Materials for Thematic Map Generation System (code, data, trained weights)

This upload contains the source code, example data, and supporting assets for an intelligent thematic mapping system that generates publication-ready maps from GeoJSON data through a hybrid cartographic workflow. The public source repository is available at [ecutxcq/Thematic-map](https://github.com/ecutxcq/Thematic-map), and the ControlNet model used by the system is available at [chou2024/thematic-map-controlnet-sd15](https://huggingface.co/chou2024/thematic-map-controlnet-sd15). The system combines deterministic geospatial preprocessing with AI-assisted visual generation, including ControlNet-guided thematic rendering, symbol synthesis and overlay, and whitespace-aware layout optimization. It supports quantitative, qualitative, and isopleth-style thematic maps, as well as point-based and regional symbolization, and provides an interactive browser interface for manual refinement before export. The software is implemented as a local browser-based application with a Python backend and a JavaScript frontend. Core processing relies on geospatial and image-processing libraries for GeoJSON handling, rasterization, coordinate transformation, symbol placement, and final layout composition. The repository also includes example GeoJSON datasets, training-data resources, and the lightweight LoRA symbol model used by the system. Large pretrained base-model weights are not included in this archive. Users should download the required Stable Diffusion and ControlNet model files separately from their original sources and place them in the expected local directories before running the full inference pipeline. This release is intended to support reproducibility, inspection of the system workflow, and reuse of the intelligent thematic mapping framework in future cartographic research and application development.