Table des matières
Text Grid Dramen Corpus Import Tutorial
Requirements
To follow this Tutorial you will need :
- Dramen.xml TEI corpus source file available at http://textgridrep.de/browse.html?id=textgrid:r0vm.0
- TXM 0.7.5
- “ExecXSLMacro” TXM Macro (See TXM Macros Documentation (in French) for more details) or any XSLT 2.0 processor)
- XSLT stylesheets
- CSS stylesheets
- tei.css (available at TEI Stylesheet repository
- TreeTagger German parameters file (available at http://www.cis.uni-muenchen.de/~schmid/tools/TreeTagger/ if you want to annotate the text
1. Prepare files
We will use TXM “XML/w + CSV” import module which needs each text of the corpus to be stored in a separate file. So the first step is to split the teiCorpus.
- Apply txm-filter-teicorpustextgrid-xmlw.xsl stylesheet to the Dramen.xml file
- you may use the TXM “ExecXSLMacro” for that purpose
- it will create an “out” subfolder and produce 8 separate TEI files named by concatenating the following metadata :
- creationDate
- author
- title
- it will also add attributes @author, @title and @creationDate to the <text> element which will available for corpus exploitation
- it will delete the teiHeader which does not belong to the source text
2. Primary tokenization
- Copy the source files produced by the xsl transformation to a folder, e.g. “drametemp”
- Open TXM and use File / Import / XML/w + CSV menu
- Select the source directory (where you placed the files)
- Start the import (“Play” button)
- You may stop the import as soon as “Tokenizartion complete” is dispayed on the console
3. Adjust word properties and perform final import
- Get the tokenized files from the
![$TXMHOME/corpora/dramentemp/tokenized foldel and place them to a new source folder, e.g. "dramen" (by default the folder name will become the name of the corpus)
- Select the new source directory in the TXM "Import parameters of XML:w + CSV" form
- Specify the main language "de" and check "Annotate the corpus" box if you want to annotate the files
* see [[http://txm.sourceforge.net/installtreetagger_en.html|http://txm.sourceforge.net/installtreetagger_en.html]] for instruction on how to use TreeTagger with TXM
- Select the txm-filter-teitextgrid-xmlw-posttok.xsl stylesheet in the "Front XSL" field. This stylesheet will :
* add @ref to every word (<w>) for default concordance references (filename + page number)
* normalize transform <speaker> elements into @who attribute of the <sp> element (to allow comparing speakers)
* raise initial <pb> tags as high as possible in the xml structure
- Run the import process till the end
===== 4. Customize edition pages =====
- Get the XML-TXM files from the $](/wiki/txm-users/lib/plugins/latex/images/renderfail.png "Fail: triggered security filter; contains blacklisted LaTeX tags.")
TXMHOME/corpora/dramen/txm/DRAMEN folder - Run the txm-edition-xmltxm-textgrid.xsl stylesheet on all XML-TXM files
- all original xml-tei elements will be transformed into <div>, <p> or <span> HTML elements with a @class created by concatenation of the original TEI element name and, if available, its @type and @subtype
- Save or rename the standard output files with an .html extenstion
- Run the txm-edition-page-split.xsl stylesheet on every .html file with the parameter cssname=txm-textgrid
- the results will be written to the “default” subfolder
- Create a “css” subfolder in the “default” sirectory and copy the tei.css and txm-textgrid.css files there
- Replace the original “default” folder in $TXMHOME/corpora/dramen/HTML/DRAMEN with the one you have just generated
- Your TXM corpus is ready!
- you can customize the style of the TXM edition by editing the txm-textgrid.css file
