prediction-api

Archived

After looking into trade execution frameworks, it has been found that there are better options than making an api. So this project is on hold indefinitly and will be pursing other options

TODO List

• Data Structures
• Container to load the data to the database
• Queries to prepare the data into one table
• Run the Dataset through the POC juypter notebook
• (optional) build out the pytorch training in the api
  • Copy all the code from the juypter notebook
  • select the data
  • test/train split
  • instantiate the model
  • train the model
  • save the model to disk
• build out serving the model in the api
  • load in model
  • collect live data
  • transform the live data
  • make the prediction
  • reverse transform the prediction
  • return the prediction
• Build out the Trade Decision endpoint
  • Collect Data
  • Make the prediction
  • Make the trade decision
  • place trades with stop loss and take profit
  • write the trade to the database
  • return the trade to the user
• Orchistration in GH Actions
• CI/CD in GH Actions
• Deploy