Adding base setup of the OpenEnv

README.md

@@ -1,11 +0,0 @@
----
-title: Voice Agent
-emoji: 🌍
-colorFrom: gray
-colorTo: gray
-sdk: docker
-pinned: false
-license: mit
----
-
-Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference

customer_env/README.md

@@ -0,0 +1,255 @@
+---
+title: Customer Env Environment Server
+emoji: 🕰️
+colorFrom: red
+colorTo: blue
+sdk: docker
+pinned: false
+app_port: 8000
+base_path: /web
+tags:
+  - openenv
+---
+
+# Customer Env Environment
+
+A simple test environment that echoes back messages. Perfect for testing the env APIs as well as demonstrating environment usage patterns.
+
+## Quick Start
+
+The simplest way to use the Customer Env environment is through the `CustomerEnv` class:
+
+```python
+from customer_env import CustomerAction, CustomerEnv
+
+try:
+    # Create environment from Docker image
+    customer_envenv = CustomerEnv.from_docker_image("customer_env-env:latest")
+
+    # Reset
+    result = customer_envenv.reset()
+    print(f"Reset: {result.observation.echoed_message}")
+
+    # Send multiple messages
+    messages = ["Hello, World!", "Testing echo", "Final message"]
+
+    for msg in messages:
+        result = customer_envenv.step(CustomerAction(message=msg))
+        print(f"Sent: '{msg}'")
+        print(f"  → Echoed: '{result.observation.echoed_message}'")
+        print(f"  → Length: {result.observation.message_length}")
+        print(f"  → Reward: {result.reward}")
+
+finally:
+    # Always clean up
+    customer_envenv.close()
+```
+
+That's it! The `CustomerEnv.from_docker_image()` method handles:
+- Starting the Docker container
+- Waiting for the server to be ready
+- Connecting to the environment
+- Container cleanup when you call `close()`
+
+## Building the Docker Image
+
+Before using the environment, you need to build the Docker image:
+
+```bash
+# From project root
+docker build -t customer_env-env:latest -f server/Dockerfile .
+```
+
+## Deploying to Hugging Face Spaces
+
+You can easily deploy your OpenEnv environment to Hugging Face Spaces using the `openenv push` command:
+
+```bash
+# From the environment directory (where openenv.yaml is located)
+openenv push
+
+# Or specify options
+openenv push --namespace my-org --private
+```
+
+The `openenv push` command will:
+1. Validate that the directory is an OpenEnv environment (checks for `openenv.yaml`)
+2. Prepare a custom build for Hugging Face Docker space (enables web interface)
+3. Upload to Hugging Face (ensuring you're logged in)
+
+### Prerequisites
+
+- Authenticate with Hugging Face: The command will prompt for login if not already authenticated
+
+### Options
+
+- `--directory`, `-d`: Directory containing the OpenEnv environment (defaults to current directory)
+- `--repo-id`, `-r`: Repository ID in format 'username/repo-name' (defaults to 'username/env-name' from openenv.yaml)
+- `--base-image`, `-b`: Base Docker image to use (overrides Dockerfile FROM)
+- `--private`: Deploy the space as private (default: public)
+
+### Examples
+
+```bash
+# Push to your personal namespace (defaults to username/env-name from openenv.yaml)
+openenv push
+
+# Push to a specific repository
+openenv push --repo-id my-org/my-env
+
+# Push with a custom base image
+openenv push --base-image ghcr.io/meta-pytorch/openenv-base:latest
+
+# Push as a private space
+openenv push --private
+
+# Combine options
+openenv push --repo-id my-org/my-env --base-image custom-base:latest --private
+```
+
+After deployment, your space will be available at:
+`https://huggingface.co/spaces/<repo-id>`
+
+The deployed space includes:
+- **Web Interface** at `/web` - Interactive UI for exploring the environment
+- **API Documentation** at `/docs` - Full OpenAPI/Swagger interface
+- **Health Check** at `/health` - Container health monitoring
+- **WebSocket** at `/ws` - Persistent session endpoint for low-latency interactions
+
+## Environment Details
+
+### Action
+**CustomerAction**: Contains a single field
+- `message` (str) - The message to echo back
+
+### Observation
+**CustomerObservation**: Contains the echo response and metadata
+- `echoed_message` (str) - The message echoed back
+- `message_length` (int) - Length of the message
+- `reward` (float) - Reward based on message length (length × 0.1)
+- `done` (bool) - Always False for echo environment
+- `metadata` (dict) - Additional info like step count
+
+### Reward
+The reward is calculated as: `message_length × 0.1`
+- "Hi" → reward: 0.2
+- "Hello, World!" → reward: 1.3
+- Empty message → reward: 0.0
+
+## Advanced Usage
+
+### Connecting to an Existing Server
+
+If you already have a Customer Env environment server running, you can connect directly:
+
+```python
+from customer_env import CustomerEnv
+
+# Connect to existing server
+customer_envenv = CustomerEnv(base_url="<ENV_HTTP_URL_HERE>")
+
+# Use as normal
+result = customer_envenv.reset()
+result = customer_envenv.step(CustomerAction(message="Hello!"))
+```
+
+Note: When connecting to an existing server, `customer_envenv.close()` will NOT stop the server.
+
+### Using the Context Manager
+
+The client supports context manager usage for automatic connection management:
+
+```python
+from customer_env import CustomerAction, CustomerEnv
+
+# Connect with context manager (auto-connects and closes)
+with CustomerEnv(base_url="http://localhost:8000") as env:
+    result = env.reset()
+    print(f"Reset: {result.observation.echoed_message}")
+    # Multiple steps with low latency
+    for msg in ["Hello", "World", "!"]:
+        result = env.step(CustomerAction(message=msg))
+        print(f"Echoed: {result.observation.echoed_message}")
+```
+
+The client uses WebSocket connections for:
+- **Lower latency**: No HTTP connection overhead per request
+- **Persistent session**: Server maintains your environment state
+- **Efficient for episodes**: Better for many sequential steps
+
+### Concurrent WebSocket Sessions
+
+The server supports multiple concurrent WebSocket connections. To enable this,
+modify `server/app.py` to use factory mode:
+
+```python
+# In server/app.py - use factory mode for concurrent sessions
+app = create_app(
+    CustomerEnvironment,  # Pass class, not instance
+    CustomerAction,
+    CustomerObservation,
+    max_concurrent_envs=4,  # Allow 4 concurrent sessions
+)
+```
+
+Then multiple clients can connect simultaneously:
+
+```python
+from customer_env import CustomerAction, CustomerEnv
+from concurrent.futures import ThreadPoolExecutor
+
+def run_episode(client_id: int):
+    with CustomerEnv(base_url="http://localhost:8000") as env:
+        result = env.reset()
+        for i in range(10):
+            result = env.step(CustomerAction(message=f"Client {client_id}, step {i}"))
+        return client_id, result.observation.message_length
+
+# Run 4 episodes concurrently
+with ThreadPoolExecutor(max_workers=4) as executor:
+    results = list(executor.map(run_episode, range(4)))
+```
+
+## Development & Testing
+
+### Direct Environment Testing
+
+Test the environment logic directly without starting the HTTP server:
+
+```bash
+# From the server directory
+python3 server/customer_env_environment.py
+```
+
+This verifies that:
+- Environment resets correctly
+- Step executes actions properly
+- State tracking works
+- Rewards are calculated correctly
+
+### Running Locally
+
+Run the server locally for development:
+
+```bash
+uvicorn server.app:app --reload
+```
+
+## Project Structure
+
+```
+customer_env/
+├── .dockerignore         # Docker build exclusions
+├── __init__.py            # Module exports
+├── README.md              # This file
+├── openenv.yaml           # OpenEnv manifest
+├── pyproject.toml         # Project metadata and dependencies
+├── uv.lock                # Locked dependencies (generated)
+├── client.py              # CustomerEnv client
+├── models.py              # Action and Observation models
+└── server/
+    ├── __init__.py        # Server module exports
+    ├── customer_env_environment.py  # Core environment logic
+    ├── app.py             # FastAPI application (HTTP + WebSocket endpoints)
+    └── Dockerfile         # Container image definition
+```

