kris/ForcePilot
1
0
Fork 0
Code Issues Pull Requests Actions 1 Packages Projects Releases Wiki Activity

fix(parser): 修复 Docling 解析文档时图片错位问题

Browse Source 
遍历所有图片生成等长替换列表，非 data URI 图片填空字符串，
确保占位符替换严格 1:1，避免混合图片类型时顺序错乱。
This commit is contained in:
Aw 2026-05-11 21:54:25 +08:00
parent f07b32b365
commit 56166abff6
1 changed files with 14 additions and 21 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

35
backend/package/yuxi/plugins/parser/unified.py
View File

@ -127,31 +127,24 @@ def _convert_with_docling(file_path: Path, params: dict | None = None) -> str:
doc = result.document doc = result.document
if hasattr(doc, "pictures") and doc.pictures: if hasattr(doc, "pictures") and doc.pictures:
image_refs: list[tuple[str, bytes]] = [] replacements: list[str] = []
for pic in doc.pictures: for pic in doc.pictures:
if hasattr(pic, "image") and hasattr(pic.image, "uri"): uri = str(pic.image.uri) if hasattr(pic, "image") and hasattr(pic.image, "uri") else ""
uri = str(pic.image.uri) if uri.startswith("data:"):
if uri.startswith("data:"): try:
image_data, mime_type = _parse_data_uri(uri) image_data, mime_type = _parse_data_uri(uri)
timestamp = int(time.time() * 1000000) filename = f"image_{int(time.time() * 1000000)}.{mime_type.split('/')[-1]}"
filename = f"image_{timestamp}.{mime_type.split('/')[-1]}" url = _upload_image_to_minio(image_data, filename, image_bucket, image_prefix)
image_refs.append((filename, image_data)) replacements.append(f"![{filename}]({url})")
except Exception as e: # noqa: BLE001
image_urls: list[str] = [] logger.error(f"上传图片失败 {filename}: {e}")
for filename, image_data in image_refs: replacements.append("")
try: else:
url = _upload_image_to_minio(image_data, filename, image_bucket, image_prefix) replacements.append("")
image_urls.append(f"![{filename}]({url})")
except Exception as e: # noqa: BLE001
logger.error(f"上传图片失败 {filename}: {e}")
image_urls.append(f"[图片: {filename}]")
markdown = doc.export_to_markdown() markdown = doc.export_to_markdown()
for replacement in replacements:
for url in image_urls: markdown = re.sub(r"<!--\s*image\s*-->", replacement, markdown, count=1)
markdown = re.sub(r"<!--\s*image\s*-->", url, markdown, count=1)
return markdown return markdown
return doc.export_to_markdown() return doc.export_to_markdown()