Feat/5058/add smoke tests for api happy path#3
Conversation
aryan-p03
left a comment
aryan-p03
left a comment
There was a problem hiding this comment.
Tests are working locally which is good. Although I have some questions in general:
- Would it be easier to have something like our Go component tests where we can do steps like "When I {{HTTP_METHOD}} {{HTTP_PATH}}"?
- These tests include files but only the metadata is registered. If we were to go and download the file in another step then that would fail as the file does not exist. Something to think about if we were to have other test scenarios where we wanted to test file access based on user permissions.
| context.response = requests.get( | ||
| f"{context.dataset_api_url}/datasets/{dataset_id}/editions/{edition_id}/versions/{version}/dimensions/{dimension}/options", | ||
| headers=context.headers, | ||
| ) |
There was a problem hiding this comment.
Both these function are called request_version_dimensions?
| @then('the response should contain the version with ID "{version}"') | ||
| def response_should_contain_version(context, version): | ||
| data = context.response.json() | ||
| assert data["version"] == int(version) | ||
|
|
||
|
|
||
| @then('the response should contain the version with state "{state}"') | ||
| def response_should_contain_version(context, state): | ||
| data = context.response.json() | ||
| assert data["state"] == state |
There was a problem hiding this comment.
Both these functions are called response_should_contain_version?
| Given the dataset API is available | ||
| When I request the options for dataset ID "cpih01", edition ID "time-series", version "1", dimension "geography" | ||
| Then the response status code should be "200" | ||
| And the response should contain the options for dimension "geography" |
There was a problem hiding this comment.
Are dataset types other than static in scope for the smoke test repo? (If not, then we don't need to worry about these endpoints)
There was a problem hiding this comment.
Will check in standup tomorrow.
| Then the response status code should be "200" | ||
| And the response should include the updated edition title "New edition title" | ||
|
|
||
| Scenario: Update state of a version |
There was a problem hiding this comment.
I think the scenario line has an extra indent
| ) | ||
| for i in range(num_versions): | ||
| _post_version(context, context.edition_id, i) | ||
| _post_file_metadata(context, i) |
There was a problem hiding this comment.
The random edition_id is defined at the start here but then passed into _post_version which will again ramdomise the edition_id. So does this mean each new version being posted is going to a different edition path?
So this isn't creating multiple versions within a single edition but rather creating multiple editions each with one version?
There was a problem hiding this comment.
Might be worth putting something in the README about how to debug tests
|
|
||
|
|
||
| def _post_file_metadata(context, version): | ||
| path = f"{context.dataset_id}/{context.edition_id}/{version}/{context.filename}" |
There was a problem hiding this comment.
In reality the file path is {{UUID}}/filename.extension which would be more accurate to use
2 participants
What
Added simple smoke tests for all Dataset API endpoints.
https://officefornationalstatistics.atlassian.net/browse/DIS-5058
Also added a
behave_debug.pyfile, which can be used to debug behaviour tests.How to review
Run
behaveand check that all tests pass. You'll need to add thedataset-editions-versions:deletepermission to the list of permissions for theadministratorrole indp-permissions-api/import-script/roles.json, if this hasn't been fixed yet.I've kept the assertions very simple for now, but these can be expanded upon over time, if necessary.
Who can review
Anyone