customer_env/__init__.py

@@ -0,0 +1,16 @@
+# Copyright (c) Meta Platforms, Inc. and affiliates.
+# All rights reserved.
+#
+# This source code is licensed under the BSD-style license found in the
+# LICENSE file in the root directory of this source tree.
+
+"""Customer Env Environment."""
+
+from .client import CustomerEnv
+from .models import CustomerAction, CustomerObservation
+
+__all__ = [
+    "CustomerAction",
+    "CustomerObservation",
+    "CustomerEnv",
+]

customer_env/client.py

@@ -0,0 +1,63 @@
+# Copyright (c) Meta Platforms, Inc. and affiliates.
+# All rights reserved.
+
+"""Customer Env Environment Client."""
+
+from typing import Dict, Any
+
+from openenv.core import EnvClient
+from openenv.core.client_types import StepResult
+from openenv.core.env_server.types import State
+
+from models import CustomerAction, CustomerObservation
+
+
+class CustomerEnv(
+        EnvClient[CustomerAction, CustomerObservation, State]
+):
+    """
+    Client for the Customer Env Environment (Banking POMDP).
+
+    This client maintains a persistent WebSocket connection to the environment server,
+    enabling efficient multi-step interactions with lower latency.
+    """
+
+    def _step_payload(self, action: CustomerAction) -> Dict[str, Any]:
+        """
+        Convert CustomerAction to JSON payload for step message.
+        """
+        return {
+            "action_type": action.action_type,
+            "content": action.content,
+            "tool_args": action.tool_args,
+        }
+
+    def _parse_result(self, payload: Dict[str, Any]) -> StepResult[CustomerObservation]:
+        """
+        Parse server response into StepResult[CustomerObservation].
+        """
+        obs_data = payload.get("observation", {})
+        
+        observation = CustomerObservation(
+            customer_reply=obs_data.get("customer_reply"),
+            tool_response=obs_data.get("tool_response"),
+            conversation_history=obs_data.get("conversation_history", ""),
+            done=payload.get("done", False),
+            reward=payload.get("reward", 0.0),
+            metadata=obs_data.get("metadata", {}),
+        )
+
+        return StepResult(
+            observation=observation,
+            reward=payload.get("reward", 0.0),
+            done=payload.get("done", False),
+        )
+
+    def _parse_state(self, payload: Dict[str, Any]) -> State:
+        """
+        Parse server response into State object.
+        """
+        return State(
+            episode_id=payload.get("episode_id"),
+            step_count=payload.get("step_count", 0),
+        )

customer_env/models.py

@@ -0,0 +1,16 @@
+from pydantic import BaseModel, Field
+from typing import Optional, Dict, Any
+
+class CustomerAction(BaseModel):
+    action_type: str = Field(..., description="Must be 'speak', 'tool_call', or 'end_call'")
+    content: str = Field(..., description="The spoken text or the name of the tool")
+    tool_args: Dict[str, Any] = Field(default_factory=dict, description="Arguments for the tool")
+
+class CustomerObservation(BaseModel):
+    customer_reply: Optional[str] = Field(None, description="What the customer said")
+    tool_response: Optional[str] = Field(None, description="Result of the tool call")
+    conversation_history: str = Field(..., description="Full transcript of the episode")
+    
+    done: bool = Field(False, description="Whether the episode has ended")
+    reward: float = Field(0.0, description="Reward received for this step")
+    metadata: Dict[str, Any] = Field(default_factory=dict)

customer_env/openenv.yaml

@@ -0,0 +1,7 @@
+spec_version: 1
+name: customer_env
+type: space
+runtime: fastapi
+app: server.app:app
+port: 8000
+

