Docker support with tests for local ollama

.dockerignore

@@ -0,0 +1,32 @@
+# Git specific
+.git
+.gitignore
+
+# Python specific
+__pycache__/
+*.pyc
+*.pyo
+*.pyd
+*.egg-info/
+.Python
+env/
+venv/
+.env
+
+# IDE/Editor specific
+.vscode/
+.idea/
+*.swp
+*.swo
+
+# Ollama cache directory (if it somehow ends up in build context)
+.ollama_cache/
+
+# Docker specific files (should not be part of the image content itself)
+Dockerfile
+docker-compose.yml
+.dockerignore
+
+# OS specific
+.DS_Store
+Thumbs.db

.env.example

@@ -0,0 +1,14 @@
+# LLM Configuration
+LLM_PROVIDER="ollama"
+LLM_BACKEND_URL="http://localhost:11434/v1" # For Ollama running in the same container, /v1 added for OpenAI compatibility
+LLM_DEEP_THINK_MODEL="qwen3:0.6b"
+LLM_QUICK_THINK_MODEL="qwen3:0.6b"
+OPENAI_API_KEY="ollama-key" # Optional, if you want to use OpenAI models or ollama models with OpenAI API compatibility
+
+# Agent Configuration
+MAX_DEBATE_ROUNDS="1"
+ONLINE_TOOLS="False" # Set to True if you want to enable tools that access the internet
+
+# Note: For local Docker Compose when Ollama runs on the host machine (not in container),
+# you might use LLM_BACKEND_URL="http://host.docker.internal:11434/v1"
+# The current docker-compose setup runs Ollama inside the app service.

.gitattributes

@@ -0,0 +1 @@
+docker-entrypoint.sh text eol=lf

Docker-readme.md

@@ -0,0 +1,94 @@
+# Local run with Docker or docker-compse
+
+## Environment Configuration with .env
+
+This project uses a `.env` file to manage environment-specific configurations for local development, especially when using Docker Compose. This allows you to customize settings without modifying version-controlled files like `docker-compose.yml`.
+
+### Setup
+1.  **Create your local `.env` file:**
+    Copy the example configuration to a new `.env` file:
+    ```bash
+    cp .env.example .env
+    ```
+2.  **Customize your `.env` file:**
+    Open the `.env` file in a text editor and modify the variables as needed for your local setup. For example, you might want to change LLM models or API keys (if applicable in the future).
+
+### How it Works with Docker Compose
+When you run `docker-compose up` or `docker-compose run`, Docker Compose automatically looks for a `.env` file in the project root directory (where `docker-compose.yml` is located) and loads the environment variables defined in it. These variables are then passed into the container environment for the `app` service.
+
+The `.env` file itself is ignored by Git (as specified in `.gitignore`), so your local configurations will not be committed to the repository.
+
+## Running with Docker
+
+This project supports running within a Docker container, which ensures a consistent environment for development and testing. 
+
+### Prerequisites
+- Docker installed and running on your system.
+
+### Build the Docker Image
+Navigate to the root directory of the project (where the `Dockerfile` is located) and run:
+```bash
+docker build -t tradingagents .
+```
+
+### Test local ollama setup
+To test ollama connectivity and local model:
+```bash
+docker run --rm \
+  -e LLM_PROVIDER="ollama" \
+  -e LLM_BACKEND_URL="http://localhost:11434/v1" \
+  -e LLM_DEEP_THINK_MODEL="qwen3:0.6b" \
+  -e LLM_QUICK_THINK_MODEL="qwen3:0.6b" \
+  -e MAX_DEBATE_ROUNDS="1" \
+  -e ONLINE_TOOLS="False" \
+  tradingagents \
+  python test_ollama_connection.py
+```
+**Note on Ollama for Local Docker:**
+The `LLM_BACKEND_URL` is set to `http://localhost:11434/v1`. This assumes you have Ollama running on your host machine and accessible at port 11434. '/v1' is added to url at the end for OpenAI api compatibility. 
+
+
+### Run the Main Application
+To run the `main.py` script (default command for the Docker image):
+```bash
+docker run --rm \
+  -e LLM_PROVIDER="ollama" \
+  -e LLM_BACKEND_URL="http://localhost:11434/v1" \
+  # Add other necessary environment variables for main.py
+  tradingagents
+```
+Adjust environment variables as needed for your local setup.
+
+### Using Docker Compose
+
+For a more streamlined local development experience, you can use Docker Compose. The `docker-compose.yml` file in the project root is configured to use the existing `Dockerfile`.
+
+**Build and Run Tests:**
+The default command in `docker-compose.yml` is set to run the test suite.
+```bash
+docker-compose up --build
+```
+This command will build the image (if it's not already built or if changes are detected) and then run the `pytest tests/test_main.py` command. The `--rm` flag is implicitly handled by `up` when the process finishes, or you can run:
+```bash
+docker-compose run --rm app # This will use the default command from docker-compose.yml
+```
+If you want to explicitly run the tests:
+```bash
+docker-compose run --rm app python test_ollama_connection.py
+```
+
+**Run the Main Application:**
+To run the `main.py` script, you can override the default command:
+```bash
+docker-compose run --rm app python main.py
+```
+Or, you can modify the `command` in `docker-compose.yml` if you primarily want `docker-compose up` to run the main application.
+
+**Environment Variables:**
+The necessary environment variables (like `LLM_PROVIDER`, `LLM_BACKEND_URL`, model names, etc.) are pre-configured in the `docker-compose.yml` for the `app` service. Ollama is started by the entrypoint script within the same container, so `LLM_BACKEND_URL` is set to `http://localhost:11434/v1`.
+
+**Live Code Reloading:**
+The current directory is mounted as a volume into the container at `/app`. This means changes you make to your local code will be reflected inside the container, which is useful for development. You might need to rebuild the image with `docker-compose build` or `docker-compose up --build` if you change dependencies in `requirements.txt` or modify the `Dockerfile` itself.
+
+**Ollama Model Caching:**
+To prevent re-downloading Ollama models, `docker-compose.yml` now mounts `./.ollama` on your host to `/app/.ollama` in the container. Models pulled by Ollama will be stored in `./.ollama/models` locally and persist across runs. Ensure this directory is in your `.gitignore`. If Docker has permission issues creating this directory, you might need to create it manually (`mkdir .ollama`).

Dockerfile

@@ -0,0 +1,82 @@
+# syntax=docker/dockerfile:1.4
+
+# Build stage for dependencies
+FROM python:3.9-slim-bookworm AS builder
+
+# Set environment variables for build
+ENV PYTHONDONTWRITEBYTECODE=1 \
+    PIP_DISABLE_PIP_VERSION_CHECK=on \
+    PIP_DEFAULT_TIMEOUT=100
+
+# Install build dependencies
+RUN --mount=type=cache,target=/var/cache/apt,sharing=locked \
+    --mount=type=cache,target=/var/lib/apt,sharing=locked \
+    apt-get update && apt-get install -y --no-install-recommends \
+    curl \
+    git \
+    && apt-get clean
+
+# Install Ollama in builder stage with cache mount for downloads
+RUN --mount=type=cache,target=/tmp/ollama-cache \
+    curl -fsSL https://ollama.com/install.sh | sh
+
+# Create virtual environment
+RUN python -m venv /opt/venv
+ENV PATH="/opt/venv/bin:$PATH"
+
+# Copy requirements and install Python dependencies
+COPY requirements.txt .
+RUN --mount=type=cache,target=/root/.cache/pip \
+    pip install --no-cache-dir -r requirements.txt
+
+# Runtime stage
+FROM python:3.9-slim-bookworm AS runtime
+
+# Set environment variables
+ENV PYTHONUNBUFFERED=1 \
+    PYTHONDONTWRITEBYTECODE=1 \
+    PIP_DISABLE_PIP_VERSION_CHECK=on \
+    PIP_DEFAULT_TIMEOUT=100 \
+    OLLAMA_HOST=0.0.0.0
+
+# Install runtime dependencies
+RUN --mount=type=cache,target=/var/cache/apt,sharing=locked \
+    --mount=type=cache,target=/var/lib/apt,sharing=locked \
+    apt-get update && apt-get install -y --no-install-recommends \
+    curl \
+    git \
+    && apt-get clean
+
+# Copy Ollama from builder stage instead of installing again
+COPY --from=builder /usr/local/bin/ollama /usr/local/bin/ollama
+
+# Copy virtual environment from builder stage
+COPY --from=builder /opt/venv /opt/venv
+ENV PATH="/opt/venv/bin:$PATH"
+
+# Create a non-root user and group
+RUN groupadd -r appuser && useradd -r -g appuser -s /bin/bash -d /app appuser
+
+# Create app directory
+WORKDIR /app
+
+# Copy the entrypoint script
+COPY docker-entrypoint.sh /usr/local/bin/docker-entrypoint.sh
+RUN chmod +x /usr/local/bin/docker-entrypoint.sh
+
+# Copy the application code
+COPY . .
+
+# Change ownership of the app directory to the non-root user
+RUN chown -R appuser:appuser /app
+
+# Switch to non-root user
+USER appuser
+
+# Set the entrypoint
+ENTRYPOINT ["docker-entrypoint.sh"]
+
+# Default command (can be overridden, e.g., by pytest command in CI)
+CMD ["python", "main.py"]
+
+EXPOSE 11434

