ckanext-report cache upgrade

[jbrown-xentity]

User Story

In order to avoid a complicate report taking too long to generate or even timing out when user is visiting the report, data.gov admins want a way to always serve cached data to report requests.

Acceptance Criteria

[ACs should be clearly demoable/verifiable whenever possible. Try specifying them using BDD.]

• GIVEN a configuration flag is set
  WHEN a report is requested from the web UI
  THEN the report is always served from the cached data.
• GIVEN a report is out-dated
  WHEN a command is issued via CLI or API call
  THEN the report is showing the latest data.

Background

#4116
#4018

Security Considerations (required)

None

Sketch

We'll plan to fork ckanext-report, make the changes, and then try to push upstream.

[FuhuXia]

After some research, all reports data are saved to DB in table data_cache. The ticket is about how to add a config so we can set desired max_age of cache instead of hard-coded 2 days. In our case, we want to set it the max_age a real long time, and fresh the reports via cli or api call.

[FuhuXia]

PR submited to upstream. We can use our forked branch for now.

After the deployment, we can set cached data not expire by giving the max age a real big value, e.g.,

ckanext-report.max_age_in_days = 365

Then use CF action to run tasks on schedule to fresh the report cache on demand:

ckan report generate [report-name]

[nickumia-reisys]

A week of successful report generation. Most recent run: https://github.com/GSA/catalog.data.gov/actions/runs/4657679668/jobs/8242511808

[FuhuXia]

We can go ahead and use default 2-day cache age. If we start to experience issues, or if upstream have the PR merged, we will revisit this ticket.