Could we make "--nodatabases" the default for ARAX pytest suite?

[saramsey]

It would be great if the database manager didn't try to download 200 GiB of databases each time I run the pytest suite, and (invariably) forget to include the --nodatabases flag to the conftest.py module. Is there any reason why we can't make --nodatabases the default? Under what circumstances do we want the database files downloaded as a side effect of running the pytest suite?

[saramsey]

@bazarkua and @edeutsch and @dkoslicki do you have an opinion on this?

[dkoslicki]

Oof, yeah, seems to make sense to make --nodatabases the default. We will pretty quickly see locally if we are missing databases anyways. I would want to verify that the GitHub CI/CD would be able to handle this change, or if it would need to be tweaked slightly to download them

[bazarkua]

I concur with @saramsey we don't need to download 200GiB just to run the pytest suite.

Production deployments have their own independent database download in the Flask __main__.py startup. This conftest change does not touch that.

ITRB CI/CD does not invoke pytest at all. The Jenkinsfile, Dockerfile, and deploy scripts have zero references to pytest. Databases are downloaded by the init container's download_database.sh.

@dkoslicki Good question I looked at the GitHub worflow.
The GitHub Actions workflow does run pytest, and it does not pass --nodatabases But it does not need to, and the workflow already downloads databases before pytest runs, in a separate step. In pytest.yaml we download databases in separate step

[bazarkua]

Changed confest.py to skip database download.
Default pytest now:

~/Work/RTX issue-2663* ❯ pytest -v code/ARAX/test/test_ARAX_synonymizer.py                                                                 venv312 06:13:51 PM
Skipping database check (pass --withdatabases to download/update databases)
KP info cache is up to date.
===================================================================== test session starts =====================================================================
platform darwin -- Python 3.12.12, pytest-8.1.1, pluggy-1.6.0 -- /Users/bazarkua/Work/venv312/bin/python3.12
cachedir: .pytest_cache
rootdir: /Users/bazarkua/Work/RTX
plugins: anyio-4.12.0, cov-4.1.0
collected 24 items                                                                                                                                            

code/ARAX/test/test_ARAX_synonymizer.py::test_example_6b PASSED                                                                                         [  4%]
code/ARAX/test/test_ARAX_synonymizer.py::test_example_9 PASSED                                                                                          [  8%]
code/ARAX/test/test_ARAX_synonymizer.py::test_example_10 PASSED                                                                                         [ 12%]
code/ARAX/test/test_ARAX_synonymizer.py::test_example_11 PASSED                                                                                         [ 16%]
code/ARAX/test/test_ARAX_synonymizer.py::test_example_12 PASSED                                                                                         [ 20%]
code/ARAX/test/test_ARAX_synonymizer.py::test_get_canonical_curies_simple PASSED                                                                        [ 25%]
code/ARAX/test/test_ARAX_synonymizer.py::test_get_canonical_curies_single_curie PASSED                                                                  [ 29%]
code/ARAX/test/test_ARAX_synonymizer.py::test_get_canonical_curies_unrecognized PASSED                                                                  [ 33%]
code/ARAX/test/test_ARAX_synonymizer.py::test_get_canonical_curies_by_names PASSED                                                                      [ 37%]
code/ARAX/test/test_ARAX_synonymizer.py::test_get_canonical_curies_single_name PASSED                                                                   [ 41%]
code/ARAX/test/test_ARAX_synonymizer.py::test_get_canonical_curies_by_names_and_curies PASSED                                                           [ 45%]
code/ARAX/test/test_ARAX_synonymizer.py::test_get_canonical_curies_return_all_categories PASSED                                                         [ 50%]
code/ARAX/test/test_ARAX_synonymizer.py::test_get_equivalent_nodes PASSED                                                                               [ 54%]
code/ARAX/test/test_ARAX_synonymizer.py::test_get_equivalent_nodes_by_name PASSED                                                                       [ 58%]
code/ARAX/test/test_ARAX_synonymizer.py::test_get_equivalent_nodes_by_curies_and_names PASSED                                                           [ 62%]
code/ARAX/test/test_ARAX_synonymizer.py::test_get_curie_names PASSED                                                                                    [ 66%]
code/ARAX/test/test_ARAX_synonymizer.py::test_get_preferred_names PASSED                                                                                [ 70%]
code/ARAX/test/test_ARAX_synonymizer.py::test_get_normalizer_results PASSED                                                                             [ 75%]
code/ARAX/test/test_ARAX_synonymizer.py::test_improper_curie_prefix_capitalization PASSED                                                               [ 79%]
code/ARAX/test/test_ARAX_synonymizer.py::test_approximate_name_based_matching PASSED                                                                    [ 83%]
code/ARAX/test/test_ARAX_synonymizer.py::test_entity_controller_input_no_format PASSED                                                                  [ 87%]
code/ARAX/test/test_ARAX_synonymizer.py::test_entity_controller_input_minimal_format PASSED                                                             [ 91%]
code/ARAX/test/test_ARAX_synonymizer.py::test_cluster_graphs PASSED                                                                                     [ 95%]
code/ARAX/test/test_ARAX_synonymizer.py::test_truncate_cluster PASSED                                                                                   [100%]

===================================================================== 24 passed in 14.57s =====================================================================

You can pass --withdatabases as optional flag to download 200GiB of databases