README.md

@@ -192,6 +192,10 @@ print(decision)
 
 You can view the full list of configurations in `tradingagents/default_config.py`.
 
+## Docker usage and local ollama tests ##
+
+See [Docker Readme](./Docker-readme.md) for details.
+
 ## Contributing
 
 We welcome contributions from the community! Whether it's fixing a bug, improving documentation, or suggesting a new feature, your input helps make this project better. If you are interested in this line of research, please consider joining our open-source financial AI research community [Tauric Research](https://tauric.ai/).

build.sh

@@ -0,0 +1,22 @@
+#!/bin/bash
+
+# Enable BuildKit for faster builds
+export DOCKER_BUILDKIT=1
+
+# Build the image with BuildKit optimizations
+echo "Building with BuildKit optimizations..."
+docker build \
+    --progress=plain \
+    --tag tradingagents:latest \
+    .
+
+echo "Build completed!"
+echo ""
+echo "To run the container:"
+echo "docker run -it tradingagents2:latest"
+echo ""
+echo "To test Ollama connection first:"
+echo "docker run --rm -it tradingagents:latest python test_ollama_connection.py"
+echo ""
+echo "To build with additional BuildKit features:"
+echo "docker build --build-arg BUILDKIT_INLINE_CACHE=1 --tag tradingagents:latest ." 

docker-compose.yml

@@ -0,0 +1,27 @@
+version: '3.8' # Specify a version
+
+services:
+  app:
+    build: . # Use the Dockerfile in the current directory
+    volumes:
+      - .:/app # Mount current directory to /app in container for live code changes
+      - ./.ollama:/app/.ollama # Cache Ollama models
+    #environment:
+    #  - LLM_PROVIDER=ollama
+    #  - LLM_BACKEND_URL=http://localhost:11434/v1
+    #  - LLM_DEEP_THINK_MODEL=qwen3:0.6b
+    #  - LLM_QUICK_THINK_MODEL=qwen3:0.6b
+    #  - MAX_DEBATE_ROUNDS=1
+    #  - ONLINE_TOOLS=False
+    # The default command in the Dockerfile is `python main.py`.
+    # For running tests with compose, one can use:
+    # docker-compose run --rm app pytest tests/test_main.py
+    # Or, we can set a default command here to run tests:
+    env_file:
+      - .env # Load environment variables from files.env
+    #command: python test_ollama_connection.py # Uncomment to run a specific test script
+    # If you want `docker-compose up` to run tests and then exit, this command is fine.
+    # If you want `docker-compose up` to run main.py, change command or remove it to use Dockerfile's CMD.
+    # For more flexibility, users can override the command when using `docker-compose run`.
+    ports:
+      - "11434:11434" # Expose port 11434 for Ollama  

