initial commit

tests/compare_link_borders.py

@@ -0,0 +1,96 @@
+#!/usr/bin/env python3
+"""Read-only link annotation evidence through the production sandboxed worker.
+
+Poppler is an independent comparison, not an all-features correctness oracle.
+Run after building pdf-worker-evidence. Requires Pillow and Poppler utils.
+"""
+import argparse
+import hashlib
+import json
+from pathlib import Path
+import subprocess
+import time
+from PIL import Image, ImageChops, ImageStat
+
+ROOT = Path(__file__).resolve().parent.parent
+
+def sha(path):
+    with path.open('rb') as source:
+        return hashlib.file_digest(source, 'sha256').hexdigest()
+
+def stats(a, b):
+    assert a.size == b.size
+    diff = ImageChops.difference(a, b)
+    pixels = list(diff.getdata())
+    return {'size': list(a.size), 'pixels': len(pixels), 'equal_pixels': sum(max(p) == 0 for p in pixels),
+            'pixels_max_delta_gt_16': sum(max(p) > 16 for p in pixels),
+            'mean_absolute_channel_error': sum(ImageStat.Stat(diff).mean) / 3}
+
+def region(rect, rotation, margin=6):
+    points = []
+    for x in [rect[0], rect[2]]:
+        for y in [rect[1], rect[3]]:
+            x0, y0 = (x - 20) * 2, (450 - y) * 2
+            points.append([(x0, y0), (840 - y0, x0), (1020 - x0, 840 - y0), (y0, 1020 - x0)][rotation // 90])
+    return (min(p[0] for p in points) - margin, min(p[1] for p in points) - margin,
+            max(p[0] for p in points) + margin, max(p[1] for p in points) + margin)
+
+def main():
+    parser = argparse.ArgumentParser(description=__doc__)
+    parser.add_argument('--build-dir', type=Path, default=ROOT / 'build')
+    parser.add_argument('--output', type=Path, default=ROOT / 'tests/results/link-borders')
+    args = parser.parse_args()
+    out = args.output.resolve(); out.mkdir(parents=True, exist_ok=True)
+    build = args.build_dir.resolve(); helper = build / 'pdf-worker-evidence'
+    fixture = ROOT / 'tests/fixtures/pdf/link-borders/styles.pdf'
+    before = sha(fixture)
+    start = time.monotonic()
+    subprocess.run([str(helper), str(fixture), str(out / 'worker')], check=True, timeout=60)
+    elapsed = time.monotonic() - start
+    reference = subprocess.run(['/usr/bin/pdftoppm', '-cropbox', '-r', '144', '-png', str(fixture), str(out / 'poppler')], capture_output=True, check=True, timeout=60)
+    (out / 'poppler-stderr.txt').write_bytes(reference.stderr)
+    manifest = json.loads((fixture.parent / 'manifest.json').read_text())['fixtures'][0]
+    common = {'default', 'solid-rgb', 'dashed-border', 'dashed-bs-odd', 'dashed-default', 'underline',
+              'bs-overrides-border', 'zero-width', 'transparent', 'hidden', 'no-view', 'print-flag-visible'}
+    records = []
+    for index, rotation in enumerate([0, 90, 180, 270], 1):
+        a = Image.open(out / f'worker/page-{index}.png').convert('RGB')
+        b = Image.open(out / f'poppler-{index}.png').convert('RGB')
+        diff = ImageChops.difference(a, b); diff.save(out / f'difference-{index}.png')
+        comparisons = []
+        for case in manifest['annotations']:
+            bounds = region(case['rect'], rotation)
+            row = {'name': case['name'], 'reference_supported': case['name'] in common, **stats(a.crop(bounds), b.crop(bounds))}
+            if case['name'] in common:
+                # At 144dpi the grid-aligned borders should match except small
+                # independent rasterizer antialias/color-rounding differences.
+                assert row['mean_absolute_channel_error'] < 3, row
+            comparisons.append(row)
+        records.append({'page': index, 'rotation': rotation, 'whole_page': stats(a, b), 'annotations': comparisons})
+    navigation = ROOT / 'tests/fixtures/pdf/navigation.pdf'
+    navigation_before = sha(navigation)
+    subprocess.run([str(helper), str(navigation), str(out / 'navigation-worker')], check=True, timeout=60)
+    subprocess.run(['/usr/bin/pdftoppm', '-cropbox', '-f', '1', '-singlefile', '-r', '144', '-png', str(navigation), str(out / 'navigation-poppler')], capture_output=True, check=True, timeout=60)
+    a = Image.open(out / 'navigation-worker/page-1.png').convert('RGB')
+    b = Image.open(out / 'navigation-poppler.png').convert('RGB')
+    # The exact region where the original default link border was absent.
+    navigation_comparison = stats(a.crop((18, 368, 282, 400)), b.crop((18, 368, 282, 400)))
+    assert navigation_comparison['mean_absolute_channel_error'] < 3, navigation_comparison
+    assert before == sha(fixture) and navigation_before == sha(navigation)
+    result = {'transport': 'production WorkerProcess, sandboxed PDFWorker and FontBroker; no direct PDFium exporter',
+              'worker_sha256': sha(build / 'docview-pdf-worker'), 'exporter_sha256': sha(helper),
+              'qpdf_version': subprocess.check_output(['qpdf', '--version'], text=True).splitlines()[0],
+              'poppler_version': subprocess.run(['/usr/bin/pdftoppm', '-v'], capture_output=True, text=True).stderr.splitlines()[0],
+              'fixture_sha256': before, 'source_unchanged': True, 'four_page_export_seconds': elapsed,
+              'reference_status': 'independent Poppler comparison; manually inspect images; not a golden-image acceptance gate',
+              'reference_limitations': manifest['reference_limitations'],
+              'known_limitations': ['NoZoom/NoRotate geometry for generated links, saved normal APs, and native Text icons is covered separately by tests/results/pdf-annotation-flags.',
+                  'Other annotation types without a supplied AP retain PDFium generation behavior; this corpus does not certify their NoZoom handling.',
+                  'Text annotation icon appearance is viewer-specific and is not a link border defect.',
+                  'This closes tested standard link border cases, not the entire G-RENDER gate.'],
+              'pages': records, 'navigation_fixture_sha256': navigation_before, 'navigation_default_link_region': navigation_comparison}
+    (out / 'comparison.json').write_text(json.dumps(result, indent=2) + '\n')
+    print(json.dumps({'source_unchanged': True, 'four_page_export_seconds': elapsed,
+                      'standard_case_comparisons': len(common) * 4, 'navigation_default_link_region': navigation_comparison}, indent=2))
+
+if __name__ == '__main__': main()