add pdftomd

2026-05-08 16:42:19 +09:00
parent 551ab50735
commit 88d6b92283
99 changed files with 47332 additions and 0 deletions
@@ -0,0 +1,118 @@
+from __future__ import annotations
+
+import json
+import os
+import re
+import subprocess
+import sys
+from pathlib import Path
+
+import pytest
+
+
+if os.environ.get("PDF2MD_RUN_MINERU_FIXTURES") != "1":
+    pytest.skip(
+        "optional local MinerU fixture evaluation is disabled; set PDF2MD_RUN_MINERU_FIXTURES=1 to run",
+        allow_module_level=True,
+    )
+
+
+REPO_ROOT = Path(__file__).resolve().parents[2]
+SAMPLES_DIR = REPO_ROOT / "samples"
+
+
+def test_optional_local_mineru_samples_produce_release_outputs(tmp_path: Path) -> None:
+    doctor = subprocess.run(
+        [sys.executable, "-m", "pdf2md.cli", "doctor"],
+        cwd=REPO_ROOT,
+        check=False,
+        capture_output=True,
+        text=True,
+    )
+    if doctor.returncode != 0:
+        pytest.skip(f"local MinerU fixture evaluation blocked by doctor:\n{doctor.stdout}\n{doctor.stderr}")
+
+    sample_pdfs = tuple(sorted(SAMPLES_DIR.glob("*.pdf"), key=lambda path: path.name.casefold()))
+    if not sample_pdfs:
+        pytest.skip(f"no local sample PDFs found under {SAMPLES_DIR}")
+
+    output_root = tmp_path / "mineru-fixture-output"
+    attempts: list[dict[str, object]] = []
+    for pdf in sample_pdfs:
+        sample_output = output_root / pdf.stem
+        completed = subprocess.run(
+            [
+                sys.executable,
+                "-m",
+                "pdf2md.cli",
+                "convert",
+                str(pdf),
+                "--out",
+                str(sample_output),
+            ],
+            cwd=REPO_ROOT,
+            check=False,
+            capture_output=True,
+            text=True,
+            timeout=1800,
+        )
+        attempts.append(
+            {
+                "source": str(pdf.relative_to(REPO_ROOT)),
+                "command": " ".join(
+                    [
+                        sys.executable,
+                        "-m",
+                        "pdf2md.cli",
+                        "convert",
+                        str(pdf),
+                        "--out",
+                        str(sample_output),
+                    ]
+                ),
+                "exit_code": completed.returncode,
+                "stdout": completed.stdout,
+                "stderr": completed.stderr,
+            }
+        )
+        assert completed.returncode == 0, json.dumps(attempts[-1], ensure_ascii=False, indent=2)
+
+        markdown_path = sample_output / f"{pdf.stem}.md"
+        metadata_path = sample_output / f"{pdf.stem}.metadata.json"
+        report_path = sample_output / f"{pdf.stem}.report.md"
+        assert markdown_path.exists()
+        assert metadata_path.exists()
+        assert report_path.exists()
+
+        metadata = json.loads(metadata_path.read_text(encoding="utf-8"))
+        summary = metadata["summary"]
+        assert metadata["engine"] == "MinerU"
+        assert summary["pages_processed"] >= 1
+        assert "warning_count" in summary
+        assert "math_render_error_count" in summary
+        assert "asset_count" in summary
+        report = report_path.read_text(encoding="utf-8")
+        assert "Output Markdown:" in report
+        assert "Metadata JSON:" in report
+        assert "Report Markdown:" in report
+        attempts[-1].update(
+            {
+                "markdown_path": str(markdown_path),
+                "metadata_path": str(metadata_path),
+                "report_path": str(report_path),
+                "warning_count": summary["warning_count"],
+                "final_status": _report_final_status(report),
+                "math_render_error_count": summary["math_render_error_count"],
+                "asset_count": summary["asset_count"],
+                "pages_processed": summary["pages_processed"],
+            }
+        )
+
+    record_path = output_root / "fixture-evaluation.json"
+    record_path.write_text(json.dumps({"attempts": attempts}, ensure_ascii=False, indent=2) + "\n", encoding="utf-8")
+    assert record_path.exists()
+
+
+def _report_final_status(report: str) -> str:
+    match = re.search(r"^- Final status: `(?P<status>[^`]+)`$", report, re.MULTILINE)
+    return match.group("status") if match else "unavailable"