Transforms complex documents like PDFs and Office docs into LLM-ready markdown/JSON for your Agentic workflows.
Python ai4sciencedocument-analysisdocxextract-datalayout-analysisocrparserpdfpdf-converterpdf-extractor-llmpdf-extractor-pretrainpdf-extractor-ragpdf-parserpptxpythonxlsx
from __future__ import annotations
from doclib_test_utils import create_parse_services
import asyncio
from io import BytesIO
from pathlib import Path
import pytest
from docvortex.analyzers.native.office.rtf.converter import extract_rtf_metadata
from docvortex.export.middle import export_middle_json
from mineru.backend.analyze import aio_doc_analyze, doc_analyze
from mineru.doclib.core import file_io
from mineru.doclib.core.db import DatabaseManager
from mineru.doclib.core.fts import FTSManager
from mineru.errors import InvalidRequestError
from mineru.parser import parse
from mineru.render import RenderMode, render_docx, render_html, render_markdown, render_structured_content
_PROJECT_ROOT = Path(__file__).resolve().parents[2]
_SEMANTIC_RTF = _PROJECT_ROOT / "tests" / "fixtures" / "rtf" / "semantic.rtf"
def _complex_rtf() -> bytes:
"""读取覆盖标题、列表、表格、公式、注释、链接和图片的确定性 RTF。"""
return _SEMANTIC_RTF.read_bytes()
def test_rtf_doc_analyze_and_renderers_share_strict_metadata() -> None:
"""验证 RTF 同步/异步入口和四类 renderer 使用同一严格 Middle JSON。"""
middle, model = doc_analyze(_complex_rtf(), effort="xhigh", parse_mode="ocr", file_suffix="rtf")
async_middle, async_model = asyncio.run(
aio_doc_analyze(_complex_rtf(), effort="medium", parse_mode="auto", file_suffix="rtf")
)
assert middle.metadata.file_suffix == model.metadata.file_suffix == "rtf"
assert middle.extensions["mineru"]["tier"] == model.extensions["mineru"]["tier"] == "flash"
assert middle.extensions["mineru"]["parse_mode"] == model.extensions["mineru"]["parse_mode"] == "txt"
assert middle.is_full_document is model.is_full_document is True
assert len(middle.pages) == len(model.pages) == 1
assert async_middle == middle
assert async_model == model
markdown = render_markdown(middle, mode=RenderMode.FULL)
html = render_html(middle, mode=RenderMode.FULL, standalone=False)
docx = render_docx(middle)
structured = render_structured_content(middle)
assert "Inherited heading" in markdown
assert "Foot body" in markdown
assert "<table" in html
assert docx.startswith(b"PK\x03\x04")
assert structured["metadata"]["file_suffix"] == "rtf"
def test_public_parser_detects_rtf_content_before_extension(tmp_path: Path) -> None:
"""验证路径 parser 用强签名路由,并把图片导出到 ParseResult cache。"""
source = tmp_path / "disguised.csv"
source.write_bytes(b"\xef\xbb\xbf \r\n" + _complex_rtf())
result = parse(source, tier="flash")
with pytest.raises(InvalidRequestError, match="only supported for PDF") as exc_info:
parse(source, tier="flash", page_range="99")
assert exc_info.value.code == "page_range_invalid"
assert result.middle_json.metadata.file_suffix == "rtf"
assert result.middle_json.is_full_document is True
assert len(result.pages) == 1
exported = export_middle_json(result.middle_json, tmp_path / "export")
assert exported.image_paths
assert all(path.exists() for path in exported.image_paths)
assert "image_base64" not in exported.middle_json.to_json()
def test_rtf_metadata_is_reused_by_doclib(tmp_path: Path) -> None:
"""验证 info destination 元数据和固定单页计数进入 doclib。"""
source = tmp_path / "sample.rtf"
source.write_bytes(_complex_rtf())
direct = extract_rtf_metadata(BytesIO(_complex_rtf()))
metadata = asyncio.run(file_io.extract_metadata(str(source)))
assert direct == {
"title": "Café",
"author": "MinerU",
"subject": "RTF",
"keywords": "alpha;beta",
}
assert metadata["page_count"] == 1
assert metadata["title"] == "Café"
def test_doclib_ingests_rtf_as_one_page_flash_document(tmp_path: Path) -> None:
"""验证 doclib 为 RTF 建立 flash parse row、独立文件类型和固定页数。"""
class _NoRulesConfig:
async def match_rules(self, path: str, rule_type: str) -> list[dict[str, object]]:
"""让测试只观察默认 RTF 入库与 tier 归一行为。"""
return []
async def run() -> None:
"""执行真实 SQLite doclib 入库并检查稳定行数据。"""
db = DatabaseManager(str(tmp_path / "doclib.db"))
await db.initialize()
service = create_parse_services(
db=db,
fts=FTSManager(db),
config_svc=_NoRulesConfig(), # type: ignore[arg-type]
data_dir=str(tmp_path / "data"),
parse_lock_timeout_sec=1800,
)
source = tmp_path / "sample.rtf"
source.write_bytes(_complex_rtf())
response = await service.queue.request_parse(str(source), tier="flash")
doc = await db.fetchone(
"SELECT file_type, page_count, title FROM docs WHERE sha256=?",
(response.sha256,),
)
parses = await db.fetchall(
"SELECT tier, page_range, status, privacy FROM parses WHERE sha256=?",
(response.sha256,),
)
assert response.status == "pending"
assert response.tier == "flash"
assert doc == {"file_type": "rtf", "page_count": 1, "title": "Café"}
assert parses == [{"tier": "flash", "page_range": "1", "status": "pending", "privacy": "local"}]
asyncio.run(run())
@pytest.mark.parametrize(
("request_kwargs", "expected_code", "expected_param"),
[
({"tier": "standard"}, "tier_unsupported_for_file_type", "tier"),
({"tier": "flash", "remote": True}, "remote_unsupported_for_file_type", "remote"),
],
)
def test_doclib_rejects_rtf_quality_tier_and_remote(
tmp_path: Path,
request_kwargs: dict[str, object],
expected_code: str,
expected_param: str,
) -> None:
"""验证 RTF 继承非 PDF/image 的严格 tier 与 remote 边界。"""
class _NoRulesConfig:
async def match_rules(self, path: str, rule_type: str) -> list[dict[str, object]]:
"""让测试只观察主动单文件请求的参数校验。"""
return []
async def run() -> None:
"""创建隔离 doclib 并断言稳定错误契约。"""
db = DatabaseManager(str(tmp_path / "doclib.db"))
await db.initialize()
service = create_parse_services(
db=db,
fts=FTSManager(db),
config_svc=_NoRulesConfig(), # type: ignore[arg-type]
data_dir=str(tmp_path / "data"),
parse_lock_timeout_sec=1800,
)
source = tmp_path / "sample.rtf"
source.write_bytes(_complex_rtf())
with pytest.raises(InvalidRequestError) as exc_info:
await service.queue.request_parse(str(source), **request_kwargs) # type: ignore[arg-type]
assert exc_info.value.code == expected_code
assert exc_info.value.param == expected_param
asyncio.run(run())