SuccessChanges

Summary

  1. Add support for loading entries directly into the usercache (commit: 4aa521caa2bab3c4a3802b210c6a0fa5f7e25d68) (details)
Commit 4aa521caa2bab3c4a3802b210c6a0fa5f7e25d68 by shankari
Add support for loading entries directly into the usercache
If the intake pipeline is not running on the server
https://github.com/e-mission/e-mission-docs/blob/master/docs/e-mission-server/deploying_your_own_server_to_production.md#the-analysis-pipeline
then data will stay in the usercache forever.
So it won't be passed through the formatter, and will not have
`data.fmt_time` and `metadata.write_fmt_time` filled in.
If we download this data (e.g. through the UI option of Profile ->
Download JSON dump), then it won't have `data.fmt_time` and
`metadata.write_fmt_time` filled in.
If we load this data directly into the timeseries, future pipeline steps
will break since they expect `data.fmt_time` entries.
Let us instead load unprocessed entries into the usercache. This will
also allow us to test the entire pipeline, including the `move to
long-term` stage and the associated formatting, as part of the load
process.
(commit: 4aa521caa2bab3c4a3802b210c6a0fa5f7e25d68)
The file was modifiedbin/debug/load_timeline_for_day_and_user.py (diff)