Update Pydantic usage for v2 compatibility #185
Conversation
|
Tests are failing as there's increased type specificity (particularly with numpy self-reporting types rather than pydantic choosing the type). These changes are work for both |
All changes are internal. The main changes are in prototypes and API errors. Prototypes now are created from dictionaries of tuples of (type, value). API errors are now a list of dictionaries, giving the type, location, error message, erroneous input, and a url to pydantic's website with more about the error.
|
For now, spacy is not compliant with pydantic v2. These changes will give some deprecation warnings in tests, but overall are v1 and v2 compatible. |
isabelizimm
changed the title
| python -m pip install xgboost | ||
| python -m pip install spacy | ||
| python -m pip install torch | ||
| python -m pip install statsmodels |
There was a problem hiding this comment.
this is effectively the same as python -m pip install -e ".[dev,all_models]", but Python 3.7 will not complete installation, ending in an OSError saying there is not enough space 😩 this is likely due to some pretty intense dependency resolution.
I'll likely want to rethink this strategy, but that can come in a separate PR.
| python -m pip install spacy | ||
| python -m pip install torch | ||
| python -m pip install statsmodels | ||
| python -m pip install typing_extensions==4.7.1 |
There was a problem hiding this comment.
Pinned due to fastapi requirements. fastapi/fastapi#9808
1 participant
resolves #184