Uploaded image for project: 'Moodle'
  1. Moodle
  2. MDL-70887

Upgrade moodle-mlbackend-python and update its reference in /lib/mlbackend/python/classes/processor.php

    XMLWordPrintable

    Details

    • Testing Instructions:
      Hide

      For this you need to have moodlemlbackend 2.5.0 installed which requires tensorflow 1.15.5 installed so you need to run

      sudo -H python3 -m pip install "moodlemlbackend==2.5.0"
      

      You might need to upgrade pip if above is fails to find tensorflow 1.15.5

      pip install --upgrade pip
      

      Then you need to get some database with enough data to run models (ping @ilyatregubov where it is stored)

      Testing:

      Before the patch:

      sudo -H python3 -m pip install "moodlemlbackend==2.4.0"
      

      • Try running models (students at risk of dropping out):

        php admin/tool/analytics/cli/evaluate_model.php --modelid=1 --analysisinterval='\core\analytics\time_splitting\deciles_accum'
        

        Run above 3 times and note the accuracy

      After the patch:

      sudo -H python3 -m pip install "moodlemlbackend==2.5.0"
      

      • Try running models (students at risk of dropping out):

        php admin/tool/analytics/cli/evaluate_model.php --modelid=1 --analysisinterval='\core\analytics\time_splitting\deciles_accum'
        

        Run above 3 times and note the accuracy

      So accuracy after patch should be similar to what it was before patch.

      Show
      For this you need to have moodlemlbackend 2.5.0 installed which requires tensorflow 1.15.5 installed so you need to run sudo -H python3 -m pip install "moodlemlbackend==2.5.0" You might need to upgrade pip if above is fails to find tensorflow 1.15.5 pip install --upgrade pip Then you need to get some database with enough data to run models (ping @ilyatregubov where it is stored) Add python to system path at http://MOODLE/admin/settings.php?section=systempaths Set default prediction processor to Python machine learning backend at http://MOODLE/admin/settings.php?section=analyticssettings Testing: Before the patch: sudo -H python3 -m pip install "moodlemlbackend==2.4.0" Try running models (students at risk of dropping out): php admin/tool/analytics/cli/evaluate_model.php --modelid= 1 --analysisinterval= '\core\analytics\time_splitting\deciles_accum' Run above 3 times and note the accuracy After the patch: sudo -H python3 -m pip install "moodlemlbackend==2.5.0" Try running models (students at risk of dropping out): php admin/tool/analytics/cli/evaluate_model.php --modelid= 1 --analysisinterval= '\core\analytics\time_splitting\deciles_accum' Run above 3 times and note the accuracy So accuracy after patch should be similar to what it was before patch.
    • Affected Branches:
      MOODLE_400_STABLE
    • Pull 3.11 Branch:
      MDL-70887-311
    • Pull Master Branch:
      MDL-70887-master
    • Sprint:
      Moppies Kanban, Moppies Kanban, Moppies Kanban

      Description

      https://github.com/moodlehq/moodle-mlbackend-python/ should be upgraded (to avoid security errors). Once a new version will be released, references in Moodle (such as /lib/mlbackend/python/classes/processor.php ) should be updated to use the latest version.
       

        Attachments

          Issue Links

            Activity

              People

              Assignee:
              ilyatregubov Ilya Tregubov
              Reporter:
              sarjona Sara Arjona (@sarjona)
              Peer reviewer:
              Carlos Escobedo Carlos Escobedo
              Participants:
              Component watchers:
              Elizabeth Dalton, Amaia Anabitarte, Carlos Escobedo, Ferran Recio, Ilya Tregubov, Sara Arjona (@sarjona), Amaia Anabitarte, Carlos Escobedo, Ferran Recio, Ilya Tregubov, Sara Arjona (@sarjona)
              Votes:
              0 Vote for this issue
              Watchers:
              6 Start watching this issue

                Dates

                Created:
                Updated:

                  Time Tracking

                  Estimated:
                  Original Estimate - 0 minutes
                  0m
                  Remaining:
                  Remaining Estimate - 0 minutes
                  0m
                  Logged:
                  Time Spent - 6 hours, 31 minutes
                  6h 31m