#!/bin/bash
# grddler.sh - GRDDL Documents Naively
# Author: Sean B. Palmer, inamidst.com

if [[ -z $1 ]]
then echo "Usage: $0 uri > output.rdf"
     echo "Requires xsltproc, md5sum, python"
     exit
fi

# Check that we have the right tools
for PROG in xsltproc md5sum python
do if ! which $PROG &> /dev/null
   then echo "Requires '$PROG'" && exit 1
   fi
done

function urijoin() {
   # Join two URIs as though the former were the base of the latter
   python -c "print __import__('urlparse').urljoin('$1', '$2')"
}

function transforms() { 
   # Get GRDDL transformations from an XHTML file
   GETTER=http://inamidst.com/proj/grddl/getTransforms.xsl
   xsltproc $GETTER $1 2> /dev/null | \
      grep '^T' | cut -d' ' -f 2
}

# Apply the transformations found
for XSL in $(transforms $1)
do XSLURI=$(urijoin $1 $XSL)
   OUTPUTLET=$(echo $XSLURI | md5sum | cut -d' ' -f 1).$$.tmp
   echo Applying $XSLURI... >&2
   xsltproc $XSLURI $1 2> /dev/null > $OUTPUTLET
   TEMPS="$TEMPS $OUTPUTLET"
done

# Create a stylesheet for merging the outputlets
cat <<EOF > agg.$$.tmp
<xsl:stylesheet xmlns:xsl="http://www.w3.org/1999/XSL/Transform" version="1.0">
<xsl:template match="/">
<rdf:RDF xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#">
EOF

for FN in $TEMPS
do echo "<xsl:for-each select=\"document('$FN')/rdf:RDF/*\">" >> agg.$$.tmp
   echo '<xsl:copy-of select="."/>' >> agg.$$.tmp
   echo '</xsl:for-each>' >> agg.$$.tmp
done

echo '</rdf:RDF>' >> agg.$$.tmp
echo '</xsl:template>' >> agg.$$.tmp
echo '</xsl:stylesheet>' >> agg.$$.tmp

# Run the stylesheet using itself as dummy input
echo Merging the results... >&2
xsltproc agg.$$.tmp agg.$$.tmp

# Clean up the stylesheet
rm -f agg.$$.tmp

# Clean up the outputlets
for FN in $TEMPS
do rm -f $FN
done

echo 'Done!' >&2

# [EOF]