MCPcopy Create free account
hub / github.com/Paper2Poster/Paper2Poster / extract_images_and_sections

Function extract_images_and_sections

utils/poster_eval_utils.py:1020–1043  ·  view source on GitHub ↗
(md)

Source from the content-addressed store, hash-verified

1018 img.save(output_path, "PDF", resolution=200.0)
1019
1020def extract_images_and_sections(md):
1021 parts = re.split(r'(## [^\n]+)', md)
1022 records = []
1023 for i in range(1, len(parts), 2):
1024 header = parts[i].strip()
1025 content = parts[i+1]
1026 # Find all image paths
1027 images = re.findall(r'!\[.*?\]\((.*?)\)', content)
1028 if images:
1029 # Remove lines that are image markdown
1030 lines = content.splitlines()
1031 cleaned = [
1032 line for line in lines
1033 if not re.match(r'!\[.*?\]\(.*?\)', line.strip())
1034 ]
1035 section_text = "\n".join(cleaned).strip()
1036 for img in images:
1037 records.append({
1038 'section': header,
1039 'image_path': unquote(img),
1040 'section_text': section_text
1041 })
1042
1043 return records
1044
1045def gen_eval_markdown(paper_name, poster_method, poster_path, figure_count_only=False):
1046 model_name="openai/clip-vit-base-patch32"

Callers 1

gen_eval_markdownFunction · 0.70

Calls

no outgoing calls

Tested by

no test coverage detected