miaoqichuan/mqc-litigation-visual-redraw

This project takes a messy or rough diagram related to a legal case, such as a hand-drawn timeline or a poorly designed chart from a court document, and redraws it into a clean, professional-looking version suitable for use in legal filings. It is built as a skill for AI coding assistants, meaning an AI model reads the source diagram, extracts its content, and fills in a structured data file, while a separate Python script handles all the actual drawing and layout. The tool supports three categories of diagrams: timelines, flowcharts, and relationship diagrams, split into seven specific layout types. Timelines can be numbered, based on real dates, or shown as proportional bars for representing legal deadlines. Flowcharts illustrate case procedures or legal reasoning paths. Relationship diagrams show connections between parties, ownership structures, or hierarchical control, and there is also a side-by-side comparison table layout for contrasting two rulings or legal strategies. A core design decision is that the AI model only produces a JSON description of the content, never the visual coordinates or layout itself. All positioning, spacing, and rendering is handled by deterministic Python scripts, which the README says makes the output consistent even when using a weaker AI model. The workflow includes a manual confirmation step, where the extracted content is shown to the user before the final image is rendered, so nothing gets misread from the source without a chance to catch it. The visual style is intentionally restrained: grayscale with a single accent color of dark red, and no blue. Output is produced as both an editable SVG file and a PNG preview. The project requires Python 3 with no third-party packages, though Graphviz is needed for flowcharts and relationship diagrams. It is released under the MIT license.