customer_env/pyproject.toml

@@ -0,0 +1,45 @@
+# Copyright (c) Meta Platforms, Inc. and affiliates.
+# All rights reserved.
+#
+# This source code is licensed under the BSD-style license found in the
+# LICENSE file in the root directory of this source tree.
+
+[build-system]
+requires = ["setuptools>=45", "wheel"]
+build-backend = "setuptools.build_meta"
+
+[project]
+name = "openenv-customer_env"
+version = "0.1.0"
+description = "Customer Env environment for OpenEnv"
+requires-python = ">=3.10"
+dependencies = [
+    # Core OpenEnv runtime (provides FastAPI server + HTTP client types)
+    # install from github
+    # "openenv-core[core] @ git+https://github.com/meta-pytorch/OpenEnv.git",
+    "openenv-core[core]>=0.2.1",
+    # Environment-specific dependencies
+    # Add all dependencies needed for your environment here
+    # Examples:
+    # "numpy>=1.19.0",
+    # "torch>=2.0.0",
+    # "gymnasium>=0.29.0",
+    # "openspiel>=1.0.0",
+    # "smolagents>=1.22.0,<2",
+]
+
+[project.optional-dependencies]
+dev = [
+    "pytest>=8.0.0",
+    "pytest-cov>=4.0.0",
+]
+
+[project.scripts]
+# Server entry point - enables running via: uv run --project . server
+# or: python -m customer_env.server.app
+server = "customer_env.server.app:main"
+
+[tool.setuptools]
+include-package-data = true
+packages = ["customer_env", "customer_env.server"]
+package-dir = { "customer_env" = ".", "customer_env.server" = "server" }

customer_env/server/Dockerfile

@@ -0,0 +1,80 @@
+# Copyright (c) Meta Platforms, Inc. and affiliates.
+# All rights reserved.
+#
+# This source code is licensed under the BSD-style license found in the
+# LICENSE file in the root directory of this source tree.
+
+# Multi-stage build using openenv-base
+# This Dockerfile is flexible and works for both:
+# - In-repo environments (with local OpenEnv sources)
+# - Standalone environments (with openenv from PyPI/Git)
+# The build script (openenv build) handles context detection and sets appropriate build args.
+
+ARG BASE_IMAGE=ghcr.io/meta-pytorch/openenv-base:latest
+FROM ${BASE_IMAGE} AS builder
+
+WORKDIR /app
+
+# Ensure git is available (required for installing dependencies from VCS)
+RUN apt-get update && \
+    apt-get install -y --no-install-recommends git && \
+    rm -rf /var/lib/apt/lists/*
+
+# Build argument to control whether we're building standalone or in-repo
+ARG BUILD_MODE=in-repo
+ARG ENV_NAME=customer_env
+
+# Copy environment code (always at root of build context)
+COPY . /app/env
+
+# For in-repo builds, openenv is already vendored in the build context
+# For standalone builds, openenv will be installed via pyproject.toml
+WORKDIR /app/env
+
+# Ensure uv is available (for local builds where base image lacks it)
+RUN if ! command -v uv >/dev/null 2>&1; then \
+        curl -LsSf https://astral.sh/uv/install.sh | sh && \
+        mv /root/.local/bin/uv /usr/local/bin/uv && \
+        mv /root/.local/bin/uvx /usr/local/bin/uvx; \
+    fi
+    
+# Install dependencies using uv sync
+# If uv.lock exists, use it; otherwise resolve on the fly
+RUN --mount=type=cache,target=/root/.cache/uv \
+    if [ -f uv.lock ]; then \
+        uv sync --frozen --no-install-project --no-editable; \
+    else \
+        uv sync --no-install-project --no-editable; \
+    fi
+
+RUN --mount=type=cache,target=/root/.cache/uv \
+    if [ -f uv.lock ]; then \
+        uv sync --frozen --no-editable; \
+    else \
+        uv sync --no-editable; \
+    fi
+
+# Final runtime stage
+FROM ${BASE_IMAGE}
+
+WORKDIR /app
+
+# Copy the virtual environment from builder
+COPY --from=builder /app/env/.venv /app/.venv
+
+# Copy the environment code
+COPY --from=builder /app/env /app/env
+
+# Set PATH to use the virtual environment
+ENV PATH="/app/.venv/bin:$PATH"
+
+# Set PYTHONPATH so imports work correctly
+ENV PYTHONPATH="/app/env:$PYTHONPATH"
+
+# Health check
+HEALTHCHECK --interval=30s --timeout=3s --start-period=5s --retries=3 \
+    CMD curl -f http://localhost:8000/health || exit 1
+
+# Run the FastAPI server
+# The module path is constructed to work with the /app/env structure
+CMD ["sh", "-c", "cd /app/env && uvicorn server.app:app --host 0.0.0.0 --port 8000"]

customer_env/server/__init__.py

@@ -0,0 +1,11 @@
+# Copyright (c) Meta Platforms, Inc. and affiliates.
+# All rights reserved.
+#
+# This source code is licensed under the BSD-style license found in the
+# LICENSE file in the root directory of this source tree.
+
+"""Customer Env environment server components."""
+
+from .customer_env import CustomerEnvironment
+
+__all__ = ["CustomerEnvironment"]

