Uploaded image for project: 'Text Classification '
  1. Text Classification
  2. TXTREC-87

Provide a way to disable text classification

    XMLWordPrintable

Details

    • Improvement
    • Resolution: Done
    • Neutral
    • 1.1.6
    • 1.1.2, 2.0
    • Yes
    • Yes
    • AuthX 22
    • 3

    Description

      Issue 1
      Provide a way to disable this feature. A customer might want to use other amazon connector modules. However, if AWS credentials are found in the Passwords app then text classification is enabled. For some customers this can slow down the start up since the system will pause until the classification process completes:

      2020-12-29 10:59:40,242 INFO  info.magnolia.ai.text.TextClassificationModule    : Text classification might take some time, please do not shut down your instance.
      2020-12-29 10:59:40,302 INFO  info.magnolia.ai.text.TextClassificationModule    : Number of untagged nodes: 42
      

      I tried changing the config files here:

      • /text-classification/config.yaml
        aggregateDefinition:
          fieldTypes: []
        termFilteringDefinition:
          excludedTerms: []
        
      • /pages-content-tags-integration/decorations/text-classification/config/config.yaml

      Nothing seems to disable the feature from running. You have to uninstall the module.

      Issue 2
      Also some users may not want to use AWS at all but still get a confusing error message in the log which they might think is causing other issues:

      ERROR info.magnolia.aws.foundation.AwsCredentialsProvider 26.01.2021 13:35:55 - AWS credentials are expected to be set in Password manager module.
      ERROR info.magnolia.ai.text.TextClassificationModule 26.01.2021 13:35:55 - Submission of text classification request has been failed at page path '/bdcwebsite/bdcchat/bdcchat2' with error code: 'null'.
      

      Proposed solution
      I would prefer to be able to hotfix /text-classification/config.yaml. The PR is attached to this ticket:

      # turn off the module with this property
      enabled: false
      aggregateDefinition:
        fieldTypes: [text, textField, richText, richTextField, composite, compositeField, switchable, switchableField]
      termFilteringDefinition:
        excludedTerms: []
      

      would result in the log:

      2020-12-29 13:11:11,177 INFO  info.magnolia.ai.text.TextClassificationModule    : Text classification module is disabled.
      

      Checklists

        Acceptance criteria

        Attachments

          Issue Links

            Activity

              People

                miguel.martinez Miguel Martinez
                rgange Richard Gange
                AuthorX
                Votes:
                0 Vote for this issue
                Watchers:
                4 Start watching this issue

                Dates

                  Created:
                  Updated:
                  Resolved:

                  Checklists

                    Task DoD

                    Time Tracking

                      Estimated:
                      Original Estimate - Not Specified
                      Not Specified
                      Remaining:
                      Remaining Estimate - 0d
                      0d
                      Logged:
                      Time Spent - 4.75h
                      4.75h