What’s New in IN-SPIRE™ 5.8

IN-SPIRE continues to push the boundaries of text analysis and visualization with the release of version 5, build 5.8. This release is packed with new features that make it faster, easier to use, more powerful and stable, and includes a robust and global capability for filtering of documents with the "focus" and "ignore" features.

User Interface Changes

Focus and Ignore

In prior releases, you could focus in on a subset of the document collection, but only certain tools would use the context of the smaller subset of data. Now, all tools within IN-SPIRE respond to Focus and Ignore. When focus is enabled, only the documents that are in the Focus panel and not in the Ignore panel will be visible in the various IN-SPIRE visualizations. Search results will be constrained to these focused documents as well.

The operations for the Focus panel were also changed to separate the Focus box from the Ignore box. The Ignore box operates independently from the Focus box, allowing you to add and remove different combinations of groups and selections from the Focus box while continuing to ignore the documents you identified as outliers or junk documents.

Additionally, new themes are calculated each time focus is enabled. This results in themes that are more relevant to only the focused document set, as opposed to themes that are reflective of the entire dataset.

Number options for Themeview Classic

For datasets that have numeric fields, the Themeview Classic visualization has added preliminary support for displaying aggregate or statistical measures based on the aggregated collection of number data. This capability should be considered experimental.

Facet sort by Association

An additional sorting option was added to Facets called " Order by Association" which provides rankings that are more relevant to the selected set of documents rather than raw counts of documents assigned to a theme. When selected, the facet items will be sorted based on association with your selected set. In the example below, the user has chosen "Order by Association" and has clicked on the "international organizations" theme. In the Locations facet, the item for "choucha" sorts above the item for "niger" because 100% of the documents in the "choucha" facet are associated with "international organizations", but only a small percentage of the "niger" documents are associated with "international organizations". If "Order by Document Count" was used instead, the rankings would place "niger" first, followed by "choucha". This effect is similar to checking the "% selected within group" checkbox when displaying Themes in the Groups Folder widget on the left side of the main IN-SPIRE windows.

Lexical labels, colors, invert axis

Enhancements were made to the Lexical Analysis Comparison tab to make the visualization more meaningful. Labels were added to the axis, and the lexical scores for each group are displayed on the chart. The color palette was also changed for this visualization, and in some cases, the vertical axis of the chart was inverted to provide more consistency between different lexical axes.

Quick "Search within Selection"

The buttons have been changed for the Quick Search panel to make it easier to operate. An explicit button has been added to perform a search, a new button has been added to perform a Search within Selection, and Advanced has been moved so it is not confused with performing a quick search.

Keyboard shortcuts

Additional keyboard shorts were added to common operations. These are documented in the online help for IN-SPIRE, and some are visible in tool tips above the button that normally performs the operation. Some new shortcuts include:

• Control-F: Focus the current selection
• Control-G: Group the current selection

• In the Quick Search panel, pressing Enter performs a search across the entire dataset, and pressing Control-Enter performs a search within the Current Selection.

Dataset Wizard field selection enhancements

It is now easier when creating or editing datasets, to change the options for which fields are Topical, which are Categorical, which field is the Title field, and which is the Date field.

Major term panel sort options (num characters, words)

Additional columns have been added to the Major Terms tool to give additional insights into the characteristics of the major terms that IN-SPIRE has selected for your dataset. These include number of characters and number of words in each term. These enhancements make it easier to triage shorter, insignificant words, or identify key phrases that IN-SPIRE's engine selected as major terms.

Ingest password protected PDF/Word files

For Microsoft Word and Adobe PDF datasets, IN-SPIRE has added support for password protected documents. When this option is selected, IN-SPIRE will prompt for a series of passwords to use to try to decrypt password protected documents. The passwords are not persisted with the dataset, although an unencrypted version of the file is saved within the dataset.

Other Changes

Separation of Programs and Data

When installing IN-SPIRE, you can now specify different locations for the "Programs and Executable" pieces of IN-SPIRE, and the "data" portions of IN-SPIRE. The "data" location must be a read/write location, but the "Programs and Executables" location can be read-only.

Entity Extraction for RSS Feeds

The RSS Feeder now supports an option for enabling GATE entity extraction for all articles harvested through an RSS Feed.

Entity Extraction for ANVR

The ANVR specification has been enhanced to support entity extraction during dataset creation.

Bundled IN-SPIRE utilities

IN-SPIRE is now bundled with additional utilities to assist importing, exporting, and sharing analytic artifacts.

• The IN-SPIRE List Importer can create Search Networks, Viewpoints and Lexicons from a text file containing a list of terms
• Meta Data Sharing can export and import Search Networks, Viewpoints and Lexicons, allowing them to be shared among users more easily
• IN-SPIRE Advanced Settings provides a means for experienced users with specialized data analysis needs to experiment with non-standard text processing settings. These include the ability to bias the selection of phrases over single words for use in clustering and keywords, and the ability to increase the number of major terms and/or clusters used in a dataset.
• VizTable Reconstructor allows you recover a corrupted viztable. The viztable is the internal data store that IN-SPIRE uses to keep track of its datasets

Fixed in 5.8.3:

INSPIRE-1771: Editing a dataset in a client/server installation was failing

INSPIRE-1773, INSPIRE-1774: Installation related issues that cause the program to not run when IN-SPIRE environment variables are missing

INSPIRE-1789: Excel/CSV was defaulting to non-case sensitive field delimiters

Release Notes:

INSPIRE-1664: IN-SPIRE may crash when attempting a focus operation on a large datasets, and the number of documents being focused on is very small. Example: focusing on 20 documents within a dataset of 100,000 documents.

INSPIRE-1665: Facets may intermittently show wrong facet counts when making multiple facet item selections within the same facet column.

INSPIRE-1688: Focus and Ignore may not work properly when attempting to Ignore all of the documents in your dataset.

INSPIRE-1748: Uninstalling IN-SPIRE removes the viztable.local file. This file should be preserved during uninstall. If you plan to reinstall IN-SPIRE, it is recommended to make a backup of this file before uninstalling a previous version.

INSPIRE-1759: The ADDTO function of ANVR is not functioning properly. This feature should be avoided. Work around is to create a standalone dataset or a real-time dataset.