TODO

[defnlnotme]

• Reduce number of packages by merging most common ones
• Improve Collections abstraction to get better static typing
• ensure the bot works with dynamic universes (universes where assets can be added or removed). Currently it should but not 100% tested.
• Rebase Watcher.jl on Rocket.jl (best to be done after ccxt julia support)
• Refactor: Instruments.Abstract/Asset to Instruments.Abstract/Instrument
• Remove vendored Zarr.jl with upstream and intead use type piracy if it seems hard to upstream current patches
• The Cli package is currently non functional, a brainstorm should be done about what features to add to the cli tool.
• The backtest might benefit from a more interactive and graphically pleasing UI by using Term.jl which we already use to show progress bars in the Scrapers package
• Similar to the above task, leverage Term.jl to prettify core data types used by Planar.
• Add support for hedged mode, and cross margin modes (this required the Collections and dynamic universes improvements).
• Abstract the exchange connector such that the bot can seamlessly work with the hummingbot gateway
• The Plotting package needs more serious work
• Do a exhaustive analysis of type instabilities (better to be done after removal of python bindings)
• rename functions like fill! and stub! (for asset instances, asset collections and strategies) to be more descriptive
• finish working on third party apis and implement watchers/scrapers (DBNomics, alpha_vantage, newsdata.io)
• finish working on blockchain features (active addresses, total value locked, stablecoin supply, supply ratio, large movements, holders in profit)
• Decide if we want to remove the Plotting module altogether or improve it
• Decide if we want a Cli package and if yes what features (currently it is an outdated OHLCV downloader)
• Gut most dependencies (make the codebase as standalone as possible, ideally without any dependency outside the repo, should be done after removal of python)