You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
We should establish automated testing that runs on a periodic basis (Scheduled GitHub Action runs) to test whether we have implemented all API parameters.
What solution would you like?
Write code to remotely fetch the API specification from wherever it resides. For example, OpenSearch APIs are published here: https://github.com/opensearch-project/opensearch-api-specification/tree/main/spec . It may be acceptable to fetch these file via wget in the GHA script to make it available locally, but if the effort to use a testing client (like we do in integ tests) is low, that would be preferable.
Use the WorkflowSteps enum to compare the schema/content/requestBody definitions with the required and optional inputs. The enum may need to be updated to include other information for comparison.
What alternatives have you considered?
Manually following change logs, documentation, and PRs
Do you have any additional context?
It may not be necessary to go deeply into object parsing; just the first level presence or absence of an object key in the request body is likely sufficient.
The text was updated successfully, but these errors were encountered:
Is your feature request related to a problem?
Workflow Steps are tightly coupled to the APIs they implement.
The APIs are hosted in external repositories, and there is not signal to our plugin that an API has added features.
OpenSearch APIs have OpenAPI spec files written for them: https://github.com/opensearch-project/opensearch-api-specification
I have opened an issue on ML Commons requesting they do the same: opensearch-project/ml-commons#2297
We should establish automated testing that runs on a periodic basis (Scheduled GitHub Action runs) to test whether we have implemented all API parameters.
What solution would you like?
What alternatives have you considered?
Manually following change logs, documentation, and PRs
Do you have any additional context?
It may not be necessary to go deeply into object parsing; just the first level presence or absence of an object key in the request body is likely sufficient.
The text was updated successfully, but these errors were encountered: