Skip to end of metadata
Go to start of metadata

You are viewing an old version of this page. View the current version.

Compare with Current View Page History

« Previous Version 11 Next »

Overview

The doc viewer has 2 components:

  • A viewer for documents loaded from the current query
  • A workspace "manager" for saving curated documents by hand or from an automated query

The following 2 sections briefly describe these functions

Document viewer

Purpose: View individual documents and their metadata, ranked by score or date, together with their entities, geotags, and associations.

  • Useful for analyzing specific documents once filtering or querying has reduced the dataset to manageable size.

Definitions:

  • Result Ranking: Ascending or descending by significance score, date, or title

  • Selected Doc Options

    • Unselect All - Deselects all selected documents

    • Select All Visible - Select all documents visible on the left side of Doc Viewer window

    • Filter Out Selected - Discards checked documents from results list AND other visualization widgets

    • Filter Out Unselected - Discards all unchecked docs, displays only checked document (all widgets)

  • Show Only:  Allows user to drill down on specific entity or document title

    • This text takes a regular expression - in addition, starting the string with "-" makes it a negative selector

  • Relevance: Calculation of entity’s linguistic importance in document

  • Frequency:  # of times an entity appears in the selected document 

  • Sentiment: Optional sentiment score provided by enrichment engine or third party 

Benefits:

  • Provides an “initial read” of document

  • Entity breakdowns w/ relevance/significance score

Limitations:

  • Does not display full-text articles, however, provides links to the original source as well as a link to the document metadata.

Dragging and dropping

The  icon can be used to drag selected (checked) documents in the Case Visualizer. The following entities will be linked:

  • Checked associations in the entity table
  • Entities from checked associations in the association table
  • Target entities, or entities already present in the case visualizer

The "Apply Text Filter To All Selected Docs" can be used to select entities across many documents (eg all "Person" entity types) before dragging multiple documents across.

The entity table or association tables can also be dragged into the Case Visualizer. This will create nodes and edges for all checked entities and associations

Extra hops

From the widget "extra options" (gear icon), it is possible to set the document viewer to grab "2-hop" and "3-hop" queries. This means that after performing the additional query:

  • (2-hops) a secondary query is performed to retrieve documents not matching the original query but containing one or more entities from the original query
  • (3-hops) a third query not matching either the original query, or the "2-hop" query, but containing one or more entities from the "2-hop" query

This capability can slow the system down a fair bit, so should be used with caution.

Workspaces

The latest (Sep 2014+) version of the doc viewer has 2 tabs in the left pane:

The left tab, provides the workspace view:

The options ("sub-tabs") underneath that tab are:

  • Buckets: lets users create lists of hand curated documents
  • Queues: lets users specify queries that will periodically fill the list. New documents can also optionally generate alerts.

To create a bucket or queue, simply type a new name in the text box of the appropriate sub-tab (it will auto complete against existing buckets/queues), and press "New".

Buckets and queues are both stored as Infinit.e shares and can be managed from the 5. File Uploader webapp:

  • deleting
  • sharing with other users/communities
  • setting the read-access to read/write or read-only

(by default when a queue/bucket is created it exists in the user's community so is not visible to anyone else)

Documents can be added to buckets (or removed from buckets or queues) using the two buttons at the bottom of the document view (right hand) pane.

For queues there are additional options:

  • To set/update the query, use the UI to create the query you want, then select "Unshared (update to current)" and press "Save"
  • The schedule will determine how often the query is run (again, "Save" to actually apply)
  • The "@" button will show an additional text field - email addresses to which alerts are sent can be entered into here (again, "Save" to actually apply)

 

Workspace Filtering

Doc Viewer: Select an entity from one of the document data tables, mouse over the  icon at the top left of the section header and select 'Filter Selected Items' from the list.

 

To remove a filter, click the  button located just below the query bar, this will revert the workspace back to the original query results.

 

Dragging to the case visualizer

Doc Viewer to Case Visualizer

  • Individual entities, geotags, or associations can be selected and dragged to the Case Visualizer by dragging directly from the list under the document details, or by checking the entity and dragging the  icon from the widget toolbar.

  • If multiple entities are dragged and dropped together, all entities will be linked by coreference.


  • An entire document (all corresponding entities, geotags, associations) can be dragged into the Case Visualizer by selecting the document and dragging the  icon from the widget toolbar 

         *Note: this is not recommended for documents with large amounts of entities or associations

  • Entities dragged from a document will contain a link to that document in the node properties

 

 

  • No labels