Robert EscrivaandAI 1cbcaa2147 [CHORE] preload default ONNX model with caching in CI workflows (#7201) ...

## Description of changes

Add a reusable GitHub Action and Python script to cache and preload
the all-MiniLM-L6-v2 ONNX model before test jobs run. This avoids
redundant downloads across CI runs by leveraging actions/cache with a
SHA256-pinned cache key.

- Add .github/actions/preload-default-onnx-model/action.yaml that
  caches ~/.cache/chroma/onnx_models/all-MiniLM-L6-v2 and runs the
  preload script
- Add bin/ci/preload_default_onnx_model.py with retry logic, SHA256
  verification, and safe tar extraction
- Insert the preload step in _python-tests.yml before every test job
  that depends on the default embedding model
- Insert the preload step in nightly-tests.yml before cluster tests

## Test plan

Visual inspection + CI

## Migration plan

N/A

## Observability plan

N/A

## Documentation Changes

N/A

Co-authored-by: AI

2026-06-03 14:59:39 -07:00

..

fixtures

[ENH](ci): Add GHA runtime report (#7158)

2026-05-29 10:55:43 -07:00

determine-tests-to-run.sh

[BLD] Better triggers for CI (#6603)

2026-03-18 12:42:23 -07:00

gha-runtime-prometheus.sh

[ENH](ci): add script to convert GHA runtime reports to Prometheus metrics (#7168)

2026-05-29 16:07:51 -07:00

gha-runtime-report.sh

[ENH](ci): Add GHA runtime report (#7158)

2026-05-29 10:55:43 -07:00

merge_python_test_inventory.py

[CHORE](ci): Add pytest test inventory (#7192)

2026-06-02 17:01:10 -07:00

preload_default_onnx_model.py

[CHORE] preload default ONNX model with caching in CI workflows (#7201)

2026-06-03 14:59:39 -07:00

pytest-shard.py

[ENH](ci): shard pytest and cache builds (#7198)

2026-06-03 18:47:44 +00:00

test-gha-runtime-report.sh

[ENH](ci): Add GHA runtime report (#7158)

2026-05-29 10:55:43 -07:00