Upload 4 files

Dockerfile

@@ -0,0 +1,8 @@
+FROM python:3.10
+
+WORKDIR /app
+COPY . .
+
+RUN pip install openai
+
+CMD ["python", "inference.py"]

README.md

@@ -1,10 +1,26 @@
----
-title: Food Delivery Env
-emoji: 👀
-colorFrom: yellow
-colorTo: purple
-sdk: docker
-pinned: false
----
-
-Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
+# Food Delivery Dispatch Environment (OpenEnv)
+
+## Description
+This project simulates a food delivery system where an AI agent assigns orders to riders to maximize on-time deliveries.
+
+## State Space
+- Orders: location, prep_time, deadline
+- Riders: location
+
+## Action Space
+- assign_<order_id>_to_<rider_id>
+
+## Reward Function
+- On-time delivery: +1
+- Slight delay: +0.3
+- Late: -0.5
+- Invalid: -1
+
+## Tasks
+- Easy: 2 orders, 1 rider
+- Medium: 4 orders, 2 riders
+- Hard: 6 orders, random setup
+
+## Run
+```bash
+python inference.py

inference.py

@@ -0,0 +1,55 @@
+import os
+from env.delivery_env import DeliveryEnv
+from graders.grader import compute_score
+
+TASK = os.getenv("TASK", "easy")
+
+env = DeliveryEnv()
+state = env.reset(TASK)
+
+print(f"[START] task={TASK} env=food_delivery model=smart-urgency-agent")
+
+rewards = []
+step_count = 0
+
+while True:
+    orders = [o for o in state["orders"] if not o["assigned"]]
+    riders = state["riders"]
+
+    if not orders:
+        break
+
+    best_score = float("inf")
+    best_action = None
+
+    # 🔥 Evaluate all (order, rider) pairs
+    for order in orders:
+        for rider in riders:
+            distance = abs(rider["location"] - order["location"])
+            urgency = order["deadline"] - state["time"]
+
+            score = distance + urgency  # lower is better
+
+            if score < best_score:
+                best_score = score
+                best_action = (order, rider)
+
+    order, rider = best_action
+    action = f"assign_{order['id']}_to_{rider['id']}"
+
+    state, reward, done, info = env.step(action)
+
+    rewards.append(reward)
+    step_count += 1
+
+    error = info["error"] if info["error"] else "null"
+
+    print(f"[STEP] step={step_count} action={action} reward={reward:.2f} done={str(done).lower()} error={error}")
+
+    if done:
+        break
+
+max_possible = len(rewards) * 1.5  # adjusted for bonus rewards
+score = compute_score(sum(rewards), max_possible)
+
+print(f"[END] success=true steps={step_count} score={score:.2f} rewards={','.join(f'{r:.2f}' for r in rewards)}")

openenv.yaml

@@ -0,0 +1,7 @@
+name: food_delivery_env
+entry_point: env.delivery_env:DeliveryEnv
+
+tasks:
+  - easy
+  - medium
+  - hard