mrg2fs.pl and negra2fs.pl -- readme
+++++++++++++++++++++++++++++++++++

Scripts mrg2fs.pl and negra2fs.pl serve for importing other treebank formats to FS format.


Supported formats
=================

Penn Treebank
-------------

Penn Treebank format is recognized by the script mrg2fs.pl:

( (S 
    (NP-SBJ 
      (NP (NNP Pierre) (NNP Vinken) )
      (, ,) 
      (ADJP 
        (NP (CD 61) (NNS years) )
        (JJ old) )
      (, ,) )
    (VP (MD will) 
      (VP (VB join) 
        (NP (DT the) (NN board) )
        (PP-CLR (IN as) 
          (NP (DT a) (JJ nonexecutive) (NN director) ))
        (NP-TMP (NNP Nov.) (CD 29) )))
    (. .) ))

Newlines are not important, as long as every sentence starts on a separate line.


Negra
-----

Negra export format is recognized by the script
negra2fs.pl. The introductory section of Negra export
format (BOT ... EOT) is ignored and only
sentences (BOS ... EOS) are converted.
Secondary edges are ignored as well.

%% word                 tag     morph           edge    parent  secedge comment
#BOS 1 1 985275570 1
Mögen                   VMFIN   3.Pl.Pres.Konj  HD      508
Puristen                NN      Masc.Nom.Pl.*   NK      505
aller                   PIDAT   *.Gen.Pl        NK      500
Musikbereiche           NN      Masc.Gen.Pl.*   NK      500
auch                    ADV     --              MO      508
...


Execution
=========

Both of the scripts share the same command line format and options:

{mrg,negra}2fs.pl flags file

The input files are processed one by one and converted to FS format. The
output files have the same names with an additional suffix .fs.


-d directory

The output directory where to save the converted files. Default: current directory.


-m output_file_name

Merge all input files together and write the output to the output_file_name.


-n number

Use in combination with -m. The option -n
limits the number of trees allowed in one output file. New files are created for the remaining trees. The
output_file_name should contain %d or a similar sprintf identifier to hold the output file number.