harshraj22/croprl-workspace/code/extras/api - Storage Bucket

CropRL REST API

Local HTTP interface to the CropRL farming environment. Not required for the hackathon submission — this is for local testing and experimentation.

Quick Start

# From project root
uvicorn extras.api.app:app --reload --port 8000

Then visit http://localhost:8000/docs for the interactive Swagger UI.

Endpoints

Method	Path	Description
GET	/	Health check
GET	/tasks	List available task difficulties
GET	/actions	List all action IDs and names
POST	/reset	Start a new episode
GET	/state	Get current observation
POST	/step	Take an action

Example Usage (curl)

# Reset environment
curl -X POST http://localhost:8000/reset \
  -H "Content-Type: application/json" \
  -d '{"task": "easy", "seed": 42}'

# Check state
curl http://localhost:8000/state

# Plant chickpea (action 3)
curl -X POST http://localhost:8000/step \
  -H "Content-Type: application/json" \
  -d '{"action_id": 3}'

# Wait (action 0)
curl -X POST http://localhost:8000/step \
  -H "Content-Type: application/json" \
  -d '{"action_id": 0}'

# Harvest & sell (action 7)
curl -X POST http://localhost:8000/step \
  -H "Content-Type: application/json" \
  -d '{"action_id": 7}'

Example Usage (Python)

import requests

BASE = "http://localhost:8000"

# Reset
r = requests.post(f"{BASE}/reset", json={"task": "easy", "seed": 42})
obs = r.json()["observation"]

# Play a few steps
for action in [3, 0, 7, 1, 0, 0, 0, 7]:
    r = requests.post(f"{BASE}/step", json={"action_id": action})
    data = r.json()
    print(f"Step {data['step']}: {data['action_name']} → R={data['reward']:.0f}, Cash=₹{data['observation']['cash_balance']:,.0f}")