fix: add mandatory /metadata and /schema endpoints for OpenEnv validation and update task metadata

env/server/app.py

@@ -18,7 +18,77 @@ from fastapi.middleware.cors import CORSMiddleware
 from pydantic import BaseModel
 
 from env.server.environment import CodeDebugEnvironment
-from env.models import CodeDebugAction
+from env.models import CodeDebugAction, CodeDebugObservation, CodeDebugState
+
+
+# --- Metadata Definitions ---
+
+TASKS_METADATA = [
+    {
+        "id": "debug-add_numbers",
+        "name": "debug-add_numbers",
+        "description": "Fix subtraction -> addition bug",
+        "max_steps": 6,
+        "reward_range": [0.01, 0.99],
+        "grader": "env.server.grader:grade",
+    },
+    {
+        "id": "debug-find_max",
+        "name": "debug-find_max",
+        "description": "Fix < -> > comparison bug",
+        "max_steps": 6,
+        "reward_range": [0.01, 0.99],
+        "grader": "env.server.grader:grade",
+    },
+    {
+        "id": "debug-reverse_string",
+        "name": "debug-reverse_string",
+        "description": "Fix slice -> reverse bug",
+        "max_steps": 6,
+        "reward_range": [0.01, 0.99],
+        "grader": "env.server.grader:grade",
+    },
+    {
+        "id": "debug-fibonacci",
+        "name": "debug-fibonacci",
+        "description": "Fix n-3 -> n-2 recursion bug",
+        "max_steps": 6,
+        "reward_range": [0.01, 0.99],
+        "grader": "env.server.grader:grade",
+    },
+    {
+        "id": "debug-count_vowels",
+        "name": "debug-count_vowels",
+        "description": "Fix missing case-insensitive bug",
+        "max_steps": 6,
+        "reward_range": [0.01, 0.99],
+        "grader": "env.server.grader:grade",
+    },
+    {
+        "id": "debug-flatten_list",
+        "name": "debug-flatten_list",
+        "description": "Fix append -> extend (recursive flatten) bug",
+        "max_steps": 6,
+        "reward_range": [0.01, 0.99],
+        "grader": "env.server.grader:grade",
+    },
+    {
+        "id": "debug-merge_sorted",
+        "name": "debug-merge_sorted",
+        "description": "Fix missing remaining elements bug",
+        "max_steps": 6,
+        "reward_range": [0.01, 0.99],
+        "grader": "env.server.grader:grade",
+    },
+    {
+        "id": "debug-remove_duplicates",
+        "name": "debug-remove_duplicates",
+        "description": "Fix inverted condition bug",
+        "max_steps": 6,
+        "reward_range": [0.01, 0.99],
+        "grader": "env.server.grader:grade",
+    },
+]
 
 
 # --- Pydantic request/response schemas ---
@@ -107,6 +177,27 @@ async def health():
     return {"status": "healthy", "service": "codesensei-env"}
 
 
+@app.get("/metadata")
+async def get_metadata():
+    """Returns environment and task metadata for OpenEnv validation."""
+    return {
+        "name": "codesensei",
+        "version": "1.0.0",
+        "description": "GRPO-trained LLM code debugging environment",
+        "tasks": TASKS_METADATA,
+    }
+
+
+@app.get("/schema")
+async def get_schema():
+    """Returns the JSON schemas for project models."""
+    return {
+        "action": CodeDebugAction.model_json_schema(),
+        "observation": CodeDebugObservation.model_json_schema(),
+        "state": CodeDebugState.model_json_schema(),
+    }
+
+
 @app.get("/")
 async def root():
     """Root endpoint with API info."""
@@ -116,7 +207,9 @@ async def root():
         "endpoints": {
             "POST /reset": "Start a new episode",
             "POST /step": "Submit a code fix",
-            "GET /state?session_id=X": "Get episode state",
+            "GET /state": "Get episode state",
+            "GET /metadata": "Environment & task metadata",
+            "GET /schema": "JSON schemas for models",
             "WS /ws": "WebSocket interface (recommended)",
             "GET /health": "Health check",
         },

openenv.yaml

@@ -89,47 +89,55 @@ tasks:
     name: debug-add_numbers
     description: "Fix subtraction → addition bug"
     max_steps: 6
+    difficulty: "easy"
     reward_range: [0.01, 0.99]
     grader: "env.server.grader:grade"
   - id: debug-find_max
     name: debug-find_max
     description: "Fix < → > comparison bug"
     max_steps: 6
+    difficulty: "easy"
     reward_range: [0.01, 0.99]
     grader: "env.server.grader:grade"
   - id: debug-reverse_string
     name: debug-reverse_string
     description: "Fix slice → reverse bug"
     max_steps: 6
+    difficulty: "easy"
     reward_range: [0.01, 0.99]
     grader: "env.server.grader:grade"
   - id: debug-fibonacci
     name: debug-fibonacci
     description: "Fix n-3 → n-2 recursion bug"
     max_steps: 6
+    difficulty: "easy"
     reward_range: [0.01, 0.99]
     grader: "env.server.grader:grade"
   - id: debug-count_vowels
     name: debug-count_vowels
     description: "Fix missing case-insensitive bug"
     max_steps: 6
+    difficulty: "easy"
     reward_range: [0.01, 0.99]
     grader: "env.server.grader:grade"
   - id: debug-flatten_list
     name: debug-flatten_list
     description: "Fix append → extend (recursive flatten) bug"
     max_steps: 6
+    difficulty: "easy"
     reward_range: [0.01, 0.99]
     grader: "env.server.grader:grade"
   - id: debug-merge_sorted
     name: debug-merge_sorted
     description: "Fix missing remaining elements bug"
     max_steps: 6
+    difficulty: "easy"
     reward_range: [0.01, 0.99]
     grader: "env.server.grader:grade"
   - id: debug-remove_duplicates
     name: debug-remove_duplicates
     description: "Fix inverted condition bug"
     max_steps: 6
+    difficulty: "easy"
     reward_range: [0.01, 0.99]
     grader: "env.server.grader:grade"

requirements.txt

@@ -3,3 +3,5 @@ uvicorn[standard]>=0.24.0
 websockets>=12.0
 pydantic>=2.0.0
 openai>=1.0.0
+
+openenv-core>=0.2.0