Update SDK Tag Filter Models to Support Multi-Type Values#11
Conversation
|
|
We should also consider adding test cases to validate the new type specifics. |
Good point. In this case, the change is limited to the generated model classes. |
This change aligns the SDK models with the actual behavior of the Infrastructure Analyze REST APIs. The Analyze endpoints already support tag filter values as string, number, and boolean, but the SDK previously enforced a more restrictive type, which could result in validation errors for otherwise valid requests. The SDK has now been updated to reflect the API contract correctly, ensuring consistency between the SDK and the documented REST behavior. These changes correspond to the Infrastructure Analyze endpoints documented in IBM API Hub: I have also updated the title now. |
|
LGTM |
Previously, the value field generated by the SDK was defined as a generic type, but in practice this led to Pydantic validation failures when users provided numeric or boolean inputs.
With the introduction of tag_filter_all_of_value, the field is now explicitly designed to accept string, number, and boolean values, eliminating those validation errors and making input handling more flexible and predictable.
This change aligns with the future OpenAPI specification, and the same behavior will be automatically generated from upcoming openapi.yaml versions.
Manual Changes
Version bumps across:
Improved sanity_test.sh to support easier and more reliable local sanity testing