fix: Add missing group_id to RecipeTag and TagBase schemas

[robertdanahome]

Problem:

When updating recipe tags via the API (PATCH /api/recipes/{recipe_slug} or using bulk actions like /api/recipes/bulk-actions/tag), requests can fail with either a TypeError: __init__() missing 1 required positional argument: 'group_id' or a SQL Integrity Error (violating the unique constraint on tags.slug + tags.group_id).

Root Cause:

There is an inconsistency between the Pydantic schemas used for API validation and the underlying SQLAlchemy database model for Tags:

• The Tag database model (mealie/db/models/recipe/tag.py) correctly requires group_id during initialization (__init__).
• However, the Pydantic schemas RecipeTag (mealie/schema/recipe/recipe.py) and TagBase (mealie/schema/recipe/recipe_category.py), which are used to validate the tags array in API request payloads, do not define a group_id field.
• Consequently, when a payload containing group_id within the tag objects is sent, Pydantic strips this field during validation.
• The backend processing logic then receives tag data missing the required group_id, leading to errors when attempting to interact with the Tag database model.

Solution:

This PR adds the missing group_id: UUID4 | None = None field to the RecipeTag and CategoryBase (which TagBase inherits from) Pydantic schemas.

• mealie/schema/recipe/recipe.py (modified RecipeTag)
• mealie/schema/recipe/recipe_category.py (modified CategoryBase)

This ensures that group_id, if provided in an API request for associating tags, is correctly validated and passed through to the backend database logic, resolving the TypeError and SQL Integrity Error. The field is made optional (| None = None) to minimize potential side effects in other parts of the code that might use these schemas without providing a group ID.

However, this change broke tests. Fixing it required adding the required group_id field to the RecipeToolOut schema. This ensures group_id is present when validating Recipe objects during creation in tests, resolving Pydantic validation errors and downstream IntegrityErrors that occurred during test setup fixtures.

Testing:

Tested locally by applying these schema changes and using an external script to PATCH recipes with tags including id, name, slug, and group_id. The previous TypeError and SQL Integrity Error were resolved, and tags were associated successfully.

Ran the python test suite locally, all tests are now passing.

Fixes #5286

[robertdanahome]

Full disclosure: I diagnosed and fixed this using AI. The fix works in my environment, but it's been years since I've slung code and I don't know this project well enough to confirm whether or not there are other things that might need to be changed along with this fix.

[robertdanahome]

looking into the test failures

[robertdanahome]

My first try at this broke tests. I made the fix for that, then squashed that fix into the original fix. This is now passing the python tests locally.

[robertdanahome]

fixed the formatting issue from the last test run.

[michael-genson]

Thanks!