Squirro 3.2.7 - Release Notes

Owned by Saurabh Jain, created
Last updated: Mar 26, 2021 by Patrice Neff
3 min read

This is the 7th minor release as the follow up to the last major Squirro release 3.2.0. This release was released on Mar 17, 2021

Contents - What's in the release?

Improvements

Newsletters

  • We have improved the scheduling of email newsletters under the hood. This allows for a
    Newsletter scheduling interface now includes a help link to an online crontab editor to help with the crafting of cron syntax.

  • Links in the newsletter items now redirect the user to a Squirro dashboard instead of the original source.

  • We have fixed the image uploading capability for the newsletter template editing.

Cognitive Search

  • Breadcrumb widget now displays the community query instead of the community name in Cognitive search application.

  • The default sort order for the search results have been changed from "By Time" to "By Relevance".

Typeahead

  • We have made the content typeahead introduced in the last release smarter by leveraging the textrank algorithm for extracting top phrases from a document.

  • We have also laid down the groundwork to extract relevant tracking info from our log files. In future releases, we intend to use this to provide additional typeahead features like popular searches, recent search and many more.

  • Typeahead on the content now also works for PDF documents.

  • We have also fixed a bug around PDF Highlighting of Query Matches.

  • And of-course the default project template for Cognitive Search application comes pre-configured with all these nice features out of the box.

AI Studio

  • We now allow for multiple strategies for Sentence Splitting. This is a useful feature that allows us to use different sentence splitting strategies for different types of datasources. E.g - Machine Readable news vs RSS Feeds. In addition, we now also make sure that the same sentence splitting strategy is used during the GroundTruth labeling, model training as well as inferences. This new sentence splitting functionality can also be leveraged in a machine learning workflow for advanced users with a new lib-nlp step.

  • We have fixed a bug where a hard reload of the GroundTruth screen was needed after adding a new Candidate Set.

  • AI studio help button now takes you to the right documentation space.
    We have fixed a bug where retraining with a changed GroundTruth Set was not working.

Bug Fixes

  • We have fixed multiple issues around rendering of datasource configuration form for 1-click connectors.

  • We have also "migrated" the migration scripts such that they connect to MariaDB in the same fashion as the actual Squirro services does. This respects any SSL, encryption configuration for MariaDB on the server.

Breaking Changes

  • Newsletters have been moved to the server space from the project space as the newsletters are not project specific but rather server specific.

Fresh Installation Instructions

Please follow the regular installation steps

Upgrade Instructions

Please ensure that your current version is 3.1.0 or later. If you are on a version older than 3.1.0, please contact support.

With release 3.2.0 onwards, we have officially dropped the support for Python 2. If your custom Python plugins are not compatible with Python 3, please refrain from upgrading to any of the Squirro 3.2.x release until you have migrated the custom Python plugins to be Python 3 compatible.

yum update java-1.8.0-openjdk yum update squirro-storage-node-users # update storage node yum update squirro-storage-node yum update squirro-cluster-node-users yum update squirro-* # Resolve any `rpmnew` files. We anticipate `/etc/squirro/storage.ini`, `/etc/squirro/machinelearning.ini` and `/etc/nginx/conf.d/frontend.conf` to at least be resolved. systemctl reload nginx squirro_restart # Remove all orphaned python 2 packages yum erase squirro-python27*

Resolve all the .rpmnew files in /etc/squirro. This process involves merging the changes between the .ini (e.g., storage.ini) and .ini.rpmnew(e.g., storage.ini.rpmnew) files and then eventually deleting the .ini.rpmnew files. Finally, restart the services with merged .ini files (or just use squirro_restart to restart all services).

 

Upgrade the storage node by running:

yum update java-1.8.0-openjdk yum update squirro-storage-node-users # update storage node yum update squirro-storage-node

Upgrade the cluster node by running:

yum update java-1.8.0-openjdk yum update squirro-cluster-node-users yum update squirro-* # Resolve any `rpmnew` files. We anticipate `/etc/squirro/storage.ini`, `/etc/squirro/machinelearning.ini` and `/etc/nginx/conf.d/frontend.conf` to at least be resolved. systemctl reload nginx squirro_restart # Remove all orphaned python 2 packages yum erase squirro-python27*

Resolve all the .rpmnew files in /etc/squirro. This process involves merging the changes between the .ini (e.g., storage.ini) and .ini.rpmnew(e.g., storage.ini.rpmnew) files and then eventually deleting the .ini.rpmnew files. Finally, restart the services with merged .ini files (or just use squirro_restart to restart all services).

 

 

Upgrade every storage nodes (one by one) by running:

Upgrade each cluster nodes by running:

Resolve all the .rpmnew files in /etc/squirro. This process involves merging the changes between the .ini (e.g., storage.ini) and .ini.rpmnew(e.g., storage.ini.rpmnew) files and then eventually deleting the .ini.rpmnew files. Finally, restart the services with merged .ini files (or just use squirro_restart to restart all services).