customer_env/server/app.py

@@ -0,0 +1,82 @@
+# Copyright (c) Meta Platforms, Inc. and affiliates.
+# All rights reserved.
+#
+# This source code is licensed under the BSD-style license found in the
+# LICENSE file in the root directory of this source tree.
+
+"""
+FastAPI application for the Customer Env Environment.
+
+This module creates an HTTP server that exposes the CustomerEnvironment
+over HTTP and WebSocket endpoints, compatible with EnvClient.
+
+Endpoints:
+    - POST /reset: Reset the environment
+    - POST /step: Execute an action
+    - GET /state: Get current environment state
+    - GET /schema: Get action/observation schemas
+    - WS /ws: WebSocket endpoint for persistent sessions
+
+Usage:
+    # Development (with auto-reload):
+    uvicorn server.app:app --reload --host 0.0.0.0 --port 8000
+
+    # Production:
+    uvicorn server.app:app --host 0.0.0.0 --port 8000 --workers 4
+
+    # Or run directly:
+    python -m server.app
+"""
+
+try:
+    from openenv.core.env_server.http_server import create_app
+except Exception as e:  # pragma: no cover
+    raise ImportError(
+        "openenv is required for the web interface. Install dependencies with '\n    uv sync\n'"
+    ) from e
+
+# Import from local models.py (PYTHONPATH includes /app/env in Docker)
+from models import CustomerAction, CustomerObservation
+
+from .customer_env import CustomerEnvironment
+
+
+# Create the app with web interface and README integration
+app = create_app(
+    CustomerEnvironment,
+    CustomerAction,
+    CustomerObservation,
+    env_name="customer_env",
+    max_concurrent_envs=1,  # increase this number to allow more concurrent WebSocket sessions
+)
+
+
+def main(host: str = "0.0.0.0", port: int = 8000):
+    """
+    Entry point for direct execution via uv run or python -m.
+
+    This function enables running the server without Docker:
+        uv run --project . server
+        uv run --project . server --port 8001
+        python -m customer_env.server.app
+
+    Args:
+        host: Host address to bind to (default: "0.0.0.0")
+        port: Port number to listen on (default: 8000)
+
+    For production deployments, consider using uvicorn directly with
+    multiple workers:
+        uvicorn customer_env.server.app:app --workers 4
+    """
+    import uvicorn
+
+    uvicorn.run(app, host=host, port=port)
+
+
+if __name__ == "__main__":
+    import argparse
+
+    parser = argparse.ArgumentParser()
+    parser.add_argument("--port", type=int, default=8000)
+    args = parser.parse_args()
+    main(port=args.port)

customer_env/server/basic_scenarios.csv

@@ -0,0 +1,6 @@
+intent,persona,starting_utterance
+dispute_charge: $50 at CoffeeCloud,Frustrated and rushed,What is this CoffeeCloud charge on my account?
+travel_notice: going to Japan,Polite but confused,Hi, I'm going overseas next week and need to know if my card will work.
+card_replacement: lost at gym,Panicked,I lost my wallet at the gym! Please help!
+increase_limit: needs $5000 for wedding,Direct and formal,I would like to request a credit limit increase.
+reset_password: locked out of app,Elderly and confused,The app on my phone says I'm locked out.

