New Features

• New heat map dashboard widget
• New options for the result list dashboard widget:
  • Query highlighting can be disabled
  • Smart filter explain can be enabled
  • Abstract size can be configured
  • PDF can be configured to open on the first page where a match occurs
  • Documents can be opened from its original location instead of within an overlay
  • Disable the display of matching pages of a document
  • Disable the folding of duplicates of a document
  • The timestamp can be formatted
• Added the ability to set the abstract length for result sets on the API
• Multiple values of one facet can now be selected (and OR-combined) at the same time by pressing the ctrl (windows) or cmd (mac) key while selecting them

Improvements

• Wildcard & phrase queries are now highlighted in PDFs
• Combination of highlighting queries and explaining smart filters is now possible
• Dragging and dropping of widgets in the dashboard edit mode has improved
• Items are now stored/queried by referencing their subscription_id instead of their source_id. This will allow Squirro to process items from a source differently per project/object and also allows users to change the source (e.g. include more search terms in a search). See the section Source identifier handling for how this will affect existing code.
  Migration scripts will migrate the data seamlessly.

Bug Fixes

• time_increment query parameter is working again
• Ensure all recipients of an alert receive the digest email
• Ensure the timezones are correct when sending digest emails
• Ensure facet titles are sorted correctly
• Fix caching of keywords without values

Breaking Changes

The following changes may break existing integrations.

Source identifier handling

Items are now stored/queried by referencing their subscription_id instead of their source_id. This will allow Squirro to process items from a source differently per project/object and also allows users to change the source (e.g. include more search terms in a search). This has the following two breaking side-effects:

• Queries referencing sources now need to pass in their subscription IDs
• Items no longer include the source_ids field

Pipelet config

Pipelets constructors take an argument called config that passes in the user's configuration. Up til now, the handling of the config from the local pipelet consume command and when running in production on the server was different. This has been streamlined to work the same in both cases.

As a result, change any code snippets as follows (changed code is on line 3):

class TestPipelet(PipeletV1):
    def __init__(self, config):
        self.api_key = config['config']['api_key']

    def consume(self, item):
        # Do something with self.api_key
        …

class TestPipelet(PipeletV1):
    def __init__(self, config):
        self.api_key = config['api_key']

    def consume(self, item):
        # Do something with self.api_key
        …

Squirro client changes

The following methods of SquirroClient have been removed.

• get_project_source
• modify_project_source
• delete_project_source
• get_project_sources

Other changes

• Query expansion on a project level using query_string has been removed
• Pipelet uploading/editing is disabled by default. Talk to the operator in charge if you need Pipelet access

Deprecated functions/features

• The get_items method of the SquirroClient should no longer be used. Please use the query function. The get_items method will be removed in a future version.