Enhance model repository

[samuelrince]

Description

The following improvements can be considered :

• Automatically synchronize local file with remote (right now a GitHub URL)
• Have the model repository follow its own release cycle
• Handle alias instead of duplicating models
• Support dynamic fields based on the model type

Solution

A separate repository that builds the file that is then injected in the client build?

Considerations

Any input is welcome.

Additional context

See more context here genai-impact/ecologits.js#4

[omkar-foss]

Hi @samuelrince, just wanted to highlight one thing I observed - the ModelRepository.from_csv() is loading the CSV file from disk each time it's called (code here), which is slowing down it's performance due to the increased disk I/O.

We can probably consider making the ModelRepository a singleton class, so ModelRepository.from_csv() will load the CSV from disk only the first time and return the same every time it's called. This will also increase it's performance due to much lesser disk I/O.

P.S.> I came across Ecologits recently and I think it's an amazing project and initiative, kudos to you and all other contributors for the great work! 💚

[samuelrince]

Hello @omkar-foss thanks for the feedback! It will be one of our priority in the following days or weeks. 🙏

[omkar-foss]

That's great to know @samuelrince! Let me know if I can help with any of the action items, would love to contribute.