customer_env/server/customer_env.py

@@ -0,0 +1,190 @@
+# Copyright (c) Meta Platforms, Inc. and affiliates.
+# All rights reserved.
+
+import random, csv, json
+from uuid import uuid4
+
+# Use explicit relative or local imports
+from models import CustomerAction, CustomerObservation
+from openenv.core.env_server.interfaces import Environment
+from openenv.core.env_server.types import State
+from openai import OpenAI
+
+
+local_llm = OpenAI(base_url="http://localhost:11434/v1", api_key="local-dev")
+MODEL_NAME = "llama3" 
+
+class CustomerEnvironment(Environment):
+    SUPPORTS_CONCURRENT_SESSIONS: bool = False
+
+    def __init__(self):
+        """Initialize the Customer POMDP environment."""
+        self._state = State(episode_id=str(uuid4()), step_count=0)
+        self._reset_count = 0
+        self.hidden_intent = ""
+        self.persona = ""
+        self.scenarios = []
+        # Fallback just in case the file is missing
+        default_scenario = {"intent": "unknown", "persona": "neutral", "starting_utterance": "I need help."}
+        self.conversation_history = ""
+
+
+        try:
+            with open("basic_scenarios.csv", mode="r", encoding="utf-8") as f:
+                reader = csv.DictReader(f)
+                for row in reader:
+                    self.scenarios.append(row)
+        except Exception as e:
+            print(f"Warning: Could not load scenarios.csv. {e}")
+            self.scenarios.append(default_scenario)
+
+    def reset(self) -> CustomerObservation:
+        """Reset the environment, pick a new hidden intent and persona."""
+        self._state = State(episode_id=str(uuid4()), step_count=0)
+        self._reset_count += 1
+
+        scenario = random.choice(self.scenarios)
+        self.hidden_intent = scenario["intent"]
+        self.persona = scenario["persona"]
+        start_msg = scenario["starting_utterance"]
+
+
+        self.conversation_history = f"System: Call connected.\nCustomer: {start_msg}"
+
+        return CustomerObservation(
+            customer_reply=start_msg,
+            tool_response=None,
+            conversation_history=self.conversation_history,
+            done=False,
+            reward=0.0,
+            metadata={"step": self._state.step_count}
+        )
+
+    def step(self, action: CustomerAction) -> CustomerObservation: 
+        self._state.step_count += 1
+        step_reward = 0.0
+        done = False
+        tool_response = None
+        customer_reply = None
+
+        if action.action_type == "tool_call":
+            tool_name = action.content
+            # Mocking the database lookup for now
+            if tool_name == "lookup_account":
+                tool_response = "{'status': 'verified', 'balance': '$500'}"
+                step_reward += 0.5 
+            else:
+                tool_response = f"Error: Tool '{tool_name}' not found."
+                step_reward -= 0.5 
+            
+            self.conversation_history += f"\nAgent [Action]: Used {tool_name}"
+            self.conversation_history += f"\nSystem: {tool_response}"
+
+        elif action.action_type == "speak":
+            self.conversation_history += f"\nAgent: {action.content}"
+            # Call made to the LLM
+            customer_reply = self._get_customer_reply(action.content)
+            self.conversation_history += f"\nCustomer: {customer_reply}"
+            step_reward -= 0.1 # Small penalty per turn to encourage efficiency
+
+        elif action.action_type == "end_call":
+            done = True
+            
+        if self._state.step_count >= 15:
+            done = True 
+
+        # THE JUDGE LLM EVALUATION 
+        if done:
+            final_score, reasoning = self._evaluate_with_judge()
+            step_reward += final_score
+            
+            metadata = {
+                "step": self._state.step_count,
+                "hidden_intent": self.hidden_intent,
+                "judge_reasoning": reasoning
+            }
+        else:
+            metadata = {"step": self._state.step_count}
+
+        return CustomerObservation(
+            customer_reply=customer_reply,
+            tool_response=tool_response,
+            conversation_history=self.conversation_history,
+            done=done,
+            reward=step_reward,
+            metadata=metadata
+        )
+
+    
+    def _evaluate_with_judge(self) -> tuple[float, str]:
+        """
+        Uses local LLM as a Judge to score the final transcript.
+        Returns a tuple of (score, reasoning).
+        """
+        judge_prompt = f"""You are an expert QA Judge for a banking call center.
+        Review the transcript and score the Agent's performance from -5.0 to +10.0.
+        
+        TRUE CUSTOMER INTENT: {self.hidden_intent}
+        
+        SCORING RUBRIC:
+        - +10.0: Perfect. Intent captured, correct tools used, issue resolved efficiently.
+        - +5.0: Okay. Found the intent but took too many turns or was awkward.
+        - 0.0: Neutral. Didn't solve the issue but didn't hallucinate.
+        - -5.0: Failure. Missed the intent, hallucinated tools, or was rude.
+        
+        TRANSCRIPT:
+        {self.conversation_history}
+        
+        Respond ONLY with a valid JSON object in this exact format:
+        {{"score": 8.5, "reasoning": "A brief explanation of why."}}
+        """
+
+        try:
+            response = local_llm.chat.completions.create(
+                model=MODEL_NAME,
+                messages=[{"role": "user", "content": judge_prompt}],
+                response_format={ "type": "json_object" },
+                temperature=0.0 
+            )
+            
+            result = json.loads(response.choices[0].message.content)
+            score = float(result.get("score", 0.0))
+            reasoning = result.get("reasoning", "No reasoning provided.")
+            
+            # Clamp the score just in case the LLM goes rogue
+            score = max(-5.0, min(10.0, score))
+            return score, reasoning
+            
+        except Exception as e:
+            # Fallback if the local LLM fails to generate valid JSON
+            print(f"Judge Error: {e}")
+            return -2.0, "Judge LLM failed to parse transcript."
+
+
+    def _get_customer_reply(self, agent_text: str) -> str:
+        """Uses local LLM to simulate the customer."""
+        system_prompt = f"""You are a banking customer calling support.
+        Your secret intent is: {self.hidden_intent}.
+        Your mood is: {self.persona}.
+        RULES:
+        1. Keep it under 2 sentences.
+        2. Do NOT reveal your full intent immediately. Wait for the agent to probe.
+        3. Respond naturally to what the agent just said.
+        
+        Conversation history:
+        {self.conversation_history}"""
+
+        response = local_llm.chat.completions.create(
+            model=MODEL_NAME,
+            messages=[
+                {"role": "system", "content": system_prompt},
+                {"role": "user", "content": agent_text}
+            ],
+            temperature=0.7,
+            max_tokens=60
+        )
+        return response.choices[0].message.content.strip()
+
+    @property
+    def state(self) -> State:
+        return self._state

