formex transform now at POC stage

staffanm · Jun 11, 2018 · d0dac2a · d0dac2a
1 parent 2d45762
commit d0dac2a
Show file tree

Hide file tree

Showing 2 changed files with 39 additions and 5 deletions.
diff --git a/ferenda/sources/legal/eu/eurlex.py b/ferenda/sources/legal/eu/eurlex.py
@@ -346,7 +346,6 @@ def render_xhtml_tree(self, doc):
         return doc.body
 
     def metadata_from_basefile(self, doc):
-        from pudb import set_trace; set_trace()
         desc = Describer(doc.meta, doc.uri)
         desc.rel(CDM.resource_legal_id_celex, Literal(doc.basefile))
         # the sixth letter in 
@@ -377,19 +376,31 @@ def parse(self, doc):
     def parse_formex(self, doc, source):
         parser = etree.XMLParser(remove_blank_text=True)
         sourcetree = etree.parse(source, parser).getroot()
-        from pudb import set_trace; set_trace()
         fp = self.resourceloader.openfp("xsl/formex.xsl")
         xslttree = etree.parse(fp, parser)
         transformer = etree.XSLT(xslttree)
         params = etree.XSLT
-        from pudb import set_trace; set_trace()
         resulttree = transformer(sourcetree,
                                  about=XSLT.strparam(doc.uri),
                                  rdftype=XSLT.strparam(str(doc.meta.value(URIRef(doc.uri), RDF.type))))
         return resulttree
         # re-parse to fix whitespace
         buffer = BytesIO(etree.tostring(resulttree, encoding="utf-8"))
         return etree.parse(buffer, parser)
+
+    def render_xhtml_validate(self, xhtmldoc):
+        def checknode(node):
+            if node.tag.split("}")[-1].isupper():
+                raise errors.InvalidTree("Node %s has not been properly transformed from Formex to XHTML" % node.tag)
+            for child in node:
+                if type(child).__name__ == "_Element":
+                    checknode(child)
+        try:
+            checknode(xhtmldoc.getroot())
+        except errors.InvalidTree as e:
+            return str(e)
+        return super(EURLex, self).render_xhtml_validate(xhtmldoc)
+
 
     def tabs(self):
         return []

diff --git a/ferenda/sources/legal/eu/res/xsl/formex.xsl b/ferenda/sources/legal/eu/res/xsl/formex.xsl
@@ -17,6 +17,7 @@ Note: this template expects Formex version 4 (http://formex.publications.europa.
   <xsl:param name="about"/>
   <xsl:param name="rdftype"/>
   <xsl:output indent="yes"/>
+
   <xsl:template match="/">
     <html xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
 	  xsi:schemaLocation="http://www.w3.org/1999/xhtml http://www.w3.org/MarkUp/SCHEMA/xhtml-rdfa-2.xsd"
@@ -94,6 +95,8 @@ Note: this template expects Formex version 4 (http://formex.publications.europa.
     <span class="oj-ref"><xsl:apply-templates/></span>
   </xsl:template>
 
+  <xsl:template match="QUOT.START|QUOT.END"><xsl:value-of disable-output-escaping="yes" select="concat('&amp;#x',@CODE,';')"/></xsl:template>
+
 
   <xsl:template match="CONSID">
     <!-- eg: sv: "skäl" -->
@@ -111,7 +114,13 @@ Note: this template expects Formex version 4 (http://formex.publications.europa.
   </xsl:template>
 
   <xsl:template match="DIVISION">
-    <div typeof="bibo:DocumentPart" about="$about#..." property="dcterms:title" content="{TITLE/TI/P/HT} // {TITLE/STI/P/HT}">
+    <div typeof="bibo:DocumentPart" about="{$about}#D{count(preceding-sibling::DIVISION)+1}" property="dcterms:title" content="{TITLE/TI/P/HT} // {TITLE/STI/P/HT}">
+      <xsl:apply-templates/>
+    </div>
+  </xsl:template>
+
+  <xsl:template match="DIVISION/DIVISION">
+    <div typeof="bibo:DocumentPart" about="{$about}#D{count(../preceding-sibling::DIVISION)+1}-{count(preceding-sibling::DIVISION)+1}" property="dcterms:title" content="{TITLE/TI/P/HT} // {TITLE/STI/P/HT}">
       <xsl:apply-templates/>
     </div>
   </xsl:template>
@@ -154,11 +163,25 @@ Note: this template expects Formex version 4 (http://formex.publications.europa.
     </ol>
   </xsl:template>
 
+  <xsl:template match="LIST[@TYPE='ARAB']">
+    <ol type='1'>
+      <xsl:apply-templates/>
+    </ol>
+  </xsl:template>
+
+  <xsl:template match="LIST[@TYPE='DASH']">
+    <ul class="dash">
+      <xsl:apply-templates/>
+    </ul>
+  </xsl:template>
+
   <xsl:template match="ITEM">
     <!-- fixme: find better rdf predicate -->
     <li property="rinfoex:punkt" content="{NP/NO.P}"><xsl:apply-templates select="NP/TXT"/></li>
   </xsl:template>
-
+
+  <xsl:template match="FT"><span class="ft"><xsl:apply-templates/></span></xsl:template>
+
   <xsl:template match="FINAL">
     <div class="final">
       <xsl:apply-templates/>