docker-entrypoint.sh

@@ -0,0 +1,49 @@
+#!/bin/bash
+set -e # Exit immediately if a command exits with a non-zero status.
+
+# Start Ollama serve in the background
+echo "Starting Ollama service..."
+ollama serve &
+OLLAMA_PID=$!
+
+# Wait for Ollama to be ready by checking the API endpoint
+echo "Waiting for Ollama to be ready..."
+max_attempts=30
+attempt=0
+while [ $attempt -lt $max_attempts ]; do
+    if curl -s http://localhost:11434/api/tags > /dev/null 2>&1; then
+        echo "Ollama is ready!"
+        break
+    fi
+    echo "Waiting for Ollama... (attempt $((attempt + 1))/$max_attempts)"
+    sleep 2
+    attempt=$((attempt + 1))
+done
+
+if [ $attempt -eq $max_attempts ]; then
+    echo "Error: Ollama failed to start within the expected time"
+    exit 1
+fi
+
+# Pull the required model. Use LLM_DEEP_THINK_MODEL, default to qwen:0.5b if not set.
+MODEL_TO_PULL=${LLM_DEEP_THINK_MODEL:-"qwen3:0.6b"}
+echo "Pulling Ollama model: $MODEL_TO_PULL..."
+ollama pull "$MODEL_TO_PULL"
+echo "Model $MODEL_TO_PULL pulled."
+ollama list # List models for verification
+
+# Test the connection before running the main application
+# TODO: run based on flag for testing
+echo "Testing Ollama connection..."
+python test_ollama_connection.py
+if [ $? -ne 0 ]; then
+    echo "Error: Ollama connection test failed"
+    exit 1
+fi
+
+echo "Ollama setup complete. Executing command: $@"
+# Execute the CMD or the command passed to docker run
+exec "$@"
+
+# Optional: clean up Ollama server on exit (might be complex with exec)
+# trap "echo 'Stopping Ollama service...'; kill $OLLAMA_PID; exit 0" SIGINT SIGTERM

main.py

@@ -1,21 +1,46 @@
+import os
 from tradingagents.graph.trading_graph import TradingAgentsGraph
 from tradingagents.default_config import DEFAULT_CONFIG
+from dotenv import load_dotenv
 
-# Create a custom config
+def run_analysis(config_overrides=None):
-config = DEFAULT_CONFIG.copy()
+    """
-config["llm_provider"] = "google"  # Use a different model
+    Initializes and runs a trading cycle with configurable overrides.
-config["backend_url"] = "https://generativelanguage.googleapis.com/v1"  # Use a different backend
+    """
-config["deep_think_llm"] = "gemini-2.0-flash"  # Use a different model
+    load_dotenv() # Load .env file variables
-config["quick_think_llm"] = "gemini-2.0-flash"  # Use a different model
-config["max_debate_rounds"] = 1  # Increase debate rounds
-config["online_tools"] = True  # Increase debate rounds
 
-# Initialize with custom config
+    config = DEFAULT_CONFIG.copy()
-ta = TradingAgentsGraph(debug=True, config=config)
 
-# forward propagate
+    # Override with environment variables if set
-_, decision = ta.propagate("NVDA", "2024-05-10")
+    config["llm_provider"] = os.environ.get("LLM_PROVIDER", config.get("llm_provider", "google"))
-print(decision)
+    config["backend_url"] = os.environ.get("LLM_BACKEND_URL", config.get("backend_url", "https://generativelanguage.googleapis.com/v1"))
+    config["deep_think_llm"] = os.environ.get("LLM_DEEP_THINK_MODEL", config.get("deep_think_llm", "gemini-2.0-flash"))
+    config["quick_think_llm"] = os.environ.get("LLM_QUICK_THINK_MODEL", config.get("quick_think_llm", "gemini-2.0-flash"))
+    config["max_debate_rounds"] = int(os.environ.get("MAX_DEBATE_ROUNDS", config.get("max_debate_rounds", 1)))
+    config["online_tools"] = os.environ.get("ONLINE_TOOLS", str(config.get("online_tools", True))).lower() == 'true'
 
