fix: handle async client and None reward values

.env.example

@@ -0,0 +1,15 @@
+# OpenEnv Environment Variables
+# Copy this to .env and fill in your values
+
+# Required - Hugging Face API Key
+HF_TOKEN=your_huggingface_token_here
+
+# Required - LLM Configuration
+API_BASE_URL=https://router.huggingface.co/v1
+MODEL_NAME=Qwen/Qwen2.5-72B-Instruct
+
+# Optional - Environment URL (defaults to localhost)
+ENV_URL=http://localhost:8000
+
+# Optional - For local testing without Docker
+OPENAI_API_KEY=your_openai_key_if_using_openai

data_analysis_env/.dockerignore

@@ -0,0 +1,5 @@
+__pycache__
+*.pyc
+.venv
+.git
+*.egg-info

data_analysis_env/Dockerfile

@@ -0,0 +1,17 @@
+FROM python:3.10-slim
+
+WORKDIR /app
+
+RUN pip install --no-cache-dir uv
+
+COPY server/requirements.txt /tmp/requirements.txt
+RUN pip install --no-cache-dir -r /tmp/requirements.txt
+
+COPY . /app/
+
+ENV PYTHONPATH=/app/server:/app
+ENV DATA_DIR=/app/server/data
+
+EXPOSE 8000
+
+CMD ["uvicorn", "app:app", "--host", "0.0.0.0", "--port", "8000"]

data_analysis_env/client.py

@@ -16,9 +16,8 @@ class DataAnalysisEnv(
     def __init__(
         self,
         base_url: str = "http://localhost:8000",
-        timeout: float = 30.0,
     ):
-        super().__init__(base_url=base_url, timeout=timeout)
+        super().__init__(base_url=base_url)
 
     def _step_payload(self, action: DataAnalysisAction) -> dict:
         return {

data_analysis_env/server/Dockerfile

@@ -4,14 +4,14 @@ WORKDIR /app
 
 RUN pip install --no-cache-dir uv
 
-COPY data_analysis_env/server/requirements.txt /tmp/requirements.txt
+COPY server/requirements.txt /tmp/requirements.txt
 RUN pip install --no-cache-dir -r /tmp/requirements.txt
 
-COPY data_analysis_env/ /app/
+COPY . /app/data_analysis_env/
 
-ENV PYTHONPATH=/app:$PYTHONPATH
-ENV DATA_DIR=/app/server/data
+ENV PYTHONPATH=/app
+ENV DATA_DIR=/app/data_analysis_env/data
 
 EXPOSE 8000
 
-CMD ["uvicorn", "server.app:app", "--host", "0.0.0.0", "--port", "8000"]
+CMD ["uvicorn", "data_analysis_env.server.app:app", "--host", "0.0.0.0", "--port", "8000"]

data_analysis_env/server/app.py

@@ -1,11 +1,16 @@
+import os
 from openenv.core.env_server import create_app
 
-from .data_analysis_environment import DataAnalysisEnvironment
-from ..models import DataAnalysisAction, DataAnalysisObservation
+try:
+    from server.data_analysis_environment import DataAnalysisEnvironment
+    from data_analysis_env.models import DataAnalysisAction, DataAnalysisObservation
+except ModuleNotFoundError:
+    from data_analysis_environment import DataAnalysisEnvironment
+    from models import DataAnalysisAction, DataAnalysisObservation
 
 
 def create_data_analysis_environment():
-    data_dir = "/app/data"
+    data_dir = os.getenv("DATA_DIR", "/app/server/data")
     return DataAnalysisEnvironment(data_dir=data_dir)
 
 

data_analysis_env/server/data_analysis_environment.py

@@ -8,12 +8,20 @@ import uuid
 
 from openenv.core.env_server import Environment
 
-from ..models import (
-    DataAnalysisAction,
-    DataAnalysisObservation,
-    DataAnalysisState,
-    AVAILABLE_TOOLS,
-)
+try:
+    from data_analysis_env.models import (
+        DataAnalysisAction,
+        DataAnalysisObservation,
+        DataAnalysisState,
+        AVAILABLE_TOOLS,
+    )
+except ModuleNotFoundError:
+    from models import (
+        DataAnalysisAction,
+        DataAnalysisObservation,
+        DataAnalysisState,
+        AVAILABLE_TOOLS,
+    )
 
 
 TASKS = {

inference.py

@@ -128,11 +128,12 @@ async def run_task(client: OpenAI, env: DataAnalysisEnv, task_name: str) -> dict
     rewards = []
     last_error = None
 
-    result = env.reset(task=task_name)
+    result = await env.reset(task=task_name)
     obs = result.observation
+    reward_val = obs.reward if obs.reward is not None else 0.0
 
     print(
-        f"[STEP] step={step} action=reset task={task_name} reward={obs.reward:.2f} done={result.done} error=null"
+        f"[STEP] step={step} action=reset task={task_name} reward={reward_val:.2f} done={result.done} error=null"
     )
 
     while not result.done and step < MAX_STEPS:
@@ -165,13 +166,14 @@ async def run_task(client: OpenAI, env: DataAnalysisEnv, task_name: str) -> dict
             )
             continue
 
-        result = env.step(action)
+        result = await env.step(action)
         obs = result.observation
-        rewards.append(obs.reward)
+        reward_val = obs.reward if obs.reward is not None else 0.0
+        rewards.append(reward_val)
 
         error_str = obs.error if obs.error else "null"
         print(
-            f"[STEP] step={step} action={action.tool}({action.parameters}) reward={obs.reward:.2f} done={result.done} error={error_str}"
+            f"[STEP] step={step} action={action.tool}({action.parameters}) reward={reward_val:.2f} done={result.done} error={error_str}"
         )
 
         if obs.error: