7
Integrating interMedia Text in WebDB Site Builder

The chapter contains the following sections:

7.1 Prerequisites

Oracle8i interMedia Text creates indexes that let users perform fast and accurate text searching, from traditional full-text search to document theme analysis.

The base for interMedia Text is Oracle's ConText Cartridge. For more information, see "Oracle8i interMedia Text Migration" and "Oracle8i interMedia Text Reference."

Before you integrate interMedia Text into WebDB Site Builder, you must do the following:

Install and configure the Oracle8i database to use interMedia Text. This is done using the Database Configuration Assistant. See the Oracle8i documentation for more information about installing and configuring an Oracle8i database to use interMedia Text.
Install Oracle WebDB 2.2 according to the instructions in the appropriate sections of this chapter.
Make sure that your system environment variable is set to the ORACLE_HOME in which your Oracle8i database is located. For example, C:\Oracle\Ora815.
With the Oracle Home Selector in your Oracle Installation Products program group, set your Primary Oracle Home to the ORACLE_HOME in which your Oracle8i database is located.
Create a WebDB Web site using the WebDB Site Creation Wizard. See "Creating WebDB Sites" in this guide for more information on completing this task.

Note:
Reboot your computer where appropriate when completing the above prerequisites.

7.2 Setting Up Your Environment for interMedia Text

interMedia Text in Oracle8i uses external procedures to perform document filtering. You must configure the LISTENER.ORA, TNSNAMES.ORA, and SQLNET.ORA files in your Oracle8i Home directory to support these external procedures.

The LISTENER.ORA, TNSNAMES.ORA, and SQLNET.ORA files on your computer should look similar to the following examples:

7.2.0.1 LISTENER.ORA

LISTENER = 
(DESCRIPTION_LIST =         
    (DESCRIPTION =                 
         (ADDRESS_LIST =                         
                 (ADDRESS = (PROTOCOL = IPC)(KEY = EXTPROC0))                 
         )                 
              (ADDRESS_LIST =                         
                    (ADDRESS = (PROTOCOL = TCP)(HOST = my-pc)(PORT = 1521))                 
              )         
         ) 
)

SID_LIST_LISTENER =  
   (SID_LIST = 
   (SID_DESC =  
        (SID_NAME = PLSExtProc) 
        (ORACLE_HOME = /oracle/home)
         (ENVS=LD_LIBRARY_PATH=/oracle/home/lib:/oracle/home/ctx/bin:
                              /oracle/home/bin:/oracle/home/ctx/lib)
       (PROGRAM= extproc)
)
(SID_DESC =                         
         (GLOBAL_DBNAME = ORCL)                         
         (ORACLE_HOME = /oracle/home)                         
         (SID_NAME = ORCL)                 
   )     
)

Note:
The above lines in bold apply to Solaris installations of WebDB only. Ignore these lines for Windows NT installations.

7.2.0.2 TNSNAMES.ORA

ORCL =         
   (DESCRIPTION =                 
     (ADDRESS_LIST =                         
             (ADDRESS = (PROTOCOL = TCP)(HOST = my-pc)(PORT = 1521))        
)         
     (CONNECT_DATA =                 
             (SERVICE_NAME = ORCL)         
     )
)
EXTPROC_CONNECTION_DATA =         
      (DESCRIPTION =                 
             (ADDRESS_LIST =                         
                  (ADDRESS = (PROTOCOL = IPC)(Key = EXTPROC0))         
)         
(CONNECT_DATA =                 
             (SID = PLSExtProc)         
             ) 
)

7.2.0.3 SQLNET.ORA


names.initial_retry_timeout = 30

names.max_open_connections = 3

names.message_pool_start_size = 10

Note:
See the "Net8 Administrator's Guide" included with your Oracle8i documentation for more information about creating these configuration files.

7.3 Defining DEFAULT_LEXER Values

Before creating interMedia Text indexes in Oracle WebDB, make sure that the database has the correct DEFAULT_LEXER value defined for the language.

Follow these steps to set up a DEFAULT_LEXER value:

To do so, execute the following SQL command in SQL*Plus as ctxsys user:

SELECT PRE_NAME, PRE_OBJECT FROM CTX_USER_PREFERENCES 
WHERE PRE_NAME  = `DEFAULT_LEXER';

The value for PRE_OBJECT returned by the above query may be any of the following:

LEXER TYPE Description

BASIC_LEXER

for single byte database

CHINESE_VGRAM_LEXER

for a Chinese database

JAPANESE_VGRAM_LEXER

for a Japanese database

KOREAN_LEXER

for a Korean database

(For single byte databases only) Execute the following SQL command in SQL*Plus as the ctxsys user:
```
    SELECT PRV_PREFERENCE, PRV_ATTRIBUTE,PRV_VALUE 
    FROM CTX_USER_PREFERENCE_VALUES 
    WHERE PRV_PREFERENCE LIKE `DEFAULT%'
```
Ensure that the language-specific attributes have correct values.

LEXER TYPE	Description
BASIC_LEXER	for single byte database
CHINESE_VGRAM_LEXER	for a Chinese database
JAPANESE_VGRAM_LEXER	for a Japanese database
KOREAN_LEXER	for a Korean database

If the DEFAULT_LEXER is not properly defined, it needs to be recreated. Drop the existing DEFAULT_LEXER preference by running the following commands in SQL*Plus as the ctxsys user.

BEGIN 
   CTX_DDL.DROP_PREFERENCE(`DEFAULT_LEXER'); 
   CTX_DDL.DROP_PREFERENCE(`DEFAULT_WORDLIST'); 
   CTX_DDL.DROP_STOPLIST(`DEFAULT_STOPLIST'); 
END;

Execute the appropriate language-specific script located in the <ORACLE8i_HOME>\ctx\admin\defaults directory. The format of the files in this directory is drdef<language>.sql. For example, the US script is drdefus.sql. The Japanese script is drdefja.sql. You must run the language-specific script in SQL*Plus as the ctxsys user.
Execute the following SQL command in SQL*Plus as the ctxsys user:
```
SELECT PAR_NAME,PAR_VALUE FROM CTX_PARAMETERS 
WHERE PAR_NAME in (`DEFAULT_LEXER','DEFAULT_STOPLIST','DEFAULT_WORDLIST');
```
PAR_NAME PAR_VALUE

DEFAULT_LEXER

CTXSYS.DEFAULT_LEXER

DEFAULT_STOPLIST

CTXSYS.DEFAULT_STOPLIST

DEFAULT_WORDLIST

CTXSYS.DEFAULT_WORDLIST

If the values returned by the query do not match those in the above table, use the CTX_ADM package to set them. For more information, refer to the interMedia Text reference manual.

PAR_NAME	PAR_VALUE
DEFAULT_LEXER	CTXSYS.DEFAULT_LEXER
DEFAULT_STOPLIST	CTXSYS.DEFAULT_STOPLIST
DEFAULT_WORDLIST	CTXSYS.DEFAULT_WORDLIST

7.4 Creating interMedia Text Indexes

Once you have properly set up your environment for interMedia Text, you can create interMedia Text indexes in the following way:

On the WebDB Site Builder Administration page:

Click Search under the Toolbox section.
Click the interMedia Text tab.
Configure the interMedia Text fields as required:
Check Enable interMedia Text Features.
If desired, check Enable Theme & Gists.

Note:
Theme & Gists are currently supported on WebDB sites contained in English-language databases only.
Choose a color for the Search Highlight Text Color.
Choose the number of Hits Per Page.
Specify the number of seconds for the URL Connection Time-out.
Specify the HTTP Proxy.
If necessary, specify the No Proxy Servers for Domains Beginning With information.

You can create the index in either of these ways:

Click Create Basic Index.
The interMedia Text index is created on the server.

Note:
To get help for the current page, click the Help icon on the page.
Or, you can run the following command in SQL*Plus. Log on using the user name and password for the schema that owns the WebDB site. For example, if the schema is named "SCOTT", log on with user name and password, "SCOTT".
```
ctxcrind.sql
```

The following interMedia indexes are created:

Index Name Description

CONTEXT_AUTHOR_INDEX

Index based on the items' author.

CONTEXT_DESC_INDEX

Index based on the items' description.

CONTEXT_KEYWORDS_INDEX

Index based on the items' keywords.

CONTEXT_TITLE_INDEX

Index based on the items' title.

FILE_INDEX

Index based on the items' filename.

TEXT_INDEX

Index based on the items' text or content.

URL_INDEX

Index based on the items' Uniform Resource Locator (URL).

Index Name	Description
CONTEXT_AUTHOR_INDEX	Index based on the items' author.
CONTEXT_DESC_INDEX	Index based on the items' description.
CONTEXT_KEYWORDS_INDEX	Index based on the items' keywords.
CONTEXT_TITLE_INDEX	Index based on the items' title.
FILE_INDEX	Index based on the items' filename.
TEXT_INDEX	Index based on the items' text or content.
URL_INDEX	Index based on the items' Uniform Resource Locator (URL).

Notes:

The time required for creating indexes varies depending on the number of items you have on your WebDB site. A message is displayed when the index is created.

When specifying a proxy server, do not prefix http:// to the proxy server name.

Make sure that you have the LD_LIBRARY_PATH variable correctly set for the extproc in your LISTENER.ORA file. See "Setting Up Your Environment for interMedia Text" on page 3.

7.5 Updating interMedia Text Indexes

To update the interMedia Text indexes when you add items to your WebDB site, you must have interMedia Text Server running or you can issue the following command:

ALTER INDEX with parameters ('sync')

For more information about the interMedia Text Server and the ALTER INDEX command, see the interMedia Text documentation.

7.6 Dropping interMedia Text Indexes

You can drop interMedia Text indexes in either of these ways:

On the WebDB Site Builder Administration page:
1. Click Search under the Toolbox section.
2. Click the interMedia Text tab.
3. Click Drop Basic Index.
  The interMedia Text index is deleted from the server.
Or, to drop an interMedia Text Index, run the following script in Oracle SQL*Plus logging in as the WebDB schema owner:
```
ctxdrind.sql
```

7 Integrating interMedia Text in WebDB Site Builder