-# Memorize mistakes and reflect
+
-# ta.reflect_and_remember(1000) # parameter is the position returns
+    # Apply overrides from function argument
+    if config_overrides:
+        config.update(config_overrides)
+
+    print("Using configuration:")
+    for key, value in config.items():
+        print(f"{key}: {value}")
+        
+    # Initialize with the final config
+    ta = TradingAgentsGraph(debug=True, config=config)
+
+    # Forward propagate
+    _, decision = ta.propagate("NVDA", "2024-05-10")
+    return decision
+
+if __name__ == "__main__":
+    # Example of running the trading analysis
+    # You can override specific configurations here if needed, e.g.:
+    # decision = run_trading_cyrun_analysiscle(config_overrides={"max_debate_rounds": 2})
+    decision = run_analysis()
+    print(decision)
+
+    # Memorize mistakes and reflect
+    # ta.reflect_and_remember(1000) # parameter is the position returns

requirements.txt

@@ -1,6 +1,8 @@
 typing-extensions
 langchain-openai
 langchain-experimental
+langchain_anthropic
+langchain_google_genai
 pandas
 yfinance
 praw
@@ -22,3 +24,7 @@ redis
 chainlit
 rich
 questionary
+ollama
+pytest
+python-dotenv
+

test_ollama_connection.py

@@ -0,0 +1,71 @@
+#!/usr/bin/env python3
+"""
+Simple test script to verify Ollama connection is working.
+"""
+
+import os
+import requests
+import time
+from openai import OpenAI
+
+def test_ollama_connection():
+    """Test if Ollama is accessible and responding."""
+    
+    # Get configuration from environment
+    backend_url = os.environ.get("LLM_BACKEND_URL", "http://localhost:11434/v1")
+    model = os.environ.get("LLM_DEEP_THINK_MODEL", "qwen2.5")
+    
+    print(f"Testing Ollama connection:")
+    print(f"  Backend URL: {backend_url}")
+    print(f"  Model: {model}")
+    
+    # Test 1: Check if Ollama API is responding
+    try:
+        response = requests.get(f"{backend_url.replace('/v1', '')}/api/tags", timeout=10)
+        if response.status_code == 200:
+            print("✅ Ollama API is responding")
+        else:
+            print(f"❌ Ollama API returned status code: {response.status_code}")
+            return False
+    except Exception as e:
+        print(f"❌ Failed to connect to Ollama API: {e}")
+        return False
+    
+    # Test 2: Check if the model is available
+    try:
+        response = requests.get(f"{backend_url.replace('/v1', '')}/api/tags", timeout=10)
+        models = response.json().get("models", [])
+        model_names = [m.get("name", "") for m in models]
+        
+        if model in model_names:
+            print(f"✅ Model '{model}' is available")
+        else:
+            print(f"❌ Model '{model}' not found. Available models: {model_names}")
+            return False
+    except Exception as e:
+        print(f"❌ Failed to check model availability: {e}")
+        return False
+    
+    # Test 3: Test OpenAI-compatible API
+    try:
+        client = OpenAI(base_url=backend_url, api_key="dummy")
+        response = client.chat.completions.create(
+            model=model,
+            messages=[{"role": "user", "content": "Hello, say 'test successful'"}],
+            max_tokens=50
+        )
+        print("✅ OpenAI-compatible API is working")
+        print(f"   Response: {response.choices[0].message.content}")
+        return True
+    except Exception as e:
+        print(f"❌ OpenAI-compatible API test failed: {e}")
+        return False
+
+if __name__ == "__main__":
+    success = test_ollama_connection()
+    if success:
+        print("\n🎉 All tests passed! Ollama is ready.")
+        exit(0)
+    else:
+        print("\n💥 Tests failed! Check Ollama configuration.")
+        exit(1)