Fail to dedupe with 403 error on /api/action/package_update

[FuhuXia]

The dedupe script is not working any more.

Run dedupe GH action at https://github.com/GSA/datagov-dedupe/actions/workflows/run_dedupe.yml, or run local command

pipenv run python duplicates-identifier-api.py ca-gov --api-key some-api-key --newest

It gives error:

...
  File "/home/fxia/git/datagov/datagov-dedupe/dedupe/ckan_api.py", line 342, in update_package
    self.request("POST", "/action/package_update", json=package)
  File "/home/fxia/git/datagov/datagov-dedupe/dedupe/ckan_api.py", line 102, in request
    raise CkanApiStatusException(
dedupe.ckan_api.CkanApiStatusException: Unsuccessful status code 403

On the server side, there is 403 error.

# on catalog-admin
 [ckan.config.middleware.flask_app]  403 /api/action/package_update render time 0.407 seconds

# on catalog-proxy
NginxLog "POST /api/action/package_update HTTP/1.1" 403 268

Sketch

[Notes or a checklist reflecting our understanding of the selected approach]

[FuhuXia]

I assume we are using the user geo.gov's credential to run the dedupe script. But geo.gov has no api-token. Not sure why but if we create a new one then update the GH secret, it will be working again.

[FuhuXia]

Token github-dedupe was created for user geo.gov, used in GH secret. Dedupe script runs fine now.