customer_env/server/requirements.txt

@@ -0,0 +1,6 @@
+openenv[core]>=0.2.0
+fastapi>=0.115.0
+uvicorn>=0.24.0
+
+
+

customer_env/test_script.py

@@ -0,0 +1,58 @@
+
+from client import CustomerEnv
+from models import CustomerAction
+
+def run_test():
+    print("🔌 Connecting to Local OpenEnv Server at http://127.0.0.1:8000...\n")
+
+    # Initialize the client, then call .sync() to use it in a standard 'with' block
+    client = CustomerEnv(base_url="http://127.0.0.1:8000")
+    
+    with client.sync() as env:
+        
+        # --- 1. RESET ---
+        print("--- NEW EPISODE ---")
+        result = env.reset()
+        print(f"Customer: {result.observation.customer_reply}")
+        print(f"Initial Reward: {result.reward}")
+
+        # --- 2. STEP 1: PROBING ---
+        print("\n--- AGENT ACTION 1: SPEAK ---")
+        action1 = CustomerAction(
+            action_type="speak", 
+            content="I can help with that. Could I please get your account name?"
+        )
+        print(f"Agent: {action1.content}")
+        result = env.step(action1)
+        print(f"Customer Reply: {result.observation.customer_reply}")
+        print(f"Reward (Should be slightly negative for turns): {result.reward}")
+
+        # --- 3. STEP 2: TOOL USAGE ---
+        print("\n--- AGENT ACTION 2: TOOL CALL ---")
+        action2 = CustomerAction(
+            action_type="tool_call",
+            content="lookup_account",
+            tool_args={"name": "John Doe"}
+        )
+        print(f"Agent [Action]: Using tool '{action2.content}'")
+        result = env.step(action2)
+        print(f"System Response: {result.observation.tool_response}")
+        print(f"Reward (Should increase for tool usage): {result.reward}")
+
+        # --- 4. STEP 3: END CALL (Triggers Judge) ---
+        print("\n--- AGENT ACTION 3: END EPISODE ---")
+        action3 = CustomerAction(
+            action_type="end_call",
+            content="Thank you, your issue is resolved."
+        )
+        result = env.step(action3)
+        print(f"Episode Done: {result.done}")
+        print(f"Final Step Reward: {result.reward}")
+        print(f"Hidden Intent was: {result.observation.metadata.get('hidden_intent')}")
+
+        # Print the full transcript tracked by the environment
+        print("\n--- FULL CONVERSATION TRANSCRIPT ---")
+        print(result.observation.conversation_history)
+
+if __name__ == "__main__":
+    run_test()