CI Failures due to Upstream OpenAPI Spec Changes

[rwjblue-glean]

CI is failing since #38. The failures have not been causing CI to fail on subsequent Speakeasy PRs because of the patch-speakeasy-pr.yml (which pushes a custom commit; GitHub actions identifies that the commit was pushed by the same event and chooses not to run CI again -- to prevent infinite loops).

The test failures seem specifically due to changes in the upstream OpenAPI Spec, and how our tests are generated. The fix (based on reading the Speakeasy docs) is to rebuild the tests?

We have a number of customizations to .speakeasy/test.arazzo.yaml though, and I'm unsure if speakeasy configure tests --rebuild will clobber those changes (seems like it will?).

[rwjblue-glean]

See https://github.com/gleanwork/api-client-python/runs/43921116379 for the list of failed tests.

[rwjblue-glean]

Based on the speakeasy docs I did try removing these operationIds from .speakeasy/gen.lock's generatedTests section, but a subsequent speakeasy run --skip-upload-specs --skip-versioning didn't update any files other than gen.lock and workflow.lock.