TradzQAI
Trading environnement for RL agents, backtesting and training.
Live session with coinbasepro-python is finaly arrived !
-
Available sessions:
- Local
- 1M bar datasets any size.
- header should be ['Time', 'Open', 'High', 'Low', 'Close', 'Volume'] with ';' as separator
- Tick datasets any size.
- header should be ['Time', 'BID', 'ASK', 'VOL'] or ['Time', 'Price', 'Volume'] with ',' as separator
- 1M bar datasets any size.
- Live
- Gdax API
- Local
-
Available agents:
- DDPG
- DQFD
- DQN
- DQNN
- NAF
- PPO
- TRPO
- VPG
-
Available contract type:
- CFD
- Classic
TradzQAI has been inspired by q-trader.
Indicators lib come from pyti
More datasets available here
Status
Alpha in development
GUI rework on track
Getting Started
-
Dependencies :
- Tensorflow
- Tensorforce
- coinbasepro-python
- Pandas
- Numpy
- tqdm
- h5py
Installation :
pip install -r requirements.txt -
Running the project
Usage: python run.py -h (Show usage) python run.py -b agent_name (to manually build config file, it build config files from agent, default PPO) python run.py -s live (for live session) python run.py -m eval (for eval mode) python run.py -c config_dir/ # load config from directory, make sure you have agent, env and network json files in it python run.py (Run as default)When you run it for the first time, a config directory is created, you can change it to changes environnement settings and some agents settings. It save settings (env, agent, network) in a save directory, and create a new directory if make any changes.
-
Do you own decision function for maker side. For more info look at this function
from core import Local_session as Session from mymodule import myfunc session = Session(mode=args.mode, config=args.config) session.initApi(key=key, b64=b64, passphrase=passphrase, url=url, product_id=product_id) session.getApi().setBestPriceFunc(myfunc)
-
Do your own runner.
from core import Local_session as Session session = Session() # Run with default values session.loadSession() # loading environnement, worker and agent session.start() # Start the session thread
- Do your own worker.
from core import Local_env env = Local_env() # run with default values for e in episode: state = env.reset() for s in step: action = agent.act(state) next_state, terminal, reward = env.execute(action) agent.observe(reward, terminal) if terminal or env.stop: break if env.stop or e == episode - 1: env.logger._running = False #Close the logger thread break
- How to use networks.
-
You have to define your input to fit with columns name of your datasets, it will automaticaly it grab input from network and compare it with your dataset columns in getState function, it allow you to do complex network like this :
[ [ {"names": ["Price"], "type": "input"}, {"activation": "relu", "size": 8, "type": "dense"}, {"activation": "relu", "size": 8, "type": "dense"}, {"name": "pricenet", "type": "output"} ], [ {"names": ["Volume"], "type": "input"}, {"activation": "relu", "size": 8, "type": "dense"}, {"activation": "relu", "size": 8, "type": "dense"}, {"name": "volnet", "type": "output"} ], [ {"names": ["pricenet", "volnet"], "type": "input"}, {"activation": "relu", "size": 64, "type": "dense"}, {"activation": "relu", "size": 32, "type": "dense"}, {"activation": "relu", "size": 8, "type": "dense"}, {"name": "prediction", "type": "output"} ] ]- Simple network are handled as well without defining any input:
[ {"activation": "relu", "size": 64, "type": "dense"}, {"activation": "relu", "size": 64, "type": "dense"} ]
-
- Also TradzQAI support pre trained keras model:
- You can build settings for your model by using
py run.py -b DEEP. Your model have to be placed in the same directory as the one you use to launch it and have to be calleddeep_model.h5.
- You can build settings for your model by using
-
Relevant project
This project isn't perfect so keep this in mind.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
1.1k Jan 2, 2023
485 Jan 9, 2023
983 Dec 23, 2022
48 Dec 30, 2022
184 Dec 31, 2022
535 Nov 15, 2022
922 Dec 27, 2022
46 Oct 30, 2022
1 Nov 18, 2021
367 Jan 9, 2023
13 Nov 4, 2022
1 Jan 3, 2022
12 Nov 7, 2022
4 May 8, 2022
5 Aug 30, 2022
3 Oct 31, 2022
195 Dec 7, 2022
72 Jan 6, 2023
1.4k Dec 22, 2022