File alignment plan

#Intro

Plan for aligner and analysis work

Goal: Send analysed texts to Oslo on order



1. find parallel texts
2. align
    several alignments of the same text
3. analyse the sme part
4. send to Oslo


Solution:

1. A document in CVS where we check in what is done
2. Have a directory on victorio (G5?) where we keep track of what is done
Store the aligned files, so that we can see how old they are
Processing phases:
corpus conversion  file.sme.xml, file.nob.xml
numbered sentence phase file.sme.xml.sent, file.nob.xml.sent
aligned  sentence phase (alignment.jar => xml_cor.sent files
analysed phase (corpus-analyze.pl =>

To be sent:
file.nob.xml.sent (analysed in oslo, for the time being, at least)
file.sme.xml.sent.analyzed
file.sme.xml_file.nob.xml.xml

cor = result of the alignment

Principled names:
file.sme_new.txt               numbered <s id="n"> nodes only. This is
                               the file which tca2 actually reads.
file.sme.xml_file.nob.xml.xml  <s> number correspondences only
file.sme.xml_cor.sent          original xml file with sme and nob <s> numbers

Example:
dc_1_01.doc.xml_sp_1_01.doc.xml.xml
dc_1_01.doc.xml_cor.sent
dc_1_01.doc.xml_new.txt

parallel/
 namm-namn/
  namn.sent
  namm.sent
  namm.sent.analyzed
  namm.xml_namn.xml.xml
 ...

1. Trond: Find texts, mark the xsl files (cf. dc_ files for model), tell Saara
2. Saara: make .sent and .sent.analyzed
   Børre: make namm_namn.xml
3. Trond: Send files to Lars