AlphaGriffin BitTensor


Tensorflow vs. Bitcoin. Currently in development by 2 guys. if you would like to help out with the project please donate Litecoin to the following address. Litecoin is our preferred transactional coin for its low fees and segwit2x.

LTC: 33jckQAojmGrK2ZG9ZnA6z9hXAjK1mfAhR

Getting Started

These instructions will get you a copy of the project up and running on your local machine for development and testing purposes. See deployment for notes on how to deploy the project on a live system.

Prerequisites

Copy the dummy_codes to access_codes and add in your API keys. Copy the dummy_secret to client_secret and add in your API keys.

cd config
cp dummy_codes.yaml access_codes.yaml
cp dummy_secret.json client_secret.json

Edit the new access_codes and client_secret file as necessary.

Deployment

This Project won't get far without data. First edit the access_codes.yaml. And add in all of the exchanges you plan to be working with. !! Important Note: Bittrex full set takes up about .75GB of data.

Important Note: Not all exchange provide historical data to commoners.

# In access_codes.yaml
exchanges: 'poloniex,bittrex'
# then run the datagrabber
python bittensor_data.py

Presentation

This is a Plotly Dash Site for data viewing.

python site.py

Tensorflow

# train model
python bittensor_train.py
# backtest model
python bittensor_test.py

TODO

most all tools are currently mid broken from the datagrabber upgrade. Everything will need to be reevaluated one peice at a time. Parts needing to be looked at:

• Paper Trader
  • Seperate Back Testing tools
  • Make it go fully online and start buying and selling to test transaction times
  • Make it work fully offline replicating what you would see if you were online
  • Keep really good track of trade history
  • Make the Paper Trader Growup to be a real trader some day.
  • PaperTrader Should otherwise function in Live/Fake Mode + strategy.
  • This is going to need its own file saving system.
  • Should be able to build this using the plotlyDash interface.
• BackTester and Paper Trader need to be 2 seperate things.
• Lots of TA tools left to build. The instructions are in the TALib file, just needs doing.
• Website:
  • Frontend for Paper Trader, Backtester
  • Navigation between Dash, Tensorboard, PaperTrader, Backtester
  • Better Graph And TA Chart building
  • Add in Coin Detail with Graph and TA charts like logo, company info, blockchain details
  • Pull the orderbook in from the web when a pair is selected
  • Full Resituation of the Main Presentation Page. Make it presentable.
• TensorFlow:
  • *better Data input Schema
  • reward mechinizm
  • Tensorboard
  • Build Price Prediction Estimator
  • Use Prediction in the Q-Learner
• DataSmith:
  • Needs need DataGrabber tools built into the namespace
  • Needs to save all the files to Google Drive.
  • Needs Other Data to grab. Twitter, Steam, Facebook, News, reddit, bitcointalk Company information, linkedin
• Blockchain Interface:
  • UNSTARTED
• Local Wallets:
  • UNSTARTED

Update Log

V 0.0.1: bittensor_data.py is now working as expected with the datasmith and the TA lib. This can be explored in the TA_Samples and DataSmith Notebooks.