Stock Trainer

This repo contains some information about stock trainer

WIP: Architecture Diagram

https://www.figma.com/file/hYHqMQO2HRvDLysBOAXN5H/Stuff?node-id=0%3A1

Work Pipeline:

• Kite API Bridge - Makes API Calls to Broker (Kite)
• Historical Data DB (Design Pending) - DB to store historical data, OHLC with timestamp, volume, OI
• Data Manager - Provides historical stock data from Broker
• Data Request MQ
• Data Notify MQ
• Backtester - The meat of the machine, parses, executes strategies on historical data and generates results
  • Can we use ready-made backtester? https://pypi.org/project/Backtesting/
  • Whats the feasibility?
• Strategy Code - Code executed by the backtester on the historical data to generate trades
  • Use Python scripts for backtesting?
  • Use Java .class with Serialization/Reflection?
  • Write Custom parser / interpreter?
• Strategy code persistence - How to store strategy code?
• Backtester results (format pending) - Data generated from executing strategy
• Strategy DB (Design Pending) - Stores data like strategies and execution results.
• Backtester MQ
• Stuff API - API to add strategy, launch backtest, fetch results etc.