zlei6/ragflow
1
0
Fork 0
mirror of https://github.com/infiniflow/ragflow.git synced 2026-06-29 23:41:12 +08:00
Code Issues Packages Projects Releases Wiki Activity

Fix: MinerU 3.x output discovery and API contract (#14282)

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

update MinerU parser to most recent minerU v3 logic

### Type of change

- [x] Bug Fix (non-breaking change which fixes an issue)
This commit is contained in:
Idriss Sbaaoui
2026-04-22 14:44:41 +08:00
committed by GitHub
parent 3ce1e44b2d
commit d5d162b374
1 changed files with 41 additions and 24 deletions
Download Patch File Download Diff File Expand all files Collapse all files

65
deepdoc/parser/mineru_parser.py
View File

@@ -288,13 +288,13 @@ class MinerUParser(RAGFlowPdfParser):
headers = {"Accept": "application/json"}
try:
self.logger.info(f"[MinerU] invoke api: {self.mineru_api}/pdf_parse backend={options.backend} server_url={data.get('server_url')}")
self.logger.info(f"[MinerU] invoke api: {self.mineru_api}/file_parse backend={options.backend} server_url={data.get('server_url')}")
if callback:
callback(0.20, f"[MinerU] invoke api: {self.mineru_api}/pdf_parse")
callback(0.20, f"[MinerU] invoke api: {self.mineru_api}/file_parse")
with open(pdf_file_path, "rb") as pdf_file:
files = {"files": (pdf_file_name + ".pdf", pdf_file, "application/pdf")}
with requests.post(
url=f"{self.mineru_api}/pdf_parse",
url=f"{self.mineru_api}/file_parse",
files=files,
data=data,
headers=headers,
@@ -303,27 +303,22 @@ class MinerUParser(RAGFlowPdfParser):
) as response:
response.raise_for_status()
content_type = response.headers.get("Content-Type", "")
if content_type.startswith("application/zip"):
self.logger.info(f"[MinerU] zip file returned, saving to {output_zip_path}...")
if callback:
callback(0.30, f"[MinerU] zip file returned, saving to {output_zip_path}...")
with open(output_zip_path, "wb") as f:
response.raw.decode_content = True
shutil.copyfileobj(response.raw, f)
self.logger.info(f"[MinerU] Unzip to {output_path}...")
self._extract_zip_no_root(output_zip_path, output_path, pdf_file_name + "/")
if callback:
callback(0.40, f"[MinerU] Unzip to {output_path}...")
else:
self.logger.warning(f"[MinerU] not zip returned from api: {content_type}")
except Exception as e:
if not content_type.startswith("application/zip"):
raise RuntimeError(f"[MinerU] not zip returned from api: {content_type}")
self.logger.info(f"[MinerU] zip file returned, saving to {output_zip_path}...")
if callback:
callback(0.30, f"[MinerU] zip file returned, saving to {output_zip_path}...")
with open(output_zip_path, "wb") as f:
response.raw.decode_content = True
shutil.copyfileobj(response.raw, f)
self.logger.info(f"[MinerU] Unzip to {output_path}...")
self._extract_zip_no_root(output_zip_path, output_path, pdf_file_name + "/")
if callback:
callback(0.40, f"[MinerU] Unzip to {output_path}...")
self.logger.info("[MinerU] Api completed successfully.")
return Path(output_path)
except requests.RequestException as e:
raise RuntimeError(f"[MinerU] api failed with exception {e}")
self.logger.info("[MinerU] Api completed successfully.")
return Path(output_path)
def __images__(self, fnm, zoomin: int = 1, page_from=0, page_to=600, callback=None):
self.page_from = page_from
@@ -517,7 +512,8 @@ class MinerUParser(RAGFlowPdfParser):
return sanitized or "unnamed"
safe_stem = _sanitize_filename(file_stem)
allowed_names = {f"{file_stem}_content_list.json", f"{safe_stem}_content_list.json"}
content_names = (f"{file_stem}_content_list.json", f"{safe_stem}_content_list.json")
allowed_names = set(content_names)
self.logger.info(f"[MinerU] Expected output files: {', '.join(sorted(allowed_names))}")
self.logger.info(f"[MinerU] Searching output in: {output_dir}")
@@ -542,6 +538,27 @@ class MinerUParser(RAGFlowPdfParser):
subdir = nested_alt.parent
json_file = nested_alt
if not json_file:
parse_subdir = None
if backend.startswith("pipeline"):
parse_subdir = method
elif backend.startswith("hybrid"):
parse_subdir = f"hybrid_{method}"
elif backend.startswith("vlm"):
parse_subdir = "vlm"
if parse_subdir:
for content_name in content_names:
for candidate in output_dir.glob(f"**/{parse_subdir}/{content_name}"):
self.logger.info(f"[MinerU] Trying parse-method path: {candidate}")
attempted.append(candidate)
if candidate.exists():
subdir = candidate.parent
json_file = candidate
break
if json_file:
break
if not json_file:
raise FileNotFoundError(f"[MinerU] Missing output file, tried: {', '.join(str(p) for p in attempted)}")