If you create a custom text analysis engine, you must add
it to the system before you can use it for search or text analysis.
Collections can use the engine to analyze and annotate documents and
improve the precision of search results.
About this task
Application developers can create a processing engine
archive (.pear) that adheres to the UIMA framework
for text analysis. The archive includes all of the resources required
to query collections. To be able to query collections with your custom
analysis algorithms, you must add the archive (which contains the
text analysis engine) to the system.
After you add a text analysis
engine to the system, you can change its display name and select an
option to view the XML source. The XML source shows you what information
is produced by this engine.
If a text analysis engine is associated
with a collection, you cannot remove the text analysis engine from
the system.
Procedure
To add a custom text analysis engine to the system:
- Click System to open the System view.
- On the Parse page, click Configure
system text analysis engines.
- On the Text Analysis Engines page,
click Add System Text Analysis Engine.
- On the Add a Text Analysis Engine page,
type a descriptive name for the new engine. The system
uses this display name to identify the text analysis engine throughout
the administration console.
- Specify whether you want to use the descriptor in your
custom processing engine archive to isolate its components from the
linguistic annotators that are provided with Watson Explorer Content Analytics. If you exported the .pear from Content Analytics Studio, you can avoid conflicts
by allowing the system to set up a localized class path and internal
resource manager for your custom text analysis engine. To enable this
option for a multiple-server configuration, you must ensure that all
servers use the same data directory path (ES_NODE_ROOT).
- Specify the location of the .pear file. If the file is smaller than 64 MB, the file can be on your local
computer and you can browse to locate the file. If the file is larger
than 64 MB, the file must be on the master server and you must type
the fully qualified path for the file.
- Click OK. Your text analysis engine is listed on the Text Analysis Engines
page.