This is the mail archive of the mailing list .

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

[docbook-apps] Internationalized index in XSL


automatic indexing which is available in the DocBook XSL stylesheets is working quite well for English, but for other languages it lacks several features.

From my own experience, following functionality is missing:

- allow to group accented letters like e, é, ë into the same group under letter "e"

- treat special letters (e.g. "ch") as one character and place them in correct position (e.g. between "h" and "i")

Attached customization layer is able to solve these two issues. If you want to try it, just import original fo/docbook.xsl into your customization layer and then include attached file. It was tested under Saxon, but should work in any processor which implements EXSLT function extension.

Current settings are suitable for Czech language, but you can easily modify them by editing content of l:letters element:

<l:letters lang="cs">
  <l i="1">A</l>
  <l i="1">a</l>
  <l i="1">Á</l>
  <l i="1">á</l>
  <l i="2">B</l>
  <l i="2">b</l>
  <l i="10">H</l>
  <l i="10">h</l>
  <l i="11">Ch</l>
  <l i="11">ch</l>
  <l i="11">cH</l>
  <l i="11">CH</l>
  <l i="12">I</l>
  <l i="12">i</l>

This snippet means that letters A, a, Á, á will be put under the same group in the index (they have same i attribute) and that letter "a" will be sorted before "b" which has i=2. Later elements define that "ch" will be placed between "h" and "i".

I will appreciate if you could:

- test solution and give me feedback
- send me <l:letter lang="xx"> table for languages you are using
- write me that this solution can't handle properly your language and write me what must be improved in order to fix it

After gathering your feedback I will add this file into standard DocBook XSL stylesheets.


  Jirka Kosek  	

<?xml version="1.0"?>
<!DOCTYPE xsl:stylesheet [

<!ENTITY lowercase "'abcdefghijklmnopqrstuvwxyz'">

<!ENTITY primary   'normalize-space(concat(primary/@sortas, primary[not(@sortas)]))'>
<!ENTITY secondary 'normalize-space(concat(secondary/@sortas, secondary[not(@sortas)]))'>
<!ENTITY tertiary  'normalize-space(concat(tertiary/@sortas, tertiary[not(@sortas)]))'>

<!ENTITY sep '" "'>
<!ENTITY scope 'count(ancestor::node()|$scope) = count(ancestor::node())'>
<xsl:stylesheet xmlns:xsl="";
                exclude-result-prefixes="func i l"

<!-- ********************************************************************
     $Id: autoidx-ng.xsl,v 1.25 2003/10/04 17:11:10 kosek Exp $

     This file is part of the DocBook XSL Stylesheet distribution.
     See ../README or for copyright
     and other information.

     ******************************************************************** -->

<!-- jkj: Temporal attempt to do better internacionalized indexing -->

<!-- Attribute i defines sort order in index and assigns differents letter to
     the same index group. In fact this table allows you to handle accented letters
     as non-accented if necessary and can handle even double character letters like
     "ch". If it accepted to normal stylesheets this table would g  -->
<l:letters lang="cs">
  <l i="-1"/>
  <l i="0">Symboly</l>
  <l i="1">A</l>
  <l i="1">a</l>
  <l i="1">Ã?</l>
  <l i="1">á</l>
  <l i="2">B</l>
  <l i="2">b</l>
  <l i="3">C</l>
  <l i="3">c</l>
  <l i="4">Ä?</l>
  <l i="4">Ä?</l>
  <l i="5">D</l>
  <l i="5">d</l>
  <l i="5">Ä?</l>
  <l i="5">Ä?</l>
  <l i="7">E</l>
  <l i="7">e</l>
  <l i="7">Ã?</l>
  <l i="7">é</l>
  <l i="7">Ä?</l>
  <l i="7">Ä?</l>
  <l i="7">Ã?</l>
  <l i="7">ë</l>
  <l i="8">F</l>
  <l i="8">f</l>
  <l i="9">G</l>
  <l i="9">g</l>
  <l i="10">H</l>
  <l i="10">h</l>
  <l i="11">Ch</l>
  <l i="11">ch</l>
  <l i="11">cH</l>
  <l i="11">CH</l>
  <l i="12">I</l>
  <l i="12">i</l>
  <l i="12">Ã?</l>
  <l i="12">í</l>
  <l i="13">J</l>
  <l i="13">j</l>
  <l i="14">K</l>
  <l i="14">k</l>
  <l i="15">L</l>
  <l i="15">l</l>
  <l i="16">M</l>
  <l i="16">m</l>
  <l i="17">N</l>
  <l i="17">n</l>
  <l i="17">Å?</l>
  <l i="17">Å?</l>
  <l i="19">O</l>
  <l i="19">o</l>
  <l i="19">Ã?</l>
  <l i="19">ó</l>
  <l i="19">Ã?</l>
  <l i="19">ö</l>
  <l i="20">P</l>
  <l i="20">p</l>
  <l i="21">Q</l>
  <l i="21">q</l>
  <l i="22">R</l>
  <l i="22">r</l>
  <l i="23">Å?</l>
  <l i="23">Å?</l>
  <l i="24">S</l>
  <l i="24">s</l>
  <l i="25">Å?</l>
  <l i="25">Å¡</l>
  <l i="26">T</l>
  <l i="26">t</l>
  <l i="26">Ť</l>
  <l i="26">Å¥</l>
  <l i="28">U</l>
  <l i="28">u</l>
  <l i="28">Ã?</l>
  <l i="28">ú</l>
  <l i="28">Å®</l>
  <l i="28">ů</l>
  <l i="28">Ã?</l>
  <l i="28">ü</l>
  <l i="29">V</l>
  <l i="29">v</l>
  <l i="30">W</l>
  <l i="30">w</l>
  <l i="31">X</l>
  <l i="31">x</l>
  <l i="32">Y</l>
  <l i="32">y</l>
  <l i="32">Ã?</l>
  <l i="32">ý</l>
  <l i="33">Z</l>
  <l i="33">z</l>
  <l i="34">Ž</l>
  <l i="34">ž</l>

<!-- Returns index group code for given term  -->
<func:function name="i:group-index">
  <xsl:param name="term"/>
  <xsl:variable name="long-letter-index" select="document('')/*/l:letters/l[. = substring($term,1,2)]/@i"/>
  <xsl:variable name="short-letter-index" select="document('')/*/l:letters/l[. = substring($term,1,1)]/@i"/>
  <xsl:variable name="letter-index">
      <xsl:when test="$long-letter-index">
        <xsl:value-of select="$long-letter-index"/>
      <xsl:when test="$short-letter-index">
        <xsl:value-of select="$short-letter-index"/>
  <func:result select="number($letter-index)"/>

<!-- Return index group letter for given group code -->
<func:function name="i:group-letter">
  <xsl:param name="index"/>
  <func:result select="document('')/*/l:letters/l[@i=$index][1]"/>

<xsl:key name="group-code"

<!-- Modified original code is using index group codes instead of just first letter 
     to gain better grouping -->
<xsl:template name="generate-index">
  <xsl:param name="scope" select="(ancestor::book|/)[last()]"/>

  <xsl:variable name="terms"
                                                i:group-index(&primary;))[&scope;][1]) = 1
                                    and not(@class = 'endofrange')]"/>

  <xsl:apply-templates select="$terms" mode="index-div">
    <xsl:with-param name="scope" select="$scope"/>
    <xsl:sort select="i:group-index(&primary;)" data-type="number"/>

<xsl:template match="indexterm" mode="index-div">
  <xsl:param name="scope" select="."/>

  <xsl:variable name="key"

  <xsl:if test="key('group-code', $key)[&scope;]
                [count(.|key('primary', &primary;)[&scope;][1]) = 1]">
      <xsl:call-template name="indexdiv.title">
        <xsl:with-param name="titlecontent">
          <xsl:value-of select="i:group-letter($key)"/>
        <xsl:apply-templates select="key('group-code', $key)[&scope;]
                                     [count(.|key('primary', &primary;)[&scope;][1])=1]"
          <xsl:sort select="translate(&primary;, &lowercase;, &uppercase;)"/>
          <xsl:with-param name="scope" select="$scope"/>


Attachment: smime.p7s
Description: S/MIME Cryptographic Signature

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]