Loading...

XML

Word

Printable

Affected Branches:
MOODLE_34_STABLE
Fixed Branches:
MOODLE_34_STABLE
Epic Link:
Analytics 3.4 follow ups
Pull from Repository:
git://github.com/dmonllao/moodle.git
Pull Main Branch:
~~MDL-59630~~_master
Pull Main Diff URL:
https://github.com/dmonllao/moodle/compare/MDL-59630_master~4...MDL-59630_master
Testing Instructions:

Hide

All this issue changes are covered by unit tests.

Show
All this issue changes are covered by unit tests.

Analytics tables and files needs a cleanup process like everything in software, we need to define what and when.

What, must be cleaned:

Unexisting contexts predictions
Unexisting predictions actions (~~MDL-58835~~) both when the prediction context does not exist and when the model is deleted
Deleted models logs (we already delete all other model data)
Deleted models files (all analysable files + merged files)
Calculations belonging to contexts that do not exist anymore (analytics_indicator_calc)

What we can discuss if it should be cleaned or not:

Deleted analysable records (analytics_predict_samples, analytics_train_samples) once the analysable is not available anymore
Deleted analysable files (evaluation and prediction files) once the analysable is not available anymore

What I would vote to keep:

blocks

MDL-59988 Discard datasets that are not ready for training / prediction and append them later to other datasets when ready

MDL-59694 Limit the amount of analysables that are processed during one train() and predict() execution

Discovered while testing

MDL-57791 Implement an analytics engine in Moodle

is blocked by

MDL-58835 Ability to hide predictions (flag them as not useful or acknowledged / fixed)

MDL-59067 Calculate indicators way before the target is available

Participants:: Andrew Lyons, CiBoT, Damyon Wiese, David Monllaó, Eloy Lafuente (stronk7), Jun Pataleta