feat: 新增 .txt 纯文本文件支持
GraphRAG CI/CD / build-and-deploy (push) Successful in 5m4s

- 添加 txt 到允许上传格式,跳过 MinerU OCR
- 直接读取 UTF-8 文本,按 3000 字符/页分页后送入 LangExtract
- 适合有中文文本的规范文件,避免扫描件 OCR 乱码问题
This commit is contained in:
2026-06-18 17:00:40 +08:00
parent a46ceb94de
commit 71676dba9e
3 changed files with 84 additions and 49 deletions
+1
View File
@@ -110,6 +110,7 @@ async def list_formats():
{"ext": "jpg", "description": "JPEG 图片(单页)", "max_size_mb": 200, "max_pages": 1, "requires_ocr": True},
{"ext": "jpeg", "description": "JPEG 图片(单页)", "max_size_mb": 200, "max_pages": 1, "requires_ocr": True},
{"ext": "html", "description": "HTML 文件", "max_size_mb": 200, "max_pages": 600, "requires_ocr": False},
{"ext": "txt", "description": "纯文本文件(UTF-8 编码)", "max_size_mb": 200, "max_pages": 0, "requires_ocr": False},
],
"ocr_languages": [
{"code": "ch", "name": "中文(默认)"},