Dir for the copyright-free corpus files compiled
at Giellatekno and Divvun.

Testing the validity of the corpus.

1. check whether each object file has exactly one meta file and vice-versa

command:
sh runObjMetaCheck.sh

result in output file:
obj-meta_check_DATE.txt

2. check whether the parallel file declared in a meta-file exists

3. ???


X. complex check after the XML-conversion:
   - parallelity
   - language flag
   - translation direction

 ==> todo: the scripts used for FAD-corpus testing have to be
     generalized for the whole corpus; for paralellity check,
     this means for ALL possible lang pairs combinations.

    Starting point: test_converted_corpus.xsl


========================
task: compile parallel corpora for sma and smj
========================

smj:
g -hr 'translated_from' .|awk '{print $3}'|tl

 222 select="'nob'"/>

para_20131125>find toktmx/smj2nob -name *.toktmx|wc -l
       8
para_20131125>find toktmx/nob2smj -name *.toktmx|wc -l
      47


 160 select="''"/>

   4 select="'swe'"/>
para_20131125>find toktmx/smj2swe -name *.toktmx|wc -l
       1
para_20131125>find toktmx/swe2smj -name *.toktmx|wc -l
       1

   3 select="'nno'"/>
para_20131125>find toktmx/nno2smj -name *.toktmx|wc -l
       7
para_20131125>find toktmx/smj2nno -name *.toktmx|wc -l
       4


sma:

  31 select="'nob'"/>

nob2sma>find . -name *.toktmx|wc -l
      24

para_20131125>find toktmx/sma2nob -name "*.toktmx"|wc -l
       4

  11 select="''"/>
   9 select="'nno'"/>

para_20131125>find toktmx/sma2nno -name "*.toktmx"|wc -l
       2
para_20131125>find toktmx/nno2sma -name "*.toktmx"|wc -l
       4

   2 select="'swe'"/>
para_20131125>find toktmx/sma2swe -name "*.toktmx"|wc -l
       1
para_20131125>find toktmx/swe2sma -name "*.toktmx"|wc -l
       1
==========================================================

ELAN and TEX file in the orig corpus

=====
Intro
=====

"spoken" is a new dir, created for (transcribed) spoken language data stored at freecorpus by the Freiburg based language documentation projects.

Naming this dir "spoken" is only a preliminary convention, because in GT prefers to make a genre distiction.

For more info see the readme file under ../sjd/spoken.