Stahování Zákony pro lidi

schema.xml

<!-- rozšíření schématu solru o specifické pole - 
     pro funkčnost není nutné, ale pokud se nepoužije, je třeba odpovídajícím
     způsobem upravit šablonu
-->

<!-- Zakony pro lidi -->
<field name="law_doctype" type="string" indexed="true" stored="true" multiValued="true"/>

zakony-pro-lidi-indexer.sh

#!/bin/bash

apikey=test
collection=cs

cd harvest
mkdir zakony

#výchozí začátek je rok 1945, pokud už se stahovalo, naváže posledním staženým rokem...
start=1945
if [ -f zakony-date ]
then
        start=$(date -r zakony-date +'%Y')
fi
end=$(date +'%Y')

# časová značka posledního stahování
touch zakony-date

cd zakony
rm *.xml

for (( i=$start; i<=$end; i++ ))
do
        wget "http://www.zakonyprolidi.cz/api/v1/data.xml/YearDocList?apikey=$apikey&Collection=$collection&Year=$i" -O "zakony$i.xml"
        echo "Rozděluji soubor pro rok $i"
        xml_split -l 2 zakony$i.xml
        rm "zakony$i.xml"
        rm "zakony$i-00.xml"
done

echo "Mažu zrušené zákony"
for xmlfile in $(ls)
do
        zruseno=$(tail -n 1 $xmlfile | sed 's/.*\(EffectTill\).*/1/')
        if [ "$zruseno" = "1" ] ; then
                rm $xmlfile
        fi;
done;


cd ..

./batch-import-xsl.sh zakony zakony.properties

cd ..

zakony.properties

; local/import/zakony.properties
; předpis pro importér

; XSLT Import Settings for Directory of Open Access Journals' doajArticle format
; (see http://www.doaj.org/schemas/doajArticle.xsd for details)
[General]
; REQUIRED: Name of XSLT file to apply.  Path is relative to the import/xsl directory
; of the VuFind installation (or your local override directory).
xslt = zakony.xsl
; OPTIONAL: PHP function(s) to register for use within XSLT file.  You may repeat
; this line to register multiple PHP functions.
;php_function[] = str_replace
; OPTIONAL: PHP class filled with public static functions for use by the XSLT file.
; The class must live within a PHP namespace.  You may specify a fully-qualified
; name; if you do not include a namespace, the class will automatically be assumed
; to live in the \VuFind\XSLT\Import namespace.
custom_class[] = VuFind
; OPTIONAL: If true, all custom_class settings above will be passed to the XSLT with
; their namespaces stripped off; for example, \VuFind\XSLT\Import\VuFind would be
; treated as \VuFind in XSLT files.  This allows more compact syntax within XSLT
; files, but it can lead to name conflicts if used carelessly.  If set to false, you
; must use fully-qualified names in your XSLT.  The false setting is recommended, but
; the current default is "true" for compatibility with legacy configurations.
truncate_custom_class = true

; XSLT parameters -- any key/value pairs set here will be passed as parameters to
; the XSLT file, allowing local values to be set without modifying XSLT code.
[Parameters]
institution = "ZakonyProLidi"
collection = "Law"

zakony.xsl

<!-- local/import/xsl/zakony.xsl
      šablona pro transformaci xml dodaného ze zakonyprolidi.cz API
-->

<xsl:stylesheet version="1.0"
    xmlns:xsl="http://www.w3.org/1999/XSL/Transform"
    xmlns:php="http://php.net/xsl"
    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
    xmlns:xlink="http://www.w3.org/2001/XMLSchema-instance">
    <xsl:output method="xml" indent="yes" encoding="utf-8"/>
    <xsl:param name="institution">ZakonyProLidi</xsl:param>
    <xsl:param name="collection">Law</xsl:param>
    <xsl:template match="DocInfo">
        <add>
            <doc>
                <!-- ID -->
                <field name="id">zakony<xsl:value-of select="/DocInfo/@DocId"/></field>

                <!-- RECORDTYPE -->
                <field name="recordtype">law</field>

                <!-- FULLRECORD -->
                <field name="fullrecord">
                    <xsl:copy-of select="php:function('VuFind::xmlAsText', /DocInfo)"/>
                </field>

                <!-- ALLFIELDS -->
                <field name="allfields">
                    <xsl:value-of select="normalize-space(string(/DocInfo))"/>
                </field>

                <!-- INSTITUTION -->
                <field name="institution">
                    <xsl:value-of select="$institution" />
                </field>

                                <!-- BUILDING -->
                <field name="building">zakony</field>

                <!-- COLLECTION -->
                <field name="collection">
                    <xsl:value-of select="$collection" />
                </field>

                <!-- ITEMTYPE -->
                <field name="itemtype">LG</field>

                <!-- LANGUAGE -->
                <field name="language">Czech</field>

                <!-- FORMAT -->
                <field name="format">Legislative document</field>

                <!--  LAW DOCTYPE -->
                <field name="law_doctype">
                    <xsl:value-of select="/DocInfo/@DocType"/>
                </field>

                <!-- AUTHOR -->
                <field name="author">Česká Republika</field>

                <!-- TITLE -->
                <field name="title"><xsl:value-of select="/DocInfo/@Title"/>. <xsl:value-of select="/DocInfo/@Quote"/></field>
                <field name="title_short"><xsl:value-of select="/DocInfo/@Title[normalize-space()]"/>. <xsl:value-of select="/DocInfo/@Quote"/></field>
                <field name="title_full"><xsl:value-of select="/DocInfo/@Title[normalize-space()]"/>. <xsl:value-of select="/DocInfo/@Quote"/></field>
                <field name="title_sort"><xsl:value-of select="php:function('VuFind::stripArticles', string(/DocInfo/@Title[normalize-space()]))"/>. <xsl:value-of select="php:function('VuFind::stripArticles', string(/DocInfo/@Quote[normalize-space()]))"/></field>

                <!-- PUBLISHER -->
                <field name="publisher">Česká Republika</field>

                <!-- PUBLISHDATE -->
                <field name="publishDate">
                        <xsl:value-of select="/DocInfo/@Year"/>
                </field>
                <field name="publishDateSort">
                        <xsl:value-of select="/DocInfo/@Year"/>
                </field>

                <!-- URL -->
                <field name="url">http://www.zakonyprolidi.cz<xsl:value-of select="/DocInfo/@Href[normalize-space()]"/></field>
            </doc>
        </add>
    </xsl:template>
</xsl:stylesheet>