fastr

Langue: en

Version: fastr main 2.04 (mandriva - 01/05/08)

Section: 1 (Commandes utilisateur)

NAME

fastr - FASt Term Recognizer

SYNOPSIS

fastr  [ -C conf-file ][ -v ] -z
fastr  [ -C conf-file ][ -v ] -c name
fastr  [ -C conf-file ][ -v ] -i -I -x -X [ name ]
fastr  [ -C conf-file ][ -v ] -p
fastr  [ -C conf-file ][ -v ] -l -L -t [ category inflection ]*
fastr  [ -C conf-file ][ -v ] -m
fastr  [ -C conf-file ][ -v ] -s terms column
fastr  [ -C conf-file ][ -v ] -V

For ease of access, the fastr manual has been split up into four sections:

     fastr        fastr overview (this section)
     fastrconf    fastr configuration file (computational parameters)
     fastrlang    fastr language file (linguistic parameters)
     fastrdata    fastr data (data structures)
 
 
 

(If you're intending to read these straight through for the first time, the suggested order will tend to reduce the number of forward references.)

DESCRIPTION

fastr is a multi-lingual tool for automatic indexing. fastr retrieves occurrences of terms and their variants from a full-text or tagged document. fastr must be provided with a list of controlled terms and a set of variation patterns. The terms are compiled into a dictionary while the variation patterns are included into the language file and re-compiled systematically.

The two steps of the computation are the following:

     term compiling (switches -z, -c)
     indexing (switches -i, -I, -x, and -X)
 
 
 
Additional utilities are provided in order to visualize compiled data:
     visualization of words, terms, and rules (switches -l, -L, and -t)
     visualization of meta-rules (switches -m)
 
 
 

to project lexical information on a corpus:

     addition of lexical features (switch -p)
 
 
 

and to reformat the index file:

     replacement of identifiers with strings (switch -s)
 
 
 

If the switch -C is provided, the parameters of the session are given by the file conf-file. Otherwise, the default configuration file is ./fastr.conf. The switch -v shifts to the verbose mode and outputs a trace (see the Log file in the fastrconf section).

SWITCHES

-c, --compile
Compiles the single-word list or the term list name into the current dictionary.
-i, --index-variants
Indexes the given file name, or the standard input if none is given. An index is an occurrence of a term of the current dictionary, or an occurrence of one of its variants.
-I, --index-exhaustively
Same as -i but reports all the occurrences of terms instead of reporting only their first occurrence in each sentence.
-l, --lists lexicon
Lists the lexicon entries. Chooses the first inflection if no inflection number is provided for this category in the command line.
-L, --lists lexicon and links
Same as -l but reports additionally the whole derivational and semantic families of each word.
-m, --lists metarules
Lists the metarules (see the Language file in the fastrlang section).
-p, --projection of lexical features
Analyzes morphologically the given file name, or the standard input if none is given. Ouputs the morphological features of each word in the input file.
-s, --strings
Substitutes the labels of terms in the column column of the standard input, by their string found in file terms (see the Path Label path in the fastrlang section).
-t, --lists lexicon and terms
Same as -l but reports additionally the rules linked to each word.
-V, --version number
About fastr.
-x, --index-exclusively-terms
Same as -i but only reports occurrences of terms which are not variants.
-X, --index-exclusively-variants
Same as -i but only reports occurrences of terms which are variants.
-z, --resets
Resets the current dictionary.

AUTHOR

Christian Jacquemin <jacquemin@limsi.fr>, LIMSI, Orsay, France... and many friendly collaborations.

NOTES

Versions 2.x differ significantly from versions 1.x. We apologize for the inconveniencies that may result from these transformations. These changes are intended to avoid future modifications.