NAME

ixBuild − build an index to a directory

SYNOPSIS

ixBuild [-i index][-lxXvVLFrRC][-a argStr][-P #][-d cmd][-D str][-A #][-f str][-c cmd][-S cmd][-s file][-p pri][-[nNtT] regex][[directory | file] ...]

DESCRIPTION

ixBuild creates or adds to an index for the specified files and directories, or for "." by default.  Each directory supplied to ixBuild is considered a master directory for the purposes of index construction.  ixBuild visits all of the files in the directory tree rooted at a master directory, extracting keywords and a description from each file visited.  With this information, it creates or updates an index located in a subdirectory called .index, which resides in the master directory. 
  An index contains two extensibly hashed databases built with db(3).  One associates keywords with the names of the files from which they were extracted.  The other provides a file name registration service used to determine the currency of the index with respect to the modification date of each file. 
 
By default, a filetype-specific keyword extraction program is used to obtain the keywords from each file visited. Similarly, a filetype-specific description generation program is used to obtain a description of each file visited. For most plain text documents, the default keyword extractor uses a prepared knowledge of English to filter out words that occur with about-average frequency, and to weight unusual words heavily.
 
For a complete discussion of the index construction process, see the file IndexingPaper in the directory /NextLibrary/Documentation/NeXT/RelNotes. 
 

OPTIONS

-i index
Causes ixBuild to use index instead of looking in the .index subdirectories of the master directories. 

-l Prevents ixBuild from following symbolic links. 

-x Causes ixBuild to print the information gathered for each file visited, but doesn’t add keys or registration information to the index structures. 

-X Causes ixBuild to do nothing other than determining the eligibility and currency of each file visited. 

-v Causes ixBuild to print a verbose trace of the indexing process on the standard error stream. 

-V Causes ixBuild to print the type and name of each file visited on the standard output.  This option is used by the Digital Librarian to update the display during index construction. 

-L Prevents ixBuild from folding keys to lower case. 

-F Prevents ixBuild from considering file names to be keywords. 

-r Causes ixBuild to read the names of the files and directories to be processed, and optionally their descriptions, separated from the names by white space, from the standard input, one name and optional description per line. 

-R Prevents ixBuild from removing extinct references from the index structures before adding new information. 

-C Prevents ixBuild from compressing indices following index construction. 

-a argString
argString is an optional argument list to be passed to pword(1) through the environment variable PWORD. 

-P percent
percent is the percentage of keywords to be extracted.  The value should be a floating point number between 0.0 and 1.0.  The default value is 1.0. 

-d cmd
Causes ixBuild to use cmd to generate a description for each file visited, rather than the default.  cmd is invoked with one argument, the name of the file in process, and should produce a description no more than 80 characters long on its standard output. 

-D str
Causes ixBuild to use str as the description of every file visited. 

-A #
Causes ixBuild to flush the index files under construction to disk every # minutes. The default value is 10. 

-f str
Causes ixBuild to use str as the file type of every file visited. 

-c cmd
Causes ixBuild to use cmd to generate keywords for each file visited, rather than the default.  cmd is invoked with one argument, the name of the file in process, and should produce a list of white space separated word/weight pairs on its standard output, one pair per line, as in pword(1). 

-S cmd
When this flag is supplied, the names of all files found to be both eligible for indexing and in need of indexing are piped to the standard input of cmd, one file name per line.  The files are not indexed by ixBuild. 

-s file
Causes ixBuild to use file as a stop list instead of /usr/lib/indexing/files/wf/DefaultEnglish.wf. 

-p pri
When this flag is supplied, the task scheduling priority - as in setpriority(2) - of the process group containing ixBuild is set to pri.  This is useful for reducing the proportion of system resources consumed by the indexing process, so that other applications are more responsive.  The default value is 20. 

-n str
Causes ixBuild to consider only files with names matching regex.  regex is a file name specification in Bourne shell syntax. 

-N str
Causes ixBuild to ignore files with names matching regex.  regex is a file name specification in Bourne shell syntax. 

-t str
Causes ixBuild to consider only files with types matching regex.  regex is a regular expression in the style of ex(1). 

-T str
Causes ixBuild to ignore files with types matching regex.  regex is a regular expression in the style of ex(1). 

FILES

 .index/{index,index.[DL],index.Registry.[DL],index.external}

− the index structures

SEE ALSO

ixClean(1), pword(1), wfcomp(1), ix(3), text(3)
/NextLibrary/Documentation/RelNotes/IndexingPaper

NeXT, Inc.  —  July 7, 1989