Merge pull request #25 from semorrison/master

comparediffs, a tool to download, scrub, and compare PDFs

.gitignore

@@ -13,3 +13,6 @@ pdfparanoia.egg-info/
 
 # ignore pdfs in the top-level dir
 /*.pdf
+
+# temporary pdfs in tests/diff/
+tests/diff/pdf

tests/diff/README.md

@@ -0,0 +1,17 @@
+`comparediffs` provides a tool which
+* downloads a PDF from two different sources,
+* runs `pdfparanoia` on both files, and
+* compares the outputs byte-for-byte.
+
+Typical usage is to first establish two `ssh` tunnels to hosts with access to the literature, e.g. via
+`ssh -D 1080 host1` and `ssh -D 1081 host2`. You can then invoke `comparediffs` via
+
+    ./comparediffs localhost:1080 localhost:1081 < urls
+
+where `urls` is a file containing one URL per line (e.g. the example file in this directory).
+
+`comparediffs` creates a subdirectory `pdf/`, in which is stores PDFs. It won't try to download the same PDF twice, so if you make changes to `pdfparanoia` you'll
+want to clean out some or all of this subdirectory.
+
+It's easy to see which PDFs `pdfparanoia` failed on afterwards, as it leaves copies of the scrubbed files with suffixes `.1.cleaned.pdf` and `.2.cleaned.pdf`.
+When `pdfparanoia` succeeds (or isn't even needed, because the downloaded files were identical), the scrubbed files are removed.

tests/diff/comparediffs

@@ -0,0 +1,69 @@
+#!/bin/bash
+if [ "$#" != "2" ]; then
+	echo "Please supply the addresses of two SOCKS proxies:"
+	echo "compare host1:port1 host2:port2"
+	exit 1
+fi
+
+proxy1="$1"
+proxy2="$2"
+
+curl -s --socks $proxy1 http://google.com | grep "The document has moved" > /dev/null
+if [ "$?" != "0" ]; then
+	echo "SOCKS proxy $proxy1 doesn't seem to be working."
+	exit 2
+fi
+curl -s --socks $proxy2 http://google.com | grep "The document has moved" > /dev/null
+if [ "$?" != "0" ]; then
+	echo "SOCKS proxy $proxy2 doesn't seem to be working."
+	exit 3
+fi
+
+echo "--------------------------------------------------------------------------------------------------------------------------------"
+echo "Comparing PDFs downloaded via '$proxy1' and '$proxy2'"
+echo "Please enter URLs, one per line."
+echo 
+echo "PDFs which still differ after pdfparanoia will be left in the pdf/ subdirectory, with suffixes .1.cleaned.pdf and .2.cleaned.pdf"
+echo "--------------------------------------------------------------------------------------------------------------------------------"
+echo
+
+mkdir -p pdf
+
+while read url; do
+	hash=`echo $url | openssl sha1 | cut -d ' ' -f 2`
+	echo "Retrieving $url, with hash $hash";
+	if [ ! -s pdf/$hash.1.pdf ]; then
+		curl --socks $proxy1 -L --cookie cookie.jar -A "Mozilla/5.0 (compatible; MSIE 9.0; Windows NT 6.1; Trident/5.0)" -H "Accept: text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8" $url > pdf/$hash.1.pdf
+	fi
+	if [ "`head -c 4 pdf/$hash.1.pdf`" == "%PDF" ]; then
+		if [ ! -s pdf/$hash.2.pdf ]; then
+			curl --socks $proxy2 -L --cookie cookie.jar -A "Mozilla/5.0 (compatible; MSIE 9.0; Windows NT 6.1; Trident/5.0)" -H "Accept: text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8" $url > pdf/$hash.2.pdf
+		fi
+		if [ "`head -c 4 pdf/$hash.2.pdf`" == "%PDF" ]; then
+			if cmp -s pdf/$hash.1.pdf pdf/$hash.2.pdf; then
+				echo "PDFs are identical already, no need to use pdfparanoia"
+			else
+				echo "PDFs differ, running pdfparanoia"
+				if [ ! -s pdf/$hash.1.cleaned.pdf ]; then
+					cat pdf/$hash.1.pdf | pdfparanoia > pdf/$hash.1.cleaned.pdf
+				fi
+				if [ ! -s pdf/$hash.2.cleaned.pdf ]; then
+					cat pdf/$hash.2.pdf | pdfparanoia > pdf/$hash.2.cleaned.pdf
+				fi
+				if cmp -s pdf/$hash.1.cleaned.pdf pdf/$hash.2.cleaned.pdf; then
+					echo "pdfparanoia successful scrubbed the PDFs"
+					rm pdf/$hash.*.cleaned.pdf
+				else
+					echo "pdfparanoia failed!"
+				fi
+			fi
+		else
+			echo "Download failed from source 2"
+			rm pdf/$hash.*.pdf
+		fi
+	else
+		echo "Download failed from source 1"
+		rm pdf/$hash.*.pdf
+	fi
+	echo
+done

tests/diff/urls

@@ -0,0 +1,20 @@
+http://annals.math.princeton.edu/wp-content/uploads/annals-v176-n2-p11-s.pdf
+http://www.ams.org/journals/mcom/2012-81-278/S0025-5718-2011-02542-1/S0025-5718-2011-02542-1.pdf
+http://www.ams.org/journals/bull/2010-47-04/S0273-0979-10-01296-6/S0273-0979-10-01296-6.pdf
+http://nyjm.albany.edu/j/2009/15-14p.pdf
+http://link.springer.com/content/pdf/10.1007/s00440-011-0397-9
+http://www.worldscientific.com/doi/pdf/10.1142/S2010326311500018
+http://www.sciencedirect.com/science?_ob=MiamiImageURL&_cid=272585&_user=994540&_pii=S0001870812001806&_check=y&_origin=article&_zone=toolbar&_coverDate=10-Sep-2012&view=c&originContentFamily=serial&wchp=dGLzVlt-zSkWb&md5=bfeb5e0619d45362640529aff02baeda&pid=1-s2.0-S0001870812001806-main.pdf
+http://imrn.oxfordjournals.org/content/2002/4/165.full.pdf
+http://www.tandfonline.com/doi/pdf/10.2991/jnmp.2006.13.3.2
+http://www.jstor.org/stable/pdfplus/121105?acceptTC=true
+http://muse.jhu.edu/journals/american_journal_of_mathematics/v118/118.5bernstