ContraxSuite 1.1.4 Release Update

Client demand for document and contract analytics tools like ContraxSuite continues to rise. We continue to improve ContraxSuite, including building functionality for integration with document management systems.

Release Notes

Interest in ContraxSuite continues to grow, which is why we recently decided to shrink the interval between our usual monthly releases. We are releasing ContraxSuite version 1.1.4, the fifteenth open source release of ContraxSuite, generally available on October 1st, 2018. Release 1.1.4 focuses on several updates around document list APIs, as well as improvements in Docker deployment.

Detailed Changelog

New in Release 1.1.4:

  • Added ability to install additional Ubuntu and Python libraries into Jupyter Docker container. A user can now put debian-requirements.txt and requirements.txt files into /data/docker/volumes/contraxsuite_contraxsuite_jupyter_add_req/_data/ and the new libraries will be available in Jupyter
  • Improved performance of document list API (for new React UI):
    • Implemented caching of aggregations used in document fields in batch analysis projects
    • Disabled the return of batch analysis fields for contract analysis projects, and vice versa
    • Overall refactored field caching and document list API
  • Made several fixes and improvements in document fields calculations
  • Refactored deployment scripts for setting up a free version of ContraxSuite from public repositories
    • Refactored and simplified installation process
    • Added docker-compose file selection (single-host/master + workers) for different deployment types
    • Added Let’s Encrypt certificates generation
    • Added network interface selection required for installing to Digital Ocean VMs
    • Improved installation manual (/docker/
  • Improved docker-compose files for single-host deployments: Celery concurrency level and memory/cpu limits for ContraxSuite components are now selected based on the amount of RAM and number of CPU cores of the host machine
  • Document API and processing (for new React UI): Added support for document fields without text annotations attached. Values for such fields can be stored without text annotations. Such values will not be used for training machine learning models
  • Added new field type: long text. This field type is used in the new React UI for showing a more expanded text editor than that used for shortened str fields
  • Optimized “Detect Field Values” and “Locate” Celery tasks
  • Fixed bug with duplicated records for “Assigned to” field
  • Created Geographies field instead of Geo Entity and Geo Alias
  • Implemented functionality to run a call-back code when all sub-tasks of a main task have been finished
  • Implemented check of classifier model based on test documents set
  • Added values for Completed/Executed documents into the training data set of the classifier model
  • Fixed exceptions serialization in Celery tasks
  • Added API and UI to dump model objects and restore model objects from fixtures
  • Fixed Top Entity Usage pages pagination
  • Return calculated field errors in JSON response
  • Added ability to set Category for document fields
  • Added extraction of html tags from text, and storage of document text in database without them
  • Fixed form dropdown layout for long values on admin site
  • Disabled add/edit actions on Document Field Value list page on admin site
  • Added confidence field in DocumentField model
  • Updated LexNLP to recognize MM.DD.YYYY and DD.MM.YYYY date formats

The full release notes and changelog for version 1.1.4, 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

Comments are closed, but trackbacks and pingbacks are open.