Distortion-Aware MapFix

MapFix Spatial is an interactive MVP for correcting distorted geospatial coordinates and rendering a cleaner projection.

Live GitHub Pages demo:

https://arunshar.github.io/mapfix-spatial/

The browser includes a deterministic fallback correction engine, and server.py adds a live OpenAI backend:

• /api/correct computes the MapFix correction server-side and asks GPT for a concise geospatial analysis.
• /api/render-map calls the Image API to render a clean corrected map projection preview.
• OPENAI_API_KEY stays on the server and is never exposed to browser code.

Run locally

From this folder:

OPENAI_API_KEY=your_key_here python server.py

Then open:

http://127.0.0.1:4173

Without OPENAI_API_KEY, the page still runs as a local deterministic demo and the UI will show that the key is missing.

Optional model overrides:

OPENAI_TEXT_MODEL=gpt-5.5
OPENAI_IMAGE_MODEL=gpt-image-1
OPENAI_IMAGE_SIZE=1536x1024
OPENAI_IMAGE_QUALITY=medium

Deploy

For the live backend, deploy this folder as a Python/Docker web service. The included Dockerfile and render.yaml are ready for Render-style Docker hosting.

Required environment variable:

• OPENAI_API_KEY

The static-only files can still be deployed to Vercel, Netlify, Cloudflare Pages, or GitHub Pages, but GPT/Image API features require server.py or an equivalent serverless API.

Files

• index.html - app shell
• styles.css - responsive dashboard styling
• app.js - coordinate distortion, correction, rendering, metrics, and PNG export
• server.py - OpenAI-backed API and static-file server
• Dockerfile - container entrypoint for hosted backend deploys