Web-based Carrot2 Clustering Workbench

[stanislawosinski]

Implementation of the clustering workbench as part of the frontend application delivered with Carrot2 Document Clustering Server.

• Clustering language choice
• Pass query hint to the clustering algorithms
• export of parameters JSON for pasting into code
• Custom schema results: display URL-typed fields
• Export of clusters and documents as Excel/OpenOffice/CSV
• Solr data source
  • Limit the number of values we display for multi-value fields
  • Scroll a long list of Fields to cluster
  • Improve autodetection of natural text fields
  • Show search errors returned by Solr (e.g. query syntax problems)
• ElasticSearch data source
• Rewrite the About section
  • Add a note on what happens with the data the user submits to Workbench (search results are sent to our server for clustering, data is not logged or stored on our server). Display the note only for our public demo or show the URL of the DCS where the data is processed.
• Perform clustering on Ctrl+Enter
• Filtering of advanced settings
• Review attribute documentation
• Improve initial workbench rendering time
• Fix broken visualization export
• Fix flickering loading overlay on visualizations in workbench
• Option to export only non-default parameters from the UI
• Mention workbench as part of "Getting started" information in the documentation.
• Option to override the autodetected list of natural text fields (in case we were wrong and the field the user wishes to cluster is not on the list).
• Code splitting