RE: Stylesheet Optimization -- How to Make It Faster

(a) It would be a nice courtesy if you could lay out the code so that we can
read it.

(b) What XSLT processor are you using?

(c) The most obvious inefficiency is here:
    expand="{$abbreviations[.=$abbr]/following-sibling::expanded}"
    This would benefit from use of keys.

Michael Kay
http://www.saxonica.com/


> -----Original Message-----
> From: Jeff Sese [mailto:jsese@xxxxxxxxxxxx]
> Sent: 28 November 2006 01:41
> To: Xsl-List
> Subject:  Stylesheet Optimization -- How to Make It Faster
>
> I have a stylesheet that puts mark-up to text nodes that
> matches an abbreviation in a reference xml file. Its working
> nicely but the processing time is very slow... i'm guessing
> because its processing text nodes. A 800kb file takes me
> about 25 mins to process and i have around 800 file to
> process (varying file sizes, some are relatively small and
> some are fairly large). Is there any way to optimize my
> stylesheet so that it can process the files faster?
>
> here is my stylesheet:
>
> <?xml version="1.0" encoding="UTF-8"?>
> <xsl:stylesheet version="2.0"
> xmlns:xsl="http://www.w3.org/1999/XSL/Transform"
> xmlns:xs="http://www.w3.org/2001/XMLSchema"
> xmlns:ati="http://www.asiatype.com/xslt-functions"
> exclude-result-prefixes="xs ati">
> <xsl:output method="xml" version="1.0" encoding="UTF-8"/>
> <xsl:variable name="abbreviations" as="element()+"
> select="document('publishers_data.xml')/root/publisher/abbrev"/>
> <xsl:template match="/">
> <xsl:apply-templates/>
> </xsl:template>
> <xsl:template match="text()[ancestor::ab and
> not(ancestor::note[@id and @n and @lang])]"> <xsl:variable
> name="str" as="xs:string" select="."/> <xsl:choose> <xsl:when
> test="exists($abbreviations[matches($str,concat('(^|\W)(',ati:
escape(.),')($|\W)'))])">
> <xsl:variable name="search-str" as="xs:string+"
> select="$abbreviations[matches($str,concat('(^|\W)(',ati:escap
e(.),')($|\W)'))]"/>
> <xsl:variable name="replace" as="element()*"> <xsl:for-each
> select="$search-str"> <xsl:variable name="abbr"
> as="xs:string" select="."/> <abbr type="title"
> expand="{$abbreviations[.=$abbr]/following-sibling::expanded}"
><xsl:value-of
> select="$abbr"/></abbr>
> </xsl:for-each>
> </xsl:variable>
> <xsl:sequence select="ati:replace-with-nodes($str,
> $search-str, $replace)"/> </xsl:when> <xsl:otherwise>
> <xsl:value-of select="$str"/> </xsl:otherwise> </xsl:choose>
> </xsl:template> <xsl:template
> match="@*|element()|comment()|processing-instruction()"
> mode="#all">
> <xsl:copy>
> <xsl:apply-templates select="@*|node()"/> </xsl:copy>
> </xsl:template> <xsl:function name="ati:replace-with-nodes"
> as="node()+"> <xsl:param name="input" as="xs:string"/>
> <xsl:param name="words-to-replace" as="xs:string*"/>
> <xsl:param name="replacement" as="node()*"/> <xsl:variable
> name="regex" select="string-join(for $w in $words-to-replace
> return concat('(', ati:escape($w), ')'),'|')"/>
> <xsl:analyze-string select="$input" regex="{$regex}">
> <xsl:matching-substring> <xsl:variable name="i"
> as="xs:integer" select="(1 to
> count($words-to-replace))[regex-group(.)]"/>
> <xsl:sequence select="$replacement[$i]"/>
> </xsl:matching-substring> <xsl:non-matching-substring>
> <xsl:value-of select="."/> </xsl:non-matching-substring>
> </xsl:analyze-string> </xsl:function> <xsl:function
> name="ati:escape"> <xsl:param name="s" as="xs:string"/>
> <xsl:sequence
> select="replace($s,'[\\\|\.\-\^\?\*\+\(\)\{\}\[\]\$]','\\$0')"/>
> </xsl:function>
> </xsl:stylesheet>
>
> heres a short version of the publishers_data.xml:
>
> <root>
> <publisher>
> <abbrev>Inschriften von Priene</abbrev>
> <expanded>Inschriften von Priene</expanded> </publisher>
> <publisher> <abbrev>P. Mil. Congr. XVIII</abbrev>
> <expanded>Papiri documentari dell'UniversitC Cattolica di
> Milano</expanded> </publisher> <publisher> <abbrev>P. Jud.
> Des. Misc.</abbrev> <expanded>Discoveries in the Judean
> Desert XXXVIII</expanded> </publisher>
> <!-- more publishers here -->
> </root>
>
> heres a snippet of the source xml:
>
> <!-- preceding::node() of ab -->
> <ab lang="grk" n="1">
> <foreign lang="grk">N N3N-N3N?N=N5 N:N1Oa=0 ON?a=:O
NN1O
N5a=7N?O</foreign>
> <note place="margin">a c</note> <lb n="5"/> <foreign
> lang="grk">OO
a=9N=N?OO ON?a?& N<N5Oa=0 NN1N<N2a=;ON7N=
N2N1ON9N;N5a=;ON1N=ON?O, a=ON5 N:N1a=6
> NN9N?N=a=;ON9N?O a<&N= a=
 NN9N;a=5ON9N?O</foreign> <lb/>(III),
<foreign
> lang="grk">a<Oa=6 Oa?O N>B/N5B/ a=N;ON<ON9a=1N4N?O</foreign>
(520/16)<foreign
> lang="grk">N a<1OON?O
N9N?N3O
a=1ON?O. a?>NO
a=9N4N?ON?O N4a=2 a=

