AlphaCrop

AlphaCrop is an advanced open-source platform for field-level crop analysis, combining interactive web tools with Python-based analytic backends for researchers and agricultural practitioners.

---

Capabilities

• Region Selection: Draw polygons on a basemap (Leaflet, in the web app) to define areas of interest in a field.
• Model Inference: Select classifiers for plant disease detection or crop yield forecasts.
• Data Source Integration: Toggle between satellite and IoT sensor network data as analysis sources (simulated in the PoC).
• End-to-end Analysis: Send regions and selected models/data sources to a Python backend for analysis (FastAPI/ML).
• IoT Sensor Data: Request and display simulated or real sensor readings within the selected field region.
• Extensible Models: Supports multiple underlying analysis models (e.g., Baseline Satellite or Agricultural Foundation Model).
• Proof-of-Concept Web UI: Fast, interactive interface built with React, Vite, and Leaflet.

---

Local Development Workflow

1. Run the Landing Page (repo root)

From the repository root, start a simple static server for index.html:

# Option A (Node)
npx serve . -l 8080

# Option B (Python)
python -m http.server 8080

Landing page URL: http://localhost:8080

2. Run the Field App (separate terminal)

In a second terminal:

cd field-app
npm install
npm run dev

Field app URL: http://localhost:3000/field-app/

3. Use Both Together

1. Open the landing page at http://localhost:8080.
2. Use the Field App button/link to open http://localhost:3000/field-app/ in development.
3. In the field app, draw your region of interest on the map.
4. Choose a classification model.
5. Select data sources (Satellite and/or IoT).
6. Run the analysis.

4. Backend API Connection

To connect to your Python backend:

• Create a .env.local file in field-app/ with:

  VITE_API_URL=http://localhost:8000
  

• Backend expects:

  • POST /analyze for analysis job (region, model, data sources).
  • POST /sensors for IoT sensor readings (polygon in GeoJSON).

See the minimal backend example in field-app/README.md for FastAPI code.

Production Deployment with Vercel

Instead of running two servers locally, you can deploy the entire project to Vercel:

# Install Vercel CLI
npm i -g vercel

# Deploy (from repo root)
vercel

Vercel automatically:

1. Builds the landing page (index.html) and the field app (field-app/dist/) once.
2. Routes traffic via vercel.json:
   • / → landing page
   • /field-app/* → field app
   • /images/* → public images
3. Hosts both under a single domain (e.g., alphacrop.vercel.app).

The field app's base: "/field-app/" in vite.config.js ensures all assets resolve correctly in production.

5. Models

ID	Name	Status
baseline	Baseline Satellite Model	✅ Available
afm	Agricultural Foundation Model	🚧 Under Development

---

Project Structure

• field-app/ — Modern React (Vite) front end for drawing, model selection, and analysis launch.
• (optional) backend/ — FastAPI Python backend for real or simulated model inference and sensor data.
• README.md — Main documentation (you’re here).

---

License

Proof of concept – research use only. See repository root for full license terms.

---

Further Info

• See field-app/README.md for full app setup instructions and example backend.