udata provides a command line interface for most of the administrative tasks.
You can get the documentation related to all tasks with:
$ udata -?
And then get the documentation for subtasks:
$ udata user -?
If you have some issues, start with:
$ udata info
This will display some useful details about your local configuration.
You can create a user with:
$ udata user create
You can also give a user administrative privileges with:
$ udata user set-admin <email>
Purge data flagged as deleted¶
When users delete some data in udata,
it’s only flagged as deleted and hidden in the frontend.
This allows the administrative team to undelete data in case of error.
To remove the data flagged as deleted once and for all, you need to purge them by
either launching the appropriate jobs or by executing the
$ udata purge -> Purging datasets -> Purging reuses -> Purging organizations
Sometimes you need to purge only a given type of data. You can use the appropriate flags to do so:
# purge only datasets $ udata purge --datasets -> Purging datasets # purge only reuses $ udata purge --reuses -> Purging reuses # purge only organizations $ udata purge --organizations -> Purging organizations
Warning: these operations are permanents and irreversibles
Note: Users can’t be fully purged because of the content they submitted which can’t be orphaned. This is why they are only anonymised.
Jobs are adminstrative tasks that can be run asynchronously on a worker or synchronously through the shell.
You can list available jobs with:
$ udata job list log-test purge-organizations purge-datasets bump-metrics purge-reuses error-test harvest send-frequency-reminder crawl-resources count-tags
You can launch a job with:
# Run a job synchronously $ udata job run job-name # Run a job asynchronously (needs workers) $ udata job run -d job-name
Some jobs require arguments and keywords arguments. You can pass them as arguments too:
$ udata job run job-name arg1 arg2 key1=value key2=value
Note: this is a low level command. Most of the time, you won’t need it because there will be a dedicated command to perform the task you need.
You can also schedule or unschedule jobs (and list scheduled jobs):
$ udata job scheduled # No scheduled jobs $ udata job schedule "0 * * * *" count-tags ➢ Scheduled Job count-tags with the following crontab: 0 * * * * $ udata job scheduled Count tags: count-tags ↦ 0 * * * * # Same command to reschedule $ udata job schedule "1 * * * *" count-tags ➢ Scheduled Job count-tags with the following crontab: 1 * * * * $ udata job scheduled Count tags: count-tags ↦ 1 * * * * $ udata job unschedule count-tags ➢ Unscheduled Job count-tags with the following crontab: 0 * * * * $ udata job scheduled # No scheduled jobs
Because a job can be scheduled multiple times with different parameters, you need to provide the same parameters to unschedule:
$ udata job schedule my-job "0 * * * *" arg key=value ➢ Scheduled Job my-job(arg, key=value) with the following crontab: 0 * * * * $ udata job unschedule my-job ✘ No scheduled job match Job my-job $ udata job unschedule my-job arg key=value ➢ Unscheduled Job my-job(arg, key=value) with the following crontab: 0 * * * *
Sometimes, you need to reindex data (in case of model breaking changes, workers defect…).
You can use the
udata search index command to do so.
This command supports both full reindex without arguments and partial with model names as arguments:
# Reindex everything udata search index # Only reindex reuses and organizations udata search index reuses organizations
By default the command deletes the previous index in case of success or the new unfinished index in case of error but you can ask to keep indexes with the
# Reindex everything but keep the old index udata search index -k
When used from an interactive terminal the command also prompt for deletion confirmation if an index with the same name already exists. This can be bypassed with the
# Reindex everything and delete old index udata search index -f
It’s possible to do a partial reindex by providing models (both singular and plural are supported) as arguments:
# Only reindex datasets and reuses (plural form) udata search index datasets reuses # Only reindex datasets and reuses (singular form) udata search index dataset reuse
Start a worker with:
$ udata worker start
See all waiting Celery tasks across all workers:
$ udata worker status
Display waiting tasks in a Munin plugin compatible format (you can use the provided Munin plugin):
$ udata worker status --munin -q default $ udata worker status --munin-config -q default
Flush the cache with:
$ udata cache flush