| import re | |
| from .schemas import Block, SafeError | |
| def polygon(points, width, height): | |
| p = [ | |
| [ | |
| round(min(1000, max(0, float(x) / width * 1000)), 3), | |
| round(min(1000, max(0, float(y) / height * 1000)), 3), | |
| ] | |
| for x, y in points | |
| ] | |
| return p, [ | |
| min(x[0] for x in p), | |
| min(x[1] for x in p), | |
| max(x[0] for x in p), | |
| max(x[1] for x in p), | |
| ] | |
| def grounding(text): | |
| blocks = [] | |
| for match in re.finditer(r"!\[(text|table|figure|title|formula)\]\(([^)]*)\)", text): | |
| try: | |
| box = [float(v) for v in match[2].split(",")] | |
| if ( | |
| len(box) != 4 | |
| or not all(0 <= v <= 1000 for v in box) | |
| or box[0] > box[2] | |
| or box[1] > box[3] | |
| ): | |
| raise ValueError() | |
| blocks.append(Block(type=match[1], bbox_1000=box)) | |
| except ValueError: | |
| raise SafeError("invalid_grounding") from None | |
| return blocks | |
| def markdown_pages(pages): | |
| return "\n\n".join( | |
| f"<!-- page:{p.page_index + 1} -->\n\n{p.markdown}" | |
| for p in sorted(pages, key=lambda p: p.page_index) | |
| ) | |