FinRL

https://finrl.readthedocs.io/en/latest/
Admin
FinRL is an open-source framework for financial reinforcement learning that enables the development...
Tokens:7,754
Snippets:124
Trust Score:8.3
Update:3 months ago
Show doc for...
Tokens:
Raw
### Install and Install Pre-commit Hooks

Source: https://finrl.readthedocs.io/en/latest/developer_guide/contributing.html

Install the pre-commit tool and then install the pre-commit hooks for the project. This ensures code quality and consistency before committing.

```bash
pip install pre-commit
pre-commit install
```

--------------------------------

### Install FinRL on Windows

Source: https://finrl.readthedocs.io/en/latest/start/installation.html

Installs FinRL and its dependencies after cloning the repository on a Windows system.

```bash
cd FinRL
pip install .

```

--------------------------------

### Install box2d Alternative

Source: https://finrl.readthedocs.io/en/latest/start/installation.html

Provides an alternative installation command for box2d if the initial installation results in an AttributeError.

```bash
pip install box2d box2d-kengz

```

--------------------------------

### Install Ubuntu Dependencies

Source: https://finrl.readthedocs.io/en/latest/start/installation.html

Installs essential packages required for FinRL on Ubuntu systems.

```bash
sudo apt-get update && sudo apt-get install cmake libopenmpi-dev python3-dev zlib1g-dev libgl1-mesa-glx swig

```

--------------------------------

### Test FinRL Installation

Source: https://finrl.readthedocs.io/en/latest/start/installation.html

Runs a sample stock trading script to test the FinRL installation, particularly useful if using YahooFinance in China.

```bash
python Stock_NeurIPS2018.py

```

--------------------------------

### Install System Dependencies with Homebrew

Source: https://finrl.readthedocs.io/en/latest/start/installation.html

Installs cmake and openmpi, which are required system packages for FinRL on macOS.

```bash
brew install cmake openmpi

```

--------------------------------

### Install Ubuntu Dependencies

Source: https://finrl.readthedocs.io/en/latest/start/installation.html

Installs essential packages for OpenAI and other libraries within an Ubuntu terminal.

```bash
sudo apt-get update && sudo apt-get install cmake libopenmpi-dev python3-dev zlib1g-dev libgl1-mesa-glx

```

--------------------------------

### Install FinRL Development Version

Source: https://finrl.readthedocs.io/en/latest/start/installation.html

Installs the unstable development version of FinRL directly from its GitHub repository using pip.

```bash
pip install git+https://github.com/AI4Finance-Foundation/FinRL.git

```

--------------------------------

### Install box2d-py and Box2D

Source: https://finrl.readthedocs.io/en/latest/start/installation.html

Installs the box2d-py and Box2D Python packages, along with system dependencies swig.

```bash
brew install swig

```

```bash
pip install box2d-py

```

```bash
pip install box2d

```

```bash
pip install Box2D

```

--------------------------------

### Install Homebrew Package Manager

Source: https://finrl.readthedocs.io/en/latest/start/installation.html

Installs Homebrew, a package manager for macOS, which is required for installing system dependencies.

```bash
/bin/bash -c "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/HEAD/install.sh)"

```

--------------------------------

### Create and Activate Conda Environment

Source: https://finrl.readthedocs.io/en/latest/developer_guide/development_setup.html

Create a new Conda environment named 'ai4finance' with Python 3.8 and activate it. Then, navigate to the FinRL directory and install its dependencies.

```bash
cd ~/ai4finance
conda create --name ai4finance python=3.8
conda activate ai4finance

cd FinRL
pip install -r requirements.txt

```

--------------------------------

### Verify Anaconda Python Path

Source: https://finrl.readthedocs.io/en/latest/start/installation.html

Check if your system's Python interpreter is correctly linked to the Anaconda installation.

```bash
/Users/your_user_name/opt/anaconda3/bin/python

```

--------------------------------

### Clone AI4Finance Repositories

Source: https://finrl.readthedocs.io/en/latest/developer_guide/development_setup.html

Clone the FinRL, ElegantRL, and FinRL-Meta repositories to your local machine. Ensure you replace '[your_github_username]' with your actual GitHub username.

```bash
mkdir ~/ai4finance
cd ~/ai4finance
git clone https://github.com/[your_github_username]/FinRL.git
git clone https://github.com/[your_github_username]/ElegantRL.git
git clone https://github.com/[your_github_username]/FinRL-Meta.git

```

--------------------------------

### Build and Test Docker Container

Source: https://finrl.readthedocs.io/en/latest/developer_guide/contributing.html

Build the Docker container for the project and then run the tests within that container. This provides a consistent testing environment.

```bash
./docker/bin/build_container.sh ./docker/bin/test.sh
```