a?>NN;N9-</foreign>
> <note place="margin">v</note> <lb/> <foreign
> lang="grk">N:N1O
N=N1ON5a=:O a= ON-N;N7ON1N9 ON?a=;ON?O,
N=N5a==ON5O
N?O a=$N=. N:N1a=6 a<&N=
> a<N:N?OOOa=4O N O
O	ON1N3a=9O
N?O</foreign> <note id="n7" n="7"
lang="ger">
> <foreign lang="grk">a=$N=N N3N-N3N?N=N5 N3a=0O
 N<N5Oa>=
N1a=Oa=9N=</foreign>
> A</note> <lb/> <foreign lang="grk">a=
 a?>NN:N1ON1a?N?O. OO
a?6ON?O
N4a=2
> a<1OON?O
a=7N1N= ON5N6a?6O a<N>a=5N=N5N3N:N5, OON3N3O
N1Oa=4N= N4a=2
N&N5O
N5N:a=;N4N7O</foreign>
> <note id="n8b9" n="8b9" lang="ger"> <foreign
> lang="grk">OO
a?6ON?ObN=N?N8N5a=;N5ON1N9</foreign> wiederholt s.
<foreign
> lang="grk">a=6OON?O
a?ON1N9</foreign>, s. <foreign
> lang="grk">OON3N3O
N1ON5a?O</foreign>.</note>
> <lb/>(I 3). <foreign lang="grk">Oa=0 N3a=0O

a>=NN:N?OON9N;a=1N?O</foreign>
> (<link type="boj" targets="a002" n="BOJTEXT002_T_7">2 T
> 7</link>) <foreign lang="grk">N=N?N8N5a=;N5ON1N9.</foreign> <note
> id="n9" n="9" lang="ger"> <foreign
> lang="grk">a>=NN:N?OON9N;a=1N?O</foreign> Vossius <foreign
> lang="grk">a>=NN3N7ON9N;a=1N?O</foreign> Suid</note> </ab>
> <!-- following::node() of ab -->
>
> all: ab nodes appear in the same level (same depth) though out.
>
> Any suggestions are welcome.
>
> Thanks,
> --
> Jeff