tests/test_pdf_integration.py

import importlib.util
import tempfile
import unittest
from pathlib import Path

from zsgdp.pipeline import parse_document


@unittest.skipIf(importlib.util.find_spec("fitz") is None, "PyMuPDF is not installed")
class PDFIntegrationTests(unittest.TestCase):
    def test_pymupdf_parse_exports_page_table_and_figure_assets(self):
        import fitz  # type: ignore

        with tempfile.TemporaryDirectory() as tmp:
            tmp_path = Path(tmp)
            pdf_path = tmp_path / "sample.pdf"
            output_dir = tmp_path / "out"
            doc = fitz.open()
            page = doc.new_page(width=612, height=792)
            page.insert_text((72, 72), "Annual Report", fontsize=20)
            page.insert_text((72, 120), "Revenue Summary", fontsize=14)
            page.insert_text(
                (72, 155),
                "Region    Q1    Q2\nNorth America    10    12\nEurope    8    7",
                fontsize=11,
                fontname="cour",
            )
            page.draw_rect(fitz.Rect(72, 265, 260, 360))
            doc.save(pdf_path)

            parsed = parse_document(pdf_path, output_dir, selected_parsers=["pymupdf"])

            self.assertEqual(parsed.file_type, "pdf")
            self.assertEqual(len(parsed.tables), 1)
            self.assertGreaterEqual(len(parsed.figures), 1)
            self.assertTrue((output_dir / "assets" / "pages" / "page_0001.png").exists())
            self.assertTrue((output_dir / "assets" / "tables" / "p0001_t001.png").exists())
            self.assertTrue(any((output_dir / "assets" / "figures").glob("p0001_f*.png")))
            self.assertEqual(parsed.quality_report.metrics["table_chunk_coverage"], 1.0)
            self.assertEqual(parsed.quality_report.metrics["figure_chunk_coverage"], 1.0)


if __name__ == "__main__":
    unittest.main()