Classifications 4 (1) .2.0 Troubleshooting& FAQs Guide
Classifications 4 (1) .2.0 Troubleshooting& FAQs Guide
LLESCLS040200-TGD-EN-1
Livelink ECM – Classifications
Troubleshooting and FAQs Guide
LLESCLS040200-TGD-EN-1
Rev.: 2009-Jan-22
Open Text Corporation
275 Frank Tompa Drive, Waterloo, Ontario, Canada, N2L 0A1
Tel: +1-519-888-7111
Toll Free Canada/USA: 1-800-499-6544 International: +800-4996-5440
Fax: +1-519-888-0677
E-mail: support@opentext.com
FTP: ftp://ftp.opentext.com
For more information, visit http://www.opentext.com
1 Overview ......................................................................................9
1.1 Understanding Taxonomies ..................................................................... 9
1.2 Understanding Livelink Classifications................................................... 11
• User interface
This format is used for elements of the graphical user interface (GUI), such as
buttons, names of icons, menu items, names of dialog boxes, and fields.
• Filename, command, sample data
This format is used for file names, paths, URLs, and commands in the command
line. It is also used for example data, text to be entered in text boxes, and other
literals.
Note: If a guide provides command line examples, these examples may
contain special or hidden characters in the PDF version of the guide (for
technical reasons). To copy commands to your application or command
line, use the HTML version of the guide.
• Key names
Key names appear in ALL CAPS, for example:
Press CTRL+V.
• <Variable name>
The brackets < > are used to denote a variable or placeholder. Enter the correct
value for your situation, for example: Replace <server_name> with the name of
the relevant server, for example serv01.
Tip: Tips offer information that make your work more efficient or show
alternative ways of performing a task.
Note: Notes provide information that help you avoid problems.
Important
If this important information is ignored, major problems may be
encountered.
Caution
Cautions contain very important information that, if ignored, may cause
irreversible problems. Read this information carefully and follow all
instructions!
Knowledge Center
Open Text's corporate extranet and primary site for technical support. It is the
official source for:
• Open Text products and modules.
• Documentation for all Open Text products.
• Open Text Developer Network (OTDN): developer documentation and
programming samples for Open Text products.
• Patches for Open Text products.
The following role-specific information is available:
Partners
• Information on the Open Text Partner Program.
• Programs and support for registered partners.
Business Users
• Tips, help files, and further information from Open Text staff and other users
in one of the Open Text online communities
Administrators/developers
• Downloads and patches
• Documentation
• Product information
• Discussions
• Product previews
Feedback on If you have any comments, questions, or suggestions to improve our
documentation documentation, contact us by e-mail at documentation@opentext.com.
How it Works
After you enable Intelligent Classification, the Classification Importer process runs a
scheduled sequence of events beginning with a query of the Enterprise Server
indexes. The results are then read and parsed. The parsed content is placed into a
temporary location in the database, the DStagingImport table.
Once the content is in the DStagingImport table, an Import Agent transfers content
to a Task Handler, which requests the Classification drop-in (classresult). The
Classification drop-in performs Classification Query File updates and provides
ranking of the IPool content. Classifications are ranked based on the low threshold
score for the Classification and the get result X parameter in the Classification
Query File. Classifications are then returned for each item in the IPool. The
information in the LLClassify table is updated after each run of the Classification
Importer process.
Note: There are several processes that contribute to the overall success of Intel-
ligent Classification. If you are having difficulty applying assisted or automatic
Classifications using Intelligent Classification, see“Frequently Asked Ques-
tions” on page 21 for more information about each process.
• "get results score <X>", which instructs the search engine to return only items
with a relevance score greater than the Classification relevance score threshold X
If the queries.in file is not being updated, verify that the Update Distributor is
present and the Classification drop-in is loading correctly.
Note: If you are running Livelink Classifications in a clustered environment,
the Classification Query File is located relative to the Admin server of the
Update Distributor.
Enterprise data source should have been generated during installation. If not, you
can create one. If the Enterprise data source has been deleted, you will need to create
a new one and configure the data flow Classification Importer process. The Update
Distributor is required to classify only new items being added to Enterprise Server;
therefore, any previously classified items will remain unaffected by the loss of the
data source. An Enterprise data source can be created in the System Object Volume.
For more information about adding a data source, see Livelink Admin Online Help -
Index and Search Administration (LLESWBS-H-AGD).
Failure to load the drop-in will occur if the drop-in and OSpace versions do not
match, or if the file is not found. If the drop-in fails to load, the information appears
in the thread log similar to the following:
KDropinClass::LoadPackage( classresult.DLL ) about to load.
LoadLib::Open could not load: classresult.DLL (The specified
module could not be found.)
KDropin: Could not open Dropin library: classresult.DLL
KDropinClass::LoadPackage( classresult.DLL, DLLDropinInit ) Error
loading package.
C:\OPENTEXT10_0\module\classification_3_8_0\ospace Size: 362240
If the drop-in is not being loaded properly, re-run the Livelink Classifications
module installation program to copy the proper version of the drop-in to the
Enterprise Server bin directory (<Livelink_home>/bin).
Note: Only search results with a score greater than or equal to the relevance
score threshold are returned and processed as assisted or automatic Classifica-
tions. Results scoring below the relevance score threshold do not appear in im-
ported IPool messages or log files.
Important
Executing general Classification Profile Queries against large amounts of
content at one time may increase the load on the Import Agent, slowing the
server and the classification process.
xmlexport&scope=sub&nodeinfo&versioninfo=all&content&objID=
• Save as<cats.xml>.
3. On the destination Enterprise Server system, import the category definition
XML file, <cats.xml>. To import the Category Volume:
• Using the Livelink XML import feature in the address bar of your browser
type:
func=admin.XMLImport&
4. In the Classification taxonomy XML file, replace the node IDs of the categories
in all attribute reference tags imported from the source Enterprise Server system
with the new category node IDs that resulted from importing the category
definition XML file.
The following is an example of a Profile Query from the source Enterprise
Server system:
<query>SELECT "OTObject", "OTScore" WHERE ((
[ region "OTMeta":"Attr_24496_2" ]
"one" )) AND
("<Attr_24496_3>0</Attr_24496_3>")
ORDEREDBY RANKINGEXPRESSION (( [ region "OTmeta":"
Attr_24496_2" ] " one" )) AND ("<
Attr_24496_3>0</Attr_24496_3>")</query>
Note: In this example, 24496 represents the category node ID that must be
replaced when the Profile Query is exported to the destination Enterprise
Server system.
5. On the destination Enterprise Server system, import the Classification taxonomy
XML file using the Classifications Taxonomy Import. For more information
about importing taxonomies, see Administering Classifications (LLESCLS-H-
AGD).