- Split model::predict in parts
- JS promises updated according to eslint-plugin-promise
- New API methods replacing direct DB queries
- Reduce insights nav link display cost
- Increase time limit as well as memory for big processes
- Move prediction action event to core
- Dataset write locking and others
- Refine last time range end time
- Removed dodgy splitting method id to int
- Replace admin_setting_predictor output_html overwrite for write_setting overwrite
- New APIs for access control
- Discard invalid samples also during prediction
The add/roles thing was only converted to bootstrap 2/4 markup and the yui left in place.
The modal for adding users to a course was unsavable so I rewrote it with an mform in a popup, still calling
the same (barely) modified ajax script.
The webservice for searching cohorts was taken from admin/tool/lp and moved into /cohort. I added a generic "cohort"
mform element at the same time.
The webservice for searching for users was taken from the original ajax script.
All these functions were used only by deleted upgrade steps
so it's safe to proceed with straight deletion, considering
them internal. Deletion has been documented in corresponding
upgrade.txt files:
- upgrade_mimetypes()
- upgrade_fix_missing_root_folders_draft()
- upgrade_minmaxgrade() and upgrade_minmaxgradestepignored setting
- upgrade_course_tags()
- atto_equation_update_librarygroup4_setting()
- mod_lti_upgrade_custom_separator()
These have been kept because continue being used by restore:
- upgrade_group_members_only()
- upgrade_extra_credit_weightoverride()
- upgrade_calculated_grade_items()
This just deletes all the upgrade steps previous to 3.0.0. Some
small adjustments, like adding missing MOODLE_INTERNAL or tweaking
globals can also be applied when needed.
Also includes an upgrade step to prevent upgrading from any
version < 2015111600 (v3.0.0) as anti-cheating measure.
Next commit will get rid of/deprecate all the upgradelib functions
not used anymore in codebase.
Find any events records for assign group overrides having null priority
and sets their priority to the sortorder value from the corresponding
assign_overrides table entry.
This change removes the complex and partial loading of role definition data for users.
It is replaced by keeping one system-wide definition for each role in MUC.
This is cheaper to regenerate as it avoids complex database queries (repeated LEFT JOIN to context).
Memory usage is kept low thanks to array CoW for each user access data.
There were 4 instances of the calendar refresh events task being
queued, which is redundant as we now have the "one true task" to
fix the events table, and it needs to be run for everybody.
