thdltools/Jskad
1
0
Fork 0
Code Issues Pull requests Projects Releases Wiki Activity

added saxon8 to repository to support xslt 2.0 transformations; these are used in transforming THDL's wylie transcripts into unicode transcripts

Browse source
This commit is contained in:
eg3p 2007-05-17 18:20:12 +00:00
parent 4f553caf54
commit 0440ef5ffb
3 changed files with 54 additions and 44 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

7
archive/styles/get-list-of-transcripts.xsl
View file

@ -1,6 +1,9 @@
<?xml version="1.0" encoding="UTF-8"?>
<xsl:stylesheet xmlns:xsl="http://www.w3.org/1999/XSL/Transform" version="1.0">
<xsl:stylesheet version="1.0"
xmlns:xsl="http://www.w3.org/1999/XSL/Transform"
xmlns:java="java" >
<xsl:output method="text" encoding="utf-8"/>
<xsl:template match="/">
@ -8,6 +11,6 @@
</xsl:template>
<xsl:template match="transcript">
<xsl:value-of select="."/><xsl:text> </xsl:text>
<xsl:value-of select="java:net.URLEncoder.encode(.,'UTF-8')"/><xsl:text> </xsl:text>
</xsl:template>
</xsl:stylesheet>

BIN
extensions/saxon/saxon.jar Normal file
View file

Binary file not shown.

83
lucene-thdl-build.xml
View file

@ -27,6 +27,12 @@
</fileset>
</path>
<path id="saxon.classpath">
<fileset id="saxon.extensions" dir="${ext}/saxon">
<include name="*.jar"/>
</fileset>
</path>
<!--
<target name="segment-videos">
<fileset dir="${segmentation.instructions.dir}" id="segmentation.instructions">
@ -43,6 +49,45 @@
</target>
-->
<!-- archive tasks -->
<!-- FIXME: title_metadata.xml is not being retrieved as UTF-8!! -->
<target name="archive-get-metadata">
<mkdir dir="${archive}"/>
<get src="${get.title.metadata}" dest="${archive}/title_metadata.xml" verbose="on"/>
</target>
<target name="archive-get-list-of-transcripts">
<mkdir dir="${wylie}"/>
<xslt in="${archive}/title_metadata.xml" out="${archive}/titles_as_list.txt" style="${styles}/get-list-of-transcripts.xsl"/>
</target>
<target name="archive-get-transcripts">
<loadfile property="transcript-list" srcFile="${archive}/titles_as_list.txt"/> <!-- encoding="UTF-8" -->
<foreach list="${transcript-list}" delimiter=" " param="filename" target="archive-get-one-transcript"/>
</target>
<target name="archive-get-one-transcript">
<get src="${url.to.transcripts}/${filename}" dest="${wylie}/${filename}" ignoreerrors="true"/>
</target>
<target name="archive-transcripts-to-unicode">
<mkdir dir="${unicode}"/>
<antcall target="jskad-dist">
<param name="my.jar.suffix" value=""/>
</antcall>
<!-- why not use ant task for saxon?? -->
<java classname="net.sf.saxon.Transform" fork="yes">
<arg value="-o"/>
<arg value="${unicode}"/>
<arg value="${wylie}"/>
<arg value="${styles}/qdToUnicode.xsl"/>
<classpath>
<pathelement location="${vanillalib}/Jskad.jar"/>
<path refid="saxon.classpath"/>
</classpath>
</java>
</target>
<!--
<target name="archive-transcripts-to-unicode">
@ -58,44 +103,6 @@
</java>
</target>
-->
<!-- archive tasks -->
<!-- title_metadata.xml is not being retrieved as UTF-8. need to fix this.
perhaps we don't need to decodeUTF the transcript file name???
-->
<target name="archive-get-metadata">
<mkdir dir="${archive}"/>
<get src="${get.title.metadata}" dest="${archive}/title_metadata.xml" verbose="on"/>
</target>
<target name="archive-get-list-of-transcripts">
<mkdir dir="${wylie}"/>
<xslt in="${archive}/title_metadata.xml" out="${archive}/titles_as_list.txt" style="${styles}/get-list-of-transcripts.xsl"/>
</target>
<!-- must delete following transcripts from titles_as_list.txt
02069_clip-21-husked-barle_00.xml
02116_nasal-congestion_00.xml
A_New_Script02.xml
00007_06-dawa-and-purdrön_08.xml
or, should put in protection against URLs that cannot be "got"
-->
<target name="archive-get-transcripts">
<loadfile property="transcript-list" srcFile="${archive}/titles_as_list.txt"/>
<!--<loadfile property="transcript-list" srcFile="${archive}/titles_as_list.txt" encoding="UTF-8"/>-->
<foreach list="${transcript-list}" delimiter=" " param="filename" target="archive-get-one-transcript"/>
</target>
<!-- problem: destination filename should not be URLEncoded -->
<target name="archive-get-one-transcript">
<!--<urlencode property="file.location" value="00007_06-dawa-and-purdrön_08.xml" />
<get src="${url.to.transcripts}/${file.location}" dest="${wylie}/TEST.xml"/>-->
<get src="${url.to.transcripts}/${filename}" dest="${wylie}/${filename}"/>
</target>
<!-- concordance program -->
<target name="lucene-thdl-compile" depends="init">
<mkdir dir="${lucene-thdl.bin}"/>