baram2584/PDFToMD
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

modify pdftomd

Browse Source
This commit is contained in:
김경종
2026-05-14 10:16:59 +09:00
parent 2232b51fc9
commit dc11880140
69 changed files with 7784 additions and 1150 deletions
Download Patch File Download Diff File Expand all files Collapse all files
tests/test_paths.py
+18 -16
View File
@@ -73,14 +73,14 @@ def test_discovers_directory_recursive_with_relative_parents(tmp_path: Path) ->
def test_discovery_order_is_deterministic_for_non_ascii_names(tmp_path: Path) -> None:
touch(tmp_path / "한글.pdf")
korean_pdf = touch(tmp_path / "논문.pdf")
touch(tmp_path / "Alpha.pdf")
touch(tmp_path / "beta.PDF")
first = discover_pdfs(tmp_path)
second = discover_pdfs(tmp_path)
assert [item.source_path.name for item in first] == ["Alpha.pdf", "beta.PDF", "한글.pdf"]
assert [item.source_path.name for item in first] == ["Alpha.pdf", "beta.PDF", korean_pdf.name]
assert first == second
@@ -91,22 +91,24 @@ def test_plans_all_default_output_paths_for_single_pdf(tmp_path: Path) -> None:
[plan] = plan_pdf_outputs(pdf, output_root)
assert plan.source_pdf == pdf.resolve()
assert plan.markdown_path == output_root.resolve() / "입력.md"
assert plan.assets_dir == output_root.resolve() / "입력.assets"
assert plan.metadata_path == output_root.resolve() / "입력.metadata.json"
assert plan.report_path == output_root.resolve() / "입력.report.md"
assert plan.markdown_path == output_root.resolve() / "입력" / "입력_001.md"
assert plan.assets_dir == output_root.resolve() / "입력" / "images"
assert plan.metadata_path is None
assert plan.report_path == output_root.resolve() / "입력" / "입력_report.md"
assert plan.raw_dir is None
def test_plans_optional_metadata_and_raw_outputs(tmp_path: Path) -> None:
def test_plans_metadata_flag_as_noop_and_raw_outputs(tmp_path: Path) -> None:
pdf = touch(tmp_path / "paper.pdf")
[with_metadata_flag] = plan_pdf_outputs(pdf, tmp_path / "out", metadata=True)
[without_metadata] = plan_pdf_outputs(pdf, tmp_path / "out", metadata=False)
[with_raw] = plan_pdf_outputs(pdf, tmp_path / "out", keep_raw=True)
assert with_metadata_flag.metadata_path is None
assert without_metadata.metadata_path is None
assert without_metadata.report_path == (tmp_path / "out").resolve() / "paper.report.md"
assert with_raw.raw_dir == (tmp_path / "out").resolve() / "paper.raw"
assert without_metadata.report_path == (tmp_path / "out").resolve() / "paper" / "paper_report.md"
assert with_raw.raw_dir == (tmp_path / "out").resolve() / "paper" / "raw"
def test_recursive_planning_preserves_relative_subdirectories(tmp_path: Path) -> None:
@@ -117,8 +119,8 @@ def test_recursive_planning_preserves_relative_subdirectories(tmp_path: Path) ->
plans = plan_pdf_outputs(root, tmp_path / "out", recursive=True)
assert [plan.markdown_path.relative_to((tmp_path / "out").resolve()) for plan in plans] == [
Path("nested") / "same.md",
Path("same.md"),
Path("nested") / "same" / "same_001.md",
Path("same") / "same_001.md",
]
@@ -137,21 +139,21 @@ def test_non_recursive_duplicate_output_paths_fail(tmp_path: Path) -> None:
def test_output_conflicts_report_all_existing_paths(tmp_path: Path) -> None:
pdf = touch(tmp_path / "paper.pdf")
output_root = tmp_path / "out"
(output_root / "paper.assets").mkdir(parents=True)
(output_root / "paper.md").mkdir()
touch(output_root / "paper.metadata.json")
(output_root / "paper" / "images").mkdir(parents=True)
(output_root / "paper" / "paper_001.md").mkdir()
touch(output_root / "paper" / "paper_report.md")
with pytest.raises(OutputConflictError) as error:
plan_pdf_outputs(pdf, output_root)
conflict_names = {path.name for path in error.value.conflicts}
assert conflict_names == {"paper.assets", "paper.md", "paper.metadata.json"}
assert conflict_names == {"images", "paper_001.md", "paper_report.md"}
def test_overwrite_allows_existing_paths_without_deleting(tmp_path: Path) -> None:
pdf = touch(tmp_path / "paper.pdf")
output_root = tmp_path / "out"
existing = touch(output_root / "paper.md")
existing = touch(output_root / "paper" / "paper_001.md")
[plan] = plan_pdf_outputs(pdf, output_root, overwrite=True)