--------------------------------

### FinRL Main Script for Training, Testing, and Trading

Source: https://finrl.readthedocs.io/en/latest/start/quick_start.html

This script orchestrates the FinRL workflow. It parses command-line arguments to determine the mode (train, test, or trade) and configures the environment and training parameters accordingly. Ensure necessary directories are created before execution.

```python
import os
from typing import List
from argparse import ArgumentParser
from finrl import config
from finrl.config_tickers import DOW_30_TICKER
from finrl.config import (
    DATA_SAVE_DIR,
    TRAINED_MODEL_DIR,
    TENSORBOARD_LOG_DIR,
    RESULTS_DIR,
    INDICATORS,
    TRAIN_START_DATE,
    TRAIN_END_DATE,
    TEST_START_DATE,
    TEST_END_DATE,
    TRADE_START_DATE,
    TRADE_END_DATE,
    ERL_PARAMS,
    RLlib_PARAMS,
    SAC_PARAMS,
    ALPACA_API_KEY,
    ALPACA_API_SECRET,
    ALPACA_API_BASE_URL,
)

# construct environment
from finrl.meta.env_stock_trading.env_stocktrading_np import StockTradingEnv


def build_parser():
    parser = ArgumentParser()
    parser.add_argument(
        "--mode",
        dest="mode",
        help="start mode, train, download_data" " backtest",
        metavar="MODE",
        default="train",
    )
    return parser


# "./" will be added in front of each directory
def check_and_make_directories(directories: List[str]):
    for directory in directories:
        if not os.path.exists("./" + directory):
            os.makedirs("./" + directory)



def main():
    parser = build_parser()
    options = parser.parse_args()
    check_and_make_directories([DATA_SAVE_DIR, TRAINED_MODEL_DIR, TENSORBOARD_LOG_DIR, RESULTS_DIR])

    if options.mode == "train":
        from finrl import train

        env = StockTradingEnv

        # demo for elegantrl
        kwargs = {}  # in current meta, with respect yahoofinance, kwargs is {}. For other data sources, such as joinquant, kwargs is not empty
        train(
            start_date=TRAIN_START_DATE,
            end_date=TRAIN_END_DATE,
            ticker_list=DOW_30_TICKER,
            data_source="yahoofinance",
            time_interval="1D",
            technical_indicator_list=INDICATORS,
            drl_lib="elegantrl",
            env=env,
            model_name="ppo",
            cwd="./test_ppo",
            erl_params=ERL_PARAMS,
            break_step=1e5,
            kwargs=kwargs,
        )
    elif options.mode == "test":
        from finrl import test
        env = StockTradingEnv

        # demo for elegantrl
        kwargs = {}  # in current meta, with respect yahoofinance, kwargs is {}. For other data sources, such as joinquant, kwargs is not empty

        account_value_erl = test(
            start_date=TEST_START_DATE,
            end_date=TEST_END_DATE,
            ticker_list=DOW_30_TICKER,
            data_source="yahoofinance",
            time_interval="1D",
            technical_indicator_list=INDICATORS,
            drl_lib="elegantrl",
            env=env,
            model_name="ppo",
            cwd="./test_ppo",
            net_dimension=512,
            kwargs=kwargs,
        )
    elif options.mode == "trade":
        from finrl import trade
        env = StockTradingEnv
        kwargs = {}
        trade(
            start_date=TRADE_START_DATE,
            end_date=TRADE_END_DATE,
            ticker_list=DOW_30_TICKER,
            data_source="yahoofinance",
            time_interval="1D",
            technical_indicator_list=INDICATORS,
            drl_lib="elegantrl",
            env=env,
            model_name="ppo",
            API_KEY=ALPACA_API_KEY,
            API_SECRET=ALPACA_API_SECRET,
            API_BASE_URL=ALPACA_API_BASE_URL,
            trade_mode='backtesting',
            if_vix=True,
            kwargs=kwargs,
        )
    else:
        raise ValueError("Wrong mode.")



if __name__ == "__main__":
    main()

```

--------------------------------

### Create and Checkout a New Git Branch

Source: https://finrl.readthedocs.io/en/latest/developer_guide/development_setup.html

Navigate to the FinRL directory, create a new branch named 'branch_xxx' (replace with your desired branch name), and checkout this new branch for development.

```bash
cd ~/ai4finance
cd ./FinRL
git checkout -b branch_xxx

```

--------------------------------

### FinRL Project Structure

Source: https://finrl.readthedocs.io/en/latest/developer_guide/file_architecture.html

Overview of the FinRL library's file organization, illustrating the placement of core modules and applications within a three-layer architecture.

```text
FinRL
├── finrl (the main folder)
│   ├── applications
│           ├── cryptocurrency_trading
│           ├── high_frequency_trading
│           ├── portfolio_allocation
│           └── stock_trading
│   ├── agents
│           ├── elegantrl
│           ├── rllib
│           └── stablebaseline3
│   ├── meta
│           ├── data_processors
│           ├── env_cryptocurrency_trading
│           ├── env_portfolio_allocation
│           ├── env_stock_trading
│           ├── preprocessor
│           ├── data_processor.py
│           └── finrl_meta_config.py
│   ├── config.py
│   ├── config_tickers.py
│   ├── main.py
│   ├── train.py
│   ├── test.py
│   ├── trade.py
└───└── plot.py

```

--------------------------------

### Run Unit Tests Locally

Source: https://finrl.readthedocs.io/en/latest/developer_guide/contributing.html

Execute the project's unit tests locally using the Python unittest module. This helps verify code correctness before submitting changes.

```bash
python3 -m unittest discover
```

--------------------------------

### Running FinRL from Terminal

Source: https://finrl.readthedocs.io/en/latest/start/quick_start.html

These commands demonstrate how to execute the FinRL main script with different modes. The default ticker list is DOW_30_TICKER. For trade mode, ensure your Alpaca API credentials are set in config.py.

```bash
python main.py --mode=train
```

```bash
python main.py --mode=test
```

```bash
python main.py --mode=trade
```

--------------------------------

### Clone FinRL Repository

Source: https://finrl.readthedocs.io/en/latest/start/installation.html

Clones the FinRL GitHub repository to your local machine using git.

```bash
git clone https://github.com/AI4Finance-Foundation/FinRL.git

```
Tokens:
Raw