Blog

ContraxSuite 1.1.2 Release Update

This month, we celebrate the one-year anniversary of our decision to open-source ContraxSuite. In that time, we’ve released thirteen versions of ContraxSuite, expanding features and adding other improvements with each iteration.

Release Notes

The release of ContraxSuite 1.1.2 available from August 1 marks our thirteenth open-source release. This release focused on API updates, optimizing database queries and database cleanup, and this release implements a new and improved Docker deployment for auto-scaling working groups (increasing the speed of document processing). We also improved document field detection.

Changelog

  • LexNLP now has a reduced run time for the get_sentences method
  • We’ve added the ability for APIs to export any objects list, e.g. a document list in a .csv or .xlsx file
  • Implemented role-based complex security requirements
  • Implemented the transfer of training data for document field values
  • Optimized most API SQL queries
  • Implemented extraction of table content from documents
  • Added Review Status Group model to group document and project statuses
  • Authentication is now allowed via query parameter in GET request
  • Overall Celery stability has been improved: document loading, field value detection, and other asynchronous tasks
  • Improved document field detection and models in two major ways:
    • Added a description field to the field model, allowing for storage of more human-understandable info on what each field means;
    • Added support for “object of definition” field logic. If definition terms are entered into a field detector, then it will first check if one of them is defined in a sentence and only after this check will regexps be applied
  • Added common Nginx HTTP basic authentication behind Kibana, Flower, and Jupyter. Django has its own authentication
  • Implemented support for separate Docker cluster architectures for bigger deployments for auto-scaling Celery worker groups, and smaller deployments with a single server used for all components
  • Implemented total_cleanup.sh script which allows fast deletion of all documents, field values, tasks and other data entered in the system but keeps important configurations such as field definitions, field detectors, users, and user roles
  • Implemented Dirty Field Retraining Process

The full release notes and changelog for version 1.1.2, and for all prior versions of ContraxSuite, are available at the ContraxSuite GitHub page.

To get started with ContraxSuite by LexPredict, visit our website, or drop us a line at contact@lexpredict.com.

Comments are closed, but trackbacks and pingbacks are open.