Skip to content

fix(models): derive PDP batch schema configs from institution schemas#247

Open
vishpillai123 wants to merge 3 commits into
developfrom
feat/pdp-inference-default-schema-configs
Open

fix(models): derive PDP batch schema configs from institution schemas#247
vishpillai123 wants to merge 3 commits into
developfrom
feat/pdp-inference-default-schema-configs

Conversation

@vishpillai123

Copy link
Copy Markdown
Collaborator

Summary

  • When model.schema_configs is null, PDP inference derives a default batch composition rule from inst.schemas (required one-of-each for input types like COURSE and STUDENT).
  • Explicit model-level schema_configs still take precedence when present.
  • Returns a clear 422 instead of a 500 when neither source can produce batch rules.

Fixes the Jackson State class of failures after create-model was simplified to name-only (#238): models created without schema_configs can run inference as long as the institution has standard input schemas configured.

Test plan

  • pytest src/webapp/routers/models_test.py
  • Deploy to dev and run PDP inference for a model with null schema_configs (e.g. Jackson State)
  • Confirm inference still works for models with explicit schema_configs set

Made with Cursor

Vishakh Pillai and others added 3 commits June 12, 2026 16:00
When model.schema_configs is null, PDP inference now builds a default
required COURSE+STUDENT (etc.) batch rule from inst.schemas instead of
500ing. Explicit model configs still take precedence.

Co-authored-by: Cursor <cursoragent@cursor.com>
Co-authored-by: Cursor <cursoragent@cursor.com>
When multiple dev bundle jobs match a PDP or legacy inference pipeline
substring, resolve to [dev dev_cloudrun_sa] if present; otherwise use
the first sorted match instead of failing the inference request.

Co-authored-by: Cursor <cursoragent@cursor.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant