Pydantic model for config

[peterdudfield]

Detailed Description

Our configuration is getting a bit more complicated. It might worth having a pydantic model, with some default values.

Context

• try to keep the code neat

Possible Implementation

• Change here to a pydantic model
• Load from yaml file
• Could be a list on model configruations
• different yaml file for "day_ahead"

[rudra2112]

Hi @peterdudfield, I would like to pick this issue. Can you elaborate on what needs to be done, so that I can move in the right direction and get started.

[peterdudfield]

Hi @rudra2112 thanks for getting in touch

We currently have big dictionary that reapeats it self sometimes - here

I think we need to

1. Create a pyandatic model. Perhsps the fields can be the same as the dictionary keys
2. write yaml file with these variables in it
3. load the yaml file in the main app.
4. Have a different yaml file for the day ahead configuration https://github.com/openclimatefix/pvnet_app/blob/main/pvnet_app/app.py#L161

have you worked with pydantic before?

[rudra2112]

@peterdudfield, Thanks for the reply, Will start working on it. And Yeah I have worked a bit with Pydantic Models before.

[rudra2112]

Hey @peterdudfield, Should I replace with model_dict with a Pydantic Model with the fields pvnet_v2, pvnet_v2_sat0min_v9_batches, etc. and all its occurrences or just the the model config with pydantic model? And how should I structure the additions? the yamls, and models.py

[peterdudfield]

Yea I would do something like that. Perhaps models.py could contain a pydantic model called Model and it could have the following fields

• name
• pvnet_name
• pvnet_version
• summation_name
• summation_version
• use_adjuster
• save_gsp_sum
• save_gsp_to_forecast_value_last_seven_days

you might need a second pydnatic model, call Models which is a list of model. You can then easily save the dict to yaml file and can be loaded easily.

Does this make sense?