True-African/Informal-address-resolver

Lightweight geospatial resolver for informal Rwanda delivery addresses using OpenStreetMap gazetteers, fuzzy landmark matching, text parsing, and offline-first logic. --- title: T1.2 Informal Address Resolver sdk: gradio app_file: app.py license: mit pinned: false --- # T1.2 Informal Address Resolver This project resolves informal delivery address descriptions into usable GPS coordinates. It is designed for Kigali-style logistics settings where addresses often mix English, French, and Kinyarwanda, and may describe places by nearby landmarks instead of formal street addresses. The resolver uses a local landmark gazetteer, fuzzy matching, simple direction rules, and geospatial offsets. For each input, it returns latitude, longitude, confidence, the matched landmark, and a short rationale explaining the decision. ## Output Shape ```python { "lat": float | None, "lon": float | None, "confidence": float, "matched_landmark": str | None, "rationale": str } ``` ## Setup Clone the repository and enter the project folder: ```bash git clone github.com cd Informal-address-resolver ``` If you are using the Hugging Face Space repository instead: ```bash git clone huggingface.co cd Informal-address-resolver ``` Create a virtual environment: ```bash python -m venv .venv ``` Activate it using the command for your terminal: ```bash # Windows Git Bash source .venv/Scripts/activate # Windows PowerShell .\.venv\Scripts\Activate.ps1 # Linux, macOS, or Colab source .venv/bin/activate ``` Install packages: ```bash python -m pip install --upgrade pip python -m pip install -r requirements.txt ``` If PowerShell blocks activation, run the same commands through the environment Python: ```powershell & .\.venv\Scripts\python.exe -m pip install --upgrade pip & .\.venv\Scripts\python.exe -m pip install -r requirements.txt ``` ## Required Demo ```bash python -c "from resolver import resolve; print(resolve('inyuma ya big pharmacy on RN3, red gate'))" ``` Example output: ```python { "lat": -1.9584, "lon": 30.1129, "confidence": 1.0, "matched_l …