fix(markdown): preserve distinct image assets across exports

ms_agent/utils/markdown_converter.py

@@ -1,6 +1,7 @@
 # flake8: noqa
 # isort: skip_file
 # yapf: disable
+import hashlib
 import os
 import re
 import shutil
@@ -127,6 +128,8 @@ def _copy_and_process_images(
         resources_dir = folder_path / 'resources'
 
         for img_path, alt_text in images:
+            if img_path in image_mapping:
+                continue
             # Handle relative paths
             if not os.path.isabs(img_path):
                 # First try resources folder, then folder_path
@@ -137,11 +140,15 @@ def _copy_and_process_images(
                 source_path = Path(img_path)
 
             if source_path.exists():
-                # Generate new filename
-                new_filename = f'{len(image_mapping)}_{source_path.name}'
-                new_path = output_images_dir / new_filename
-
                 try:
+                    # A folder export shares this directory across documents.
+                    # Per-document counters can overwrite another report's images.
+                    digest = hashlib.sha256()
+                    with source_path.open('rb') as source:
+                        for chunk in iter(lambda: source.read(65536), b''):
+                            digest.update(chunk)
+                    new_filename = f'{digest.hexdigest()}{source_path.suffix}'
+                    new_path = output_images_dir / new_filename
                     # Copy and possibly compress image
                     if source_path.suffix.lower() in ['.jpg', '.jpeg', '.png', '.gif', '.bmp']:
                         with Image.open(source_path) as img:

tests/utils/test_markdown_image_assets.py

@@ -0,0 +1,85 @@
+# Copyright (c) ModelScope Contributors. All rights reserved.
+"""Keep exported Markdown reports' local image references intact."""
+from pathlib import Path
+
+from bs4 import BeautifulSoup
+from PIL import Image
+
+from ms_agent.utils.markdown_converter import MarkdownConverter
+
+
+def _image(path, color):
+    path.parent.mkdir(parents=True, exist_ok=True)
+    Image.new('RGB', (4, 4), color).save(path)
+
+
+def _exported_colors(html_path):
+    soup = BeautifulSoup(html_path.read_text(encoding='utf-8'), 'html.parser')
+    paths = [html_path.parent / image['src'] for image in soup.find_all('img')]
+    colors = []
+    for path in paths:
+        with Image.open(path) as image:
+            colors.append(image.convert('RGB').getpixel((0, 0)))
+    return paths, colors
+
+
+def test_repeated_reference_does_not_overwrite_distinct_image(tmp_path):
+    source = tmp_path / 'source'
+    _image(source / 'before' / 'chart.png', 'red')
+    _image(source / 'after' / 'chart.png', 'blue')
+    document = source / 'report.md'
+    document.write_text(
+        '# Report\n\n![Before](before/chart.png)\n\n'
+        '![Before again](before/chart.png)\n\n![After](after/chart.png)\n',
+        encoding='utf-8')
+    output = tmp_path / 'export' / 'report.html'
+    MarkdownConverter.markdown_to_html(str(document), str(output))
+    paths, colors = _exported_colors(output)
+    assert colors == [(255, 0, 0), (255, 0, 0), (0, 0, 255)]
+    assert paths[0] == paths[1]
+    assert paths[0] != paths[2]
+
+
+def test_folder_export_preserves_same_named_images_across_reports(tmp_path):
+    source = tmp_path / 'source'
+    _image(source / 'before' / 'chart.png', 'red')
+    _image(source / 'after' / 'chart.png', 'blue')
+    for name, directory in [('first', 'before'), ('second', 'after')]:
+        (source / f'{name}.md').write_text(
+            f'# {name}\n\n![Chart]({directory}/chart.png)\n', encoding='utf-8')
+    output = tmp_path / 'export'
+    MarkdownConverter.markdown_to_html(str(source), str(output))
+    first_paths, first_colors = _exported_colors(output / 'first.html')
+    second_paths, second_colors = _exported_colors(output / 'second.html')
+    assert first_colors == [(255, 0, 0)]
+    assert second_colors == [(0, 0, 255)]
+    assert first_paths != second_paths
+
+
+def test_reexport_changed_image_does_not_change_previous_report(tmp_path):
+    source = tmp_path / 'source'
+    image = source / 'chart.png'
+    _image(image, 'red')
+    document = source / 'report.md'
+    document.write_text('# Report\n\n![Chart](chart.png)\n', encoding='utf-8')
+    first = tmp_path / 'export' / 'first.html'
+    second = tmp_path / 'export' / 'second.html'
+    MarkdownConverter.markdown_to_html(str(document), str(first))
+    _image(image, 'blue')
+    MarkdownConverter.markdown_to_html(str(document), str(second))
+    assert _exported_colors(first)[1] == [(255, 0, 0)]
+    assert _exported_colors(second)[1] == [(0, 0, 255)]
+
+
+def test_identical_images_share_asset_and_svg_is_preserved(tmp_path):
+    source = tmp_path / 'source'
+    _image(source / 'one.png', 'red')
+    _image(source / 'two.png', 'red')
+    svg = '<svg xmlns="http://www.w3.org/2000/svg"><title>chart</title></svg>'
+    (source / 'chart.svg').write_text(svg, encoding='utf-8')
+    output = tmp_path / 'export'
+    output.mkdir()
+    mapping = MarkdownConverter._copy_and_process_images(
+        [('one.png', ''), ('two.png', ''), ('chart.svg', '')], output, source)
+    assert mapping['one.png'] == mapping['two.png']
+    assert (output / mapping['chart.svg']).read_text(encoding='utf-8') == svg