IBM Watson™ Discovery Service Ideas

We've moved...

You'll be redirected shortly, we've moved to our new idea portal: https://ibm-watson.ideas.aha.io

Delete button for individual docs/results

Currently there is no simple or interactive way to delete documents within the WDS toolkit making it difficult for non-developers to manage the WDS corpus without having to use curl/sdk/api. While it is easy to upload individual documents via drag and drop, it should be just as easy for a user to delete a document from the corpus as well. This could perhaps be implemented as a 'bin' button that is appended to each result in the query tab and clicking on it would initiate an api call to delete the document from the corpus. 

One side feature to consider would also be an update button that allows users to update particular document metadata fields as well.

  • KENT TRAN
  • Feb 12 2018
  • Future Consideration
Why is it useful?
Who would benefit from this IDEA? Anyone who uses the toolkit to manage documents
How should it work?
Idea Priority
Priority Justification
Customer Name
Submitting Organization
Submitter Tags
  • Attach files
  • Admin
    JOHN PECORARI commented
    29 Jan 13:19

    I have quick question to those interested.  As part of implementing support for this feature, would it be acceptable if all collection assets such as training data, synonyms, etc.were all deleted when a user chose to "delete all documents" from a collection?  Note that deleting an individual document would not delete all collection assets.

  • Admin
    Phil Anderson commented
    29 Jan 13:44

    i wouldn't expect synonyms to be deleted, but i might expect training data to be deleted

  • Admin
    JOHN PECORARI commented
    29 Jan 14:07

    while it might not be expected, would it present a real problem if it was deleted?  basically, to help reduce scope for this feature, the simplest thing to do for a "delete all documents" function is for tooling delete the collection behind the scenes and then create a new one with the same name and collection id.  So while it would delete all docs and retain the original collection attributes, it would delete all collection data vs just the documents.  For the main scenario where this feature is likely helpful, the question is would that be acceptable behavior?

  • Admin
    Phil Anderson commented
    29 Jan 14:16

    Thinking very briefly about it I think that would be acceptable - i assume a use case where a user wanted to maintain the training they would upload the same docs to overwrite, since the doc ids matter

  • KENT TRAN commented
    30 Jan 23:59

    @John Pecorari @Phil Anderson

    WRT a 'Delete All' button part of the problem we had was that there would be a lot of tweaking we may have to do with a particular config just to get the desired format/tags (e.g. making sure the docs are splitting at the right point). 

    So we might throw one or two docs into a collection, test it out and find that we need to make a few tweaks to the config. At this point if we wanted to test the updated config we would have two options here:

    - Delete the collection and create a new one, we would then have to update all our connected apps to use this new collection id.

    - Run a script to empty the collection by looping through and deleting every doc individually. 

    Both of which can be a hassle to do particularly if you end up making a dozen or more config changes before settling on one. A 'Delete All' button would help solve this issue, several guys in my team think this would be a feature useful prior kicking off a the ingestion process.