pengjunjie/ragflow
1
0
Fork 1
mirror of https://github.com/infiniflow/ragflow.git synced 2025-12-08 12:32:30 +08:00
Code Issues Packages Projects Releases Wiki Activity

Move api.settings to common.settings (#11036)

Browse Source 
### What problem does this PR solve?

As title

### Type of change

- [x] Refactoring

---------

Signed-off-by: Jin Hai <haijin.chn@gmail.com>
This commit is contained in:
Jin Hai
2025-11-06 09:36:38 +08:00
committed by GitHub
parent 87c9a054d3
commit f98b24c9bf
68 changed files with 675 additions and 718 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
deepdoc/parser/pdf_parser.py
View File

@ -40,7 +40,7 @@ from deepdoc.vision import OCR, AscendLayoutRecognizer, LayoutRecognizer, Recogn
from rag.app.picture import vision_llm_chunk as picture_vision_llm_chunk
from rag.nlp import rag_tokenizer
from rag.prompts.generator import vision_llm_describe_prompt
from rag.settings import PARALLEL_DEVICES
from common import settings
LOCK_KEY_pdfplumber = "global_shared_lock_pdfplumber"
if LOCK_KEY_pdfplumber not in sys.modules:
@ -63,8 +63,8 @@ class RAGFlowPdfParser:
self.ocr = OCR()
self.parallel_limiter = None
if PARALLEL_DEVICES > 1:
self.parallel_limiter = [trio.CapacityLimiter(1) for _ in range(PARALLEL_DEVICES)]
if settings.PARALLEL_DEVICES > 1:
self.parallel_limiter = [trio.CapacityLimiter(1) for _ in range(settings.PARALLEL_DEVICES)]
layout_recognizer_type = os.getenv("LAYOUT_RECOGNIZER_TYPE", "onnx").lower()
if layout_recognizer_type not in ["onnx", "ascend"]:
@ -1113,7 +1113,7 @@ class RAGFlowPdfParser:
for i, img in enumerate(self.page_images):
chars = __ocr_preprocess()
nursery.start_soon(__img_ocr, i, i % PARALLEL_DEVICES, img, chars, self.parallel_limiter[i % PARALLEL_DEVICES])
nursery.start_soon(__img_ocr, i, i % settings.PARALLEL_DEVICES, img, chars, self.parallel_limiter[i % settings.PARALLEL_DEVICES])
await trio.sleep(0.1)
else:
for i, img in enumerate(self.page_images):

10
deepdoc/vision/ocr.py
View File

@ -23,7 +23,7 @@ from huggingface_hub import snapshot_download
from common.file_utils import get_project_base_directory
from common.misc_utils import pip_install_torch
from rag.settings import PARALLEL_DEVICES
from common import settings
from .operators import * # noqa: F403
from . import operators
import math
@ -554,10 +554,10 @@ class OCR:
"rag/res/deepdoc")
# Append muti-gpus task to the list
if PARALLEL_DEVICES > 0:
if settings.PARALLEL_DEVICES > 0:
self.text_detector = []
self.text_recognizer = []
for device_id in range(PARALLEL_DEVICES):
for device_id in range(settings.PARALLEL_DEVICES):
self.text_detector.append(TextDetector(model_dir, device_id))
self.text_recognizer.append(TextRecognizer(model_dir, device_id))
else:
@ -569,10 +569,10 @@ class OCR:
local_dir=os.path.join(get_project_base_directory(), "rag/res/deepdoc"),
local_dir_use_symlinks=False)
if PARALLEL_DEVICES > 0:
if settings.PARALLEL_DEVICES > 0:
self.text_detector = []
self.text_recognizer = []
for device_id in range(PARALLEL_DEVICES):
for device_id in range(settings.PARALLEL_DEVICES):
self.text_detector.append(TextDetector(model_dir, device_id))
self.text_recognizer.append(TextRecognizer(model_dir, device_id))
else: