'\" t
.\" $Id$
.tr ~
.TH GRIND 1 "Dec 2006" "WordNet 3.0" "WordNet\(tm User Commands"
.SH NAME
grind \- process WordNet lexicographer files
.SH SYNOPSIS
\fBgrind\fP [ \fB\-v\fP ] [ \fB\-s\fP ] [ \fB\-L\fP\fIlogfile\fP ] [ \fB\-a\fP ] [ \fB\-d\fP ] [ \fB\-i\fP ] [ \fB\-o\fP ] [ \fB\-n\fP ] \fIfilename\fP [ \fIfilename\fP\&.\|.\|. ]
.SH DESCRIPTION
\fBgrind(\|)\fP processes WordNet lexicographer files, producing
database files suitable for use with the WordNet search and interface
code and other applications.  The syntactic and structural integrity
of the input files is verified.  Warnings and errors are reported via
\fBstderr\fP and a run-time log is produced on \fBstdout\fP.  A
database is generated only if there are no errors.
.SS Input Files
Input files correspond to the syntactic categories implemented in
WordNet \-
.BR noun ", " 
.BR verb ", " 
.BR adjective " and " 
.BR adverb .
Each input lexicographer file consists of a list of synonym sets
(\fIsynsets\fP) for one part of speech.  Although the basic synset
syntax is the same for all of the parts of speech, some parts of the
syntax only apply to a particular part of speech.  See
.BR wninput (5WN) 
for a description of the input file format.

Each \fIfilename\fP specified is of the form:

.RS
.IB pathname / pos . suffix
.RE

where \fIpathname\fP is optional and \fIpos\fP is either
.BR noun ", " 
.BR verb ", " 
.BR adj " or " 
.BR adv .
\fIsuffix\fP may be used to separate groups of synsets into different
files, for example \fBnoun.animal\fP and \fBnoun.plant\fP.  One or
more input files, in any combination of syntactic categories, may be
specified.  See
.BR lexnames (5WN)
for a list of the lexicographer files used to build the complete
WordNet database.
.SS Output Files
\fBgrind(\|)\fP produces the following output files:

.TS 
center box ;
c | c
l | l.
\fBFilename	Description\fP
_
\fBindex.\fIpos\fR	Index file for each syntactic category
\fBdata.\fIpos\fR	Data file for each syntactic category
\fBindex.sense\fP	Sense index
.TE

See 
.BR wndb (5WN) 
for a description of the database file formats.

Each time \fBgrind(\|)\fP is run, any existing database files are
overwritten with the database files generated from the specified input
files.  If no input files from a syntactic category are specified,
the corresponding database files are not overwritten.
.SS Sense Numbers
Senses are generally ordered from most to least frequently used, with
the most common sense numbered \fB1\fP.  Frequency of use is
determined by the number of times a sense is tagged in the various
semantic concordance texts.  Senses that are not semantically tagged
follow the ordered senses in an arbitrary order.  
Note that this ordering is only an
estimate based on usage in a small corpus.

The \fItagsense_cnt\fP field for each
entry in the \fBindex.\fIpos\fR files indicates how many of the senses
in the list have been tagged.

The \fBcntlist\fP file provided with the database lists the number of
times each sense is tagged in the semantic concordances.
\fBgrind(\|)\fP uses the data from \fBcntlist\fP to order the senses
of each word.  When the \fBindex\fP.\fIpos\fP files are generated, the
\fIsynset_offset\fPs are output in sense number order, with sense 1
first in the list.  Senses with the same number of semantic tags are
assigned unique but consecutive sense numbers.  The WordNet
.SB OVERVIEW
search displays all senses of the specified word, in all syntactic
categories, and indicates which of the senses are represented in the
semantically tagged texts.
.SH OPTIONS
.TP 15
.B \-v
Verify integrity of input without generating database.
.TP 15
.B \-s
Suppress generation of warning messages.  Usually \fBgrind\fP is run
with this option until all syntactic and structural errors are corrected
since the warning messages may make it difficult to spot error
messages.
.TP 15
.BI \-L logfile
Write all messages to \fIlogfile\fP instead of \fBstderr\fP.
.TP 15
.B \-a
Generate statistical report on input files processed.
.TP 15
.B \-d
Generate distribution of senses by string length report on input files
processed. 
.TP 15
.B \-i
Generate sense index file.
.TP 15
.B \-o
Order senses using \fBcntlist\fP.
.TP 15
.B \-n
Generate nominalization (derivational morphology) links in database.
.TP 15
.I filename
Input file of the form described in 
.SB Input Files.
.SH FILES
.TP 20
.B \fIpos\fP.*
lexicographer files to use to build database
.TP 20
.B cntlist
file of combined semantic concordance \fBcntlist\fP files.  Used to
assign sense numbers in WordNet database
.SH SEE ALSO
.BR cntlist (5WN),
.BR lexnames (5WN),
.BR senseidx (5WN),
.BR wndb (5WN),
.BR wninput (5WN),
.BR uniqbeg (7WN),
.BR wngloss (7WN).
.SH DIAGNOSTICS
Exit status is normally 0.
Exit status is \-1 if non-specific error occurs.
If syntactic or structural errors exist, exit status is number of
errors detected.
.TP 
.B "usage: grind [\-v] [\-s] [\-Llogfile] [\-a ] [\-d] [\-i] [\-o] [\-n] filename [filename...]"
Invalid options were specified on the command line.
.TP 
.B No input files processed.
None of the filenames specified were of the appropriate form.
.TP 
.B \fIn\fP syntactic errors found.
Syntax errors were found while parsing the input files.
.TP 
.B \fIn\fP structural errors found.
Pointer errors were found that could not be automatically corrected.
.SH BUGS
Please report bugs to \fBwordnet@princeton.edu\fP.