OpenWebUI/open-webui
1
0
Fork 0
mirror of https://github.com/open-webui/open-webui synced 2025-06-26 18:26:48 +00:00
Code Issues Actions Packages Projects Releases Wiki Activity

feat: docling do picture description support

Browse Source
This commit is contained in:
Timothy Jaeryang Baek
2025-05-14 21:26:49 +04:00
parent 6e8481e157
commit de70d0cb64
5 changed files with 70 additions and 29 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
backend/open_webui/config.py
View File

@@ -1849,6 +1849,12 @@ DOCLING_OCR_LANG = PersistentConfig(
os.getenv("DOCLING_OCR_LANG", "eng,fra,deu,spa"),
)
DOCLING_DO_PICTURE_DESCRIPTION = PersistentConfig(
"DOCLING_DO_PICTURE_DESCRIPTION",
"rag.docling_do_picture_description",
os.getenv("DOCLING_DO_PICTURE_DESCRIPTION", "False").lower() == "true",
)
DOCUMENT_INTELLIGENCE_ENDPOINT = PersistentConfig(
"DOCUMENT_INTELLIGENCE_ENDPOINT",
"rag.document_intelligence_endpoint",

2
backend/open_webui/main.py
View File

@@ -210,6 +210,7 @@ from open_webui.config import (
DOCLING_SERVER_URL,
DOCLING_OCR_ENGINE,
DOCLING_OCR_LANG,
DOCLING_DO_PICTURE_DESCRIPTION,
DOCUMENT_INTELLIGENCE_ENDPOINT,
DOCUMENT_INTELLIGENCE_KEY,
MISTRAL_OCR_API_KEY,
@@ -645,6 +646,7 @@ app.state.config.TIKA_SERVER_URL = TIKA_SERVER_URL
app.state.config.DOCLING_SERVER_URL = DOCLING_SERVER_URL
app.state.config.DOCLING_OCR_ENGINE = DOCLING_OCR_ENGINE
app.state.config.DOCLING_OCR_LANG = DOCLING_OCR_LANG
app.state.config.DOCLING_DO_PICTURE_DESCRIPTION = DOCLING_DO_PICTURE_DESCRIPTION
app.state.config.DOCUMENT_INTELLIGENCE_ENDPOINT = DOCUMENT_INTELLIGENCE_ENDPOINT
app.state.config.DOCUMENT_INTELLIGENCE_KEY = DOCUMENT_INTELLIGENCE_KEY
app.state.config.MISTRAL_OCR_API_KEY = MISTRAL_OCR_API_KEY

35
backend/open_webui/retrieval/loaders/main.py
View File

@@ -126,14 +126,12 @@ class TikaLoader:
class DoclingLoader:
def __init__(
self, url, file_path=None, mime_type=None, ocr_engine=None, ocr_lang=None
):
def __init__(self, url, file_path=None, mime_type=None, params=None):
self.url = url.rstrip("/")
self.file_path = file_path
self.mime_type = mime_type
self.ocr_engine = ocr_engine
self.ocr_lang = ocr_lang
self.params = params or {}
def load(self) -> list[Document]:
with open(self.file_path, "rb") as f:
@@ -150,11 +148,19 @@ class DoclingLoader:
"table_mode": "accurate",
}
if self.ocr_engine and self.ocr_lang:
params["ocr_engine"] = self.ocr_engine
params["ocr_lang"] = [
lang.strip() for lang in self.ocr_lang.split(",") if lang.strip()
]
if self.params:
if self.params.get("do_picture_classification"):
params["do_picture_classification"] = self.params.get(
"do_picture_classification"
)
if self.params.get("ocr_engine") and self.params.get("ocr_lang"):
params["ocr_engine"] = self.params.get("ocr_engine")
params["ocr_lang"] = [
lang.strip()
for lang in self.params.get("ocr_lang").split(",")
if lang.strip()
]
endpoint = f"{self.url}/v1alpha/convert/file"
r = requests.post(endpoint, files=files, data=params)
@@ -225,8 +231,13 @@ class Loader:
url=self.kwargs.get("DOCLING_SERVER_URL"),
file_path=file_path,
mime_type=file_content_type,
ocr_engine=self.kwargs.get("DOCLING_OCR_ENGINE"),
ocr_lang=self.kwargs.get("DOCLING_OCR_LANG"),
params={
"ocr_engine": self.kwargs.get("DOCLING_OCR_ENGINE"),
"ocr_lang": self.kwargs.get("DOCLING_OCR_LANG"),
"do_picture_classification": self.kwargs.get(
"DOCLING_DO_PICTURE_DESCRIPTION"
),
},
)
elif (
self.engine == "document_intelligence"

11
backend/open_webui/routers/retrieval.py
View File

@@ -356,6 +356,7 @@ async def get_rag_config(request: Request, user=Depends(get_admin_user)):
"DOCLING_SERVER_URL": request.app.state.config.DOCLING_SERVER_URL,
"DOCLING_OCR_ENGINE": request.app.state.config.DOCLING_OCR_ENGINE,
"DOCLING_OCR_LANG": request.app.state.config.DOCLING_OCR_LANG,
"DOCLING_DO_PICTURE_DESCRIPTION": request.app.state.config.DOCLING_DO_PICTURE_DESCRIPTION,
"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,
@@ -496,6 +497,7 @@ class ConfigForm(BaseModel):
DOCLING_SERVER_URL: Optional[str] = None
DOCLING_OCR_ENGINE: Optional[str] = None
DOCLING_OCR_LANG: Optional[str] = None
DOCLING_DO_PICTURE_DESCRIPTION: Optional[bool] = None
DOCUMENT_INTELLIGENCE_ENDPOINT: Optional[str] = None
DOCUMENT_INTELLIGENCE_KEY: Optional[str] = None
MISTRAL_OCR_API_KEY: Optional[str] = None
@@ -601,6 +603,13 @@ async def update_rag_config(
if form_data.DOCLING_OCR_LANG is not None
else request.app.state.config.DOCLING_OCR_LANG
)
request.app.state.config.DOCLING_DO_PICTURE_DESCRIPTION = (
form_data.DOCLING_DO_PICTURE_DESCRIPTION
if form_data.DOCLING_DO_PICTURE_DESCRIPTION is not None
else request.app.state.config.DOCLING_DO_PICTURE_DESCRIPTION
)
request.app.state.config.DOCUMENT_INTELLIGENCE_ENDPOINT = (
form_data.DOCUMENT_INTELLIGENCE_ENDPOINT
if form_data.DOCUMENT_INTELLIGENCE_ENDPOINT is not None
@@ -813,6 +822,7 @@ async def update_rag_config(
"DOCLING_SERVER_URL": request.app.state.config.DOCLING_SERVER_URL,
"DOCLING_OCR_ENGINE": request.app.state.config.DOCLING_OCR_ENGINE,
"DOCLING_OCR_LANG": request.app.state.config.DOCLING_OCR_LANG,
"DOCLING_DO_PICTURE_DESCRIPTION": request.app.state.config.DOCLING_DO_PICTURE_DESCRIPTION,
"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,
@@ -1133,6 +1143,7 @@ def process_file(
DOCLING_SERVER_URL=request.app.state.config.DOCLING_SERVER_URL,
DOCLING_OCR_ENGINE=request.app.state.config.DOCLING_OCR_ENGINE,
DOCLING_OCR_LANG=request.app.state.config.DOCLING_OCR_LANG,
DOCLING_DO_PICTURE_DESCRIPTION=request.app.state.config.DOCLING_DO_PICTURE_DESCRIPTION,
PDF_EXTRACT_IMAGES=request.app.state.config.PDF_EXTRACT_IMAGES,
DOCUMENT_INTELLIGENCE_ENDPOINT=request.app.state.config.DOCUMENT_INTELLIGENCE_ENDPOINT,
DOCUMENT_INTELLIGENCE_KEY=request.app.state.config.DOCUMENT_INTELLIGENCE_KEY,