Versions Compared

Key

  • This line was added.
  • This line was removed.
  • Formatting was changed.


Excerpt

We're excited to announce Squirro 2.6.5, released on 26th of September 2019.

Contents -What's in the release?

Table of Contents
minLevel2
excludeContents


New Features

...

  • Global Search
    • Global Search is placed in the app bar only for the dashboard space. It runs a search on the Master dashboard (see below) to quickly find app content.
    • Global search query can be a normal search or a NLP query to search any result across the app which technically will be a search inside the Masters Dashboard.
    • It will not filter down the active dashboard tab as it is supposed to search for results not readily placed under the visible dashboards.
    • A user can type a query, upon pressing enter, the search is submitted. Any layer/s from the master dashboard that matches the visibility condition is made visible.
  • Master Dashboard
    • Master dashboard is essentially a templating dashboard where by using layers, one creates visual templates for different query parameters (including NLP) and visibility conditions.
    • This dashboard is not designed to be any of the main dashboards shared with the end-user, instead it is supposed to show layers which meet a visibility condition caused by an NLP or normal query from the Global Search bar.
    • By default, a project does not have a dashboard marked as master but any dashboard in a project can be marked as master.
    • Only one dashboard at a time can be a master.
    • Master dashboard is not visible in the Dashboard space.
    • This dashboard controls the visibility of the Global Search bar in the app bar. If a dashboard is marked as master, Global Search bar will appear, otherwise not.
  • NLP
    • We have built the foundation for a new feature we call Conversational UI. This enables an NLP driven search.
    • NLP queries are enabled from the Global Search bar and display the results in the consumption mode.
    • A Master Dashboard is built with layers as templates designed to be visible for all combinations of NLP parameters. These layers are displayed as NLP query results.
    • In order to answer to the NLP queries, the master dashboard layers should have NLP aggregations rules defined in the visibility condition.
    • It is also possible for a layer to be configured to 'Use NLP if available'. In that case, any manual configuration of the widgets inside that layer is ignored and the widgets are automatically configured to answer to NLP queries.
    • This feature is still highly experimental and therefore disabled by default. To enable it a few setup and configuration steps are required. Please reach out to Squirro if you are interested in an early access to this feature.
  • Voice Search
    • Voice search is an experimental feature enabled from the Global Search bar and only supported on Chrome browser for now.

...

  • JSON data loader
    • We have introduced a new JSON data loader via the frontend. It is placed in the Data Import section.
    • The dataloader supports:
      • Files with single JSON object (auto-detect)
      • Files with a list of JSON objects (auto-detect)
      • Files with nested JSON objects, i.e. the JSON objects are not top-level. For this, specify the item-schema option in dataloader
      • Directories (only via CLI)
    • See documentation on how to use it via CLI
  • Addition of static keyword for data source on Frontend
    • Option to add keywords to all items coming from a particular data source during the data loading process.
      On Step 4: 
  • Behaviour of duplicated source changed
    • When duplicating a source, the source is now in "paused" state, which allows editing the data source configuration before starting the data load by "Resume".

Filtering and Pagination on User Management Page

  • User management on the server now enables searching for users and a .
  • A pagination functionality for server users was also implemented.
  • Search functionality was also added to Project Member page.

...

Project Export/Import of Squirro Items

  • Project Export/Import /Export enables now the export of items from a Squirro Project and the import in a new Squirro Project.
    (Support for exportExport/import Import of binary documents is not supported.)
  • The data imported like this will all be added as a single source called "Project Import".
  • This functionality facilitates the migration of projects between servers.

APP & Nav Bar Theming


  • Functionalities:
    • Custom Logo to replace Squirro logo next to Project Name
      • Set custom logo via image-URL
      • max. size of logo should be 50 kb to render smoothly.
    • Set App/Nav bar background colour
    • Set App / Nav bar text colour
    • Set Brand colour
    • Set Highlight colour
    • Colours can be defined by Hex Colour Code or by simply writing a colour name (e.g. purple)

      Image Added


Widgets & Dashboard

New Widgets

...

  • Allow running the same pipelet multiple times with different configuration
  • Instead of using class name as pipelet key, use new field: ‘class_name'. See documentation

Data

...

Source Duplication

  • When duplicating a source, the source is now in "paused" state, which allows editing the data source configuration before starting the data load by "Resume".

Pipelet rerun

  • About 2x speed improvement
  • Remove the 10,000 items limit → able to use on now any number of squirro items.

LineChart API improvement

  • Extended the time_selection range to support per-second aggregation (time_selectiion=Xsecond).
  • Better management of the query_aggreation aggregation when time_selection is not present in the query.

...