCUMENT_INTELLIGENCE_ENDPOINT": request.app.state.config.DOCUMENT_INTELLIGENCE_ENDPOINT,
"DOCUMENT_INTELLIGENCE_KEY": request.app.state.config.DOCUMENT_INTELLIGENCE_KEY,
"MISTRAL_OCR_API_KEY": request.app.state.config.MISTRAL_OCR_API_KEY,
# MinerU settings
"MINERU_API_MODE": request.app.state.config.MINERU_API_MODE,
"MINERU_API_URL": request.app.state.config.MINERU_API_URL,
"MINERU_API_KEY": request.app.state.config.MINERU_API_KEY,
"MINERU_PARAMS": request.app.state.config.MINERU_PARAMS,
# Reranking settings
"RAG_RERANKING_MODEL": request.app.state.config.RAG_RERANKING_MODEL,
"RAG_RERANKING_ENGINE": request.app.state.config.RAG_RERANKING_ENGINE,
@@ -647,6 +652,12 @@ class ConfigForm(BaseModel):
DOCUMENT_INTELLIGENCE_KEY: Optional[str] = None
MISTRAL_OCR_API_KEY: Optional[str] = None
# MinerU settings
MINERU_API_MODE: Optional[str] = None
MINERU_API_URL: Optional[str] = None
MINERU_API_KEY: Optional[str] = None
MINERU_PARAMS: Optional[dict] = None
# Reranking settings
RAG_RERANKING_MODEL: Optional[str] = None
RAG_RERANKING_ENGINE: Optional[str] = None
@@ -886,6 +897,28 @@ async def update_rag_config(
else request.app.state.config.MISTRAL_OCR_API_KEY
)
# MinerU settings
request.app.state.config.MINERU_API_MODE = (
form_data.MINERU_API_MODE
if form_data.MINERU_API_MODE is not None
else request.app.state.config.MINERU_API_MODE
)
request.app.state.config.MINERU_API_URL = (
form_data.MINERU_API_URL
if form_data.MINERU_API_URL is not None
else request.app.state.config.MINERU_API_URL
)
request.app.state.config.MINERU_API_KEY = (
form_data.MINERU_API_KEY
if form_data.MINERU_API_KEY is not None
else request.app.state.config.MINERU_API_KEY
)
request.app.state.config.MINERU_PARAMS = (
form_data.MINERU_PARAMS
if form_data.MINERU_PARAMS is not None
else request.app.state.config.MINERU_PARAMS
)
# Reranking settings
if request.app.state.config.RAG_RERANKING_ENGINE == "":
# Unloading the internal reranker and clear VRAM memory
@@ -1150,6 +1183,11 @@ async def update_rag_config(
"DOCUMENT_INTELLIGENCE_ENDPOINT": request.app.state.config.DOCUMENT_INTELLIGENCE_ENDPOINT,
"DOCUMENT_INTELLIGENCE_KEY": request.app.state.config.DOCUMENT_INTELLIGENCE_KEY,
"MISTRAL_OCR_API_KEY": request.app.state.config.MISTRAL_OCR_API_KEY,
# MinerU settings
"MINERU_API_MODE": request.app.state.config.MINERU_API_MODE,
"MINERU_API_URL": request.app.state.config.MINERU_API_URL,
"MINERU_API_KEY": request.app.state.config.MINERU_API_KEY,
"MINERU_PARAMS": request.app.state.config.MINERU_PARAMS,
# Reranking settings
"RAG_RERANKING_MODEL": request.app.state.config.RAG_RERANKING_MODEL,
"RAG_RERANKING_ENGINE": request.app.state.config.RAG_RERANKING_ENGINE,
@@ -1560,6 +1598,10 @@ def process_file(
DOCUMENT_INTELLIGENCE_ENDPOINT=request.app.state.config.DOCUMENT_INTELLIGENCE_ENDPOINT,
DOCUMENT_INTELLIGENCE_KEY=request.app.state.config.DOCUMENT_INTELLIGENCE_KEY,
MISTRAL_OCR_API_KEY=request.app.state.config.MISTRAL_OCR_API_KEY,
MINERU_API_MODE=request.app.state.config.MINERU_API_MODE,
MINERU_API_URL=request.app.state.config.MINERU_API_URL,
MINERU_API_KEY=request.app.state.config.MINERU_API_KEY,
MINERU_PARAMS=request.app.state.config.MINERU_PARAMS,
)
docs = loader.load(
file.filename, file.meta.get("content_type"), file_path