김경종
68 lines
2.3 KiB
Python
68 lines
2.3 KiB
Python
from __future__ import annotations
|
|
|
|
from pathlib import Path
|
|
|
|
import pytest
|
|
|
|
from pdftomd.models import DocumentIdentity
|
|
from pdftomd.paths import (
|
|
OutputBundlePaths,
|
|
document_identity_from_pdf,
|
|
make_anchor,
|
|
normalize_pdf_path,
|
|
)
|
|
|
|
|
|
def test_normalize_pdf_path_accepts_korean_and_spaced_paths(tmp_path: Path) -> None:
|
|
pdf = tmp_path / "한글 경로" / "My Report 2026.pdf"
|
|
pdf.parent.mkdir()
|
|
pdf.write_bytes(b"%PDF-1.7\n")
|
|
|
|
normalized = normalize_pdf_path(pdf)
|
|
|
|
assert normalized.is_absolute()
|
|
assert normalized.name == "My Report 2026.pdf"
|
|
|
|
|
|
def test_normalize_pdf_path_rejects_non_pdf_files(tmp_path: Path) -> None:
|
|
text_file = tmp_path / "document.txt"
|
|
text_file.write_text("not a pdf", encoding="utf-8")
|
|
|
|
with pytest.raises(ValueError, match="PDF"):
|
|
normalize_pdf_path(text_file)
|
|
|
|
|
|
def test_document_identity_from_pdf_uses_stable_slug(tmp_path: Path) -> None:
|
|
pdf = tmp_path / "한글 보고서.pdf"
|
|
pdf.write_bytes(b"%PDF-1.7\n")
|
|
|
|
first = document_identity_from_pdf(pdf)
|
|
second = document_identity_from_pdf(pdf)
|
|
|
|
assert first.filename == "한글 보고서.pdf"
|
|
assert first.slug == second.slug
|
|
assert first.slug.startswith("document-")
|
|
assert first.source_path == str(normalize_pdf_path(pdf))
|
|
|
|
|
|
def test_output_bundle_paths_keep_document_and_runtime_artifacts_separate(tmp_path: Path) -> None:
|
|
document = DocumentIdentity.from_path("Example Paper.pdf")
|
|
bundle = OutputBundlePaths.from_document(tmp_path, document)
|
|
|
|
assert bundle.document_dir == tmp_path / "example-paper"
|
|
assert bundle.images_dir == tmp_path / "example-paper" / "images"
|
|
assert bundle.chunk_path(1) == tmp_path / "example-paper" / "example-paper_001.md"
|
|
assert bundle.figure_asset_path("1") == tmp_path / "example-paper" / "images" / "example-paper_fig-1.png"
|
|
assert bundle.runtime_dir == tmp_path / ".pdftomd-runtime" / "example-paper"
|
|
assert bundle.log_path.name == "conversion.log"
|
|
assert bundle.resume_state_path.name == "resume-state.json"
|
|
assert bundle.runtime_dir not in bundle.document_dir.parents
|
|
|
|
|
|
def test_make_anchor_is_deterministic_and_validates_kind() -> None:
|
|
assert make_anchor("Figure", "2 A") == "figure-2-a"
|
|
assert make_anchor("Equation", "식 3") == "equation-3"
|
|
|
|
with pytest.raises(ValueError, match="kind"):
|
|
make_anchor("", "1")
|