Page MenuHomePhabricator

LinguaLibreBot : [TBD] Find a way to catch missed recordings over extended "out of service" periods.
Open, MediumPublicFeature

Description

With the current crontab (https://github.com/lingua-libre/operations/issues/2#issuecomment-773443145), the bot is run everyday at 2 AM. Yet, the bot only queries the recordings dating up to two days from the current day.

0 2 * * * /usr/bin/python3.5 /home/www/Lingua-Libre-Bot/llbot.py simple --startdate "`date -d "-2 days" --rfc-3339=date`" > /var/log/llbot.log 2>&1

Is the bot able to handle all the recordings from a bigger period of time ? Should I add a routine to actively scan for missed recordings ?

Related: T233917 , T215056 , T232299

Event Timeline

Pamputt changed the subtype of this task from "Task" to "Feature Request".Feb 11 2021, 5:16 PM

Any idea how you would approach that ? Do you want to check somehow if it was already uploaded to the page ? If you extant the period, doesn't it risk to add several times the same item to the target wikipage ?

A precision on MediaWiki side: the recent changes are available by default during 90 days (about 3 months, it is the current value on LinguaLibre, parameter $wgRCMaxAge), and there is a script to rebuild the recent changes.