search

thesadru / genshin.py

API wrapper for HoYoLAB/Miyoushe API built on asyncio and pydantic.
https://thesadru.github.io/genshin.py
MIT License
400 stars 72 forks source link

Allow raw data to be returned #27

Open nano-shino opened 2 years ago

nano-shino commented 2 years ago

This has been bugging me for a while but can you include the raw json data in the returned pydantic model? Sometimes the assumption that you used to parse the json isn't correct, and I would have to hack into the library to get the original json to parse it myself.

Here's an example: The real-time notes API returns 'transformer': {'obtained': True, 'recovery_time': {'Day': 0, 'Hour': 5, 'Minute': 0, 'Second': 0, 'reached': True}, 'wiki': ''}. Mihoyo only returns the biggest time unit. In this case it rounds down to 5 hours so even if its 5 hours and 12 minutes it still returns 5 hours. The remaining_transformer_recovery_time as it's currently implemented gives the wrong impression that it will recover at this exact time. Meanwhile if I have the original json, I can display something like "parametric transformer will recover in 5 hours...", "parametric transformer will recover in 21 minutes..."

This is just one example. I have come across a few cases where the raw json is greatly appreciated, like the fields that you haven't yet added (e.g. data["transformer"]["obtained"])

thesadru commented 2 years ago

I'm all for making patching easier but I'm not sure if this is truly the best approach. Storing raw values should be possible with a root validator in the base class so I can do that but I'd rather allow the user to make their own models if they so wish.

For this specific case I think I should store the time as a timedelta instead of a datetime which should allow for easier usage.

The Transformer will return None if not obtained as opposed to giving 0 remaining time so that's not really a problem.

There is a TODO list just in case.

thesadru commented 2 years ago

This will be fixed by removing the pydantic dependency and switching to a custom model library. Development is slow but it's being worked on.

nano-shino commented 2 years ago

Thanks! That's great to hear. I like the idea of pretty Python models with pydantic but it can be restricting sometimes. Personally I prefer genshin.py to remain a barebone library or at least expose the basic hoyolab APIs so I can build on other things on it. The custom model library looks very promising. Can't wait to see it.