D-Net
/
dnet-hadoop
14
5
Fork 11
Code Issues 14 Pull Requests 10 Projects 2 Releases 2 Wiki Activity

initial version of the BASE Transformation rule

This commit is contained in:
Michele Artini 2024-03-05 14:18:57 +01:00
parent af58cd726e
commit 6500151c90
1 changed files with 39 additions and 50 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

89
dhp-workflows/dhp-aggregation/src/main/resources/eu/dnetlib/dhp/collection/plugin/base/tr/base-treansformation-rule.xml
View File

@ -39,45 +39,33 @@
<xsl:apply-templates select="//*[local-name() = 'header']" />
<!--
<!-- TO EVALUATE
./base_dc:authod_id
./base_dc:authod_id/base_dc:creator_id
./base_dc:authod_id/base_dc:creator_name
./base_dc:autoclasscode
./base_dc:autoclasscode/@type
./base_dc:classcode
./base_dc:classcode/@type
./base_dc:collection
./base_dc:collection/@opendoar_id
./base_dc:collection/@ror_id
./base_dc:collname
./base_dc:continent
./base_dc:country
./base_dc:doi
./base_dc:global_id
./base_dc:lang
./base_dc:link
./base_dc:oa
./base_dc:rightsnorm
./base_dc:typenorm
./base_dc:year
./dc:contributor
./dc:coverage
./dc:creator
./dc:date
./dc:description
./dc:format
./dc:identifier
./dc:language
./dc:publisher
./dc:relation
./dc:rights
./dc:source
./dc:subject
./dc:title
./dc:type
-->
<!-- NOT USED
./base_dc:collection/text()
./base_dc:collection/@ror_id
./base_dc:continent
./base_dc:country
./base_dc:year (I use dc:date)
./dc:coverage
-->
<metadata>
@ -166,6 +154,33 @@
<xsl:value-of select="vocabulary:clean( //dc:rights, 'dnet:access_modes')" />
</oaf:accessrights>
<xsl:for-each select="//base_dc:doi">
<oaf:identifier>
<xsl:attribute name="identifierType" select="'doi'" />
<xsl:value-of select="." />
</oaf:identifier>
</xsl:for-each>
<oaf:hostedBy>
<xsl:attribute name="name">
<xsl:value-of select="//base_dc:collname" />
</xsl:attribute>
<xsl:attribute name="id">
<xsl:value-of select="concat('opendoar____::', //base_dc:collection/@opendoar_id)" />
</xsl:attribute>
</oaf:hostedBy>
<oaf:collectedFrom>
<xsl:attribute name="name">
<xsl:value-of select="$varOfficialName" />
</xsl:attribute>
<xsl:attribute name="id">
<xsl:value-of select="$varDataSourceId" />
</xsl:attribute>
</oaf:collectedFrom>
<!-- TODO CONTINUE HERE -->
<dr:CobjCategory>
<xsl:variable name="varCobjCategory" select="vocabulary:clean( //dc:type, 'dnet:publication_resource')" />
<xsl:variable name="varSuperType" select="vocabulary:clean( $varCobjCategory, 'dnet:result_typologies')" />
@ -211,37 +226,11 @@
</oaf:fulltext>
</xsl:if>
<oaf:hostedBy>
<xsl:attribute name="name">
TODO
</xsl:attribute>
<xsl:attribute name="id">
<xsl:value-of select="concat('opendoar____::', //base_dc:collection/@opendoar_id)" />
</xsl:attribute>
</oaf:hostedBy>
<oaf:collectedFrom>
<xsl:attribute name="name">
<xsl:value-of select="$varOfficialName" />
</xsl:attribute>
<xsl:attribute name="id">
<xsl:value-of select="$varDataSourceId" />
</xsl:attribute>
</oaf:collectedFrom><!-- ID recognition incomplete -->
<xsl:variable name="varKnownFileEndings"
select="('.bmp', '.doc', '.docx', '.epub', '.flv', '.jpeg', '.jpg', '.m4v', '.mp4', '.mpg', '.odp', '.pdf', '.png', '.ppt', '.tiv', '.txt', '.xls', '.xlsx', '.zip')" /><!-- regular expression
for DOIs reduced here - letters like less-than and quotation marks don't work in matches, use identiferExtract when enabled -->
<xsl:variable name="varIdDoi"
select="distinct-values((//dc:identifier[starts-with(., '10.')][matches(., '(10[.][0-9]{4,}[^\s/&gt;]*/[^\s&gt;]+)')], //dc:identifier[starts-with(., 'http') and (contains(., '://dx.doi.org/10.') or contains(., '://doi.org/10.'))]/substring-after(., 'doi.org/'), //dc:identifier[starts-with(lower-case(.), 'doi:10.')]/substring-after(lower-case(.), 'doi:')))" />
<xsl:for-each select="$varIdDoi">
<oaf:identifier>
<xsl:attribute name="identifierType" select="'doi'" />
<xsl:value-of select="." />
</oaf:identifier>
</xsl:for-each>
<xsl:variable name="varIdHdl" select="distinct-values(//dc:identifier[starts-with(., 'http') and contains(., '://hdl.handle.net/')]/substring-after(., 'hdl.handle.net/'))" />
<xsl:for-each select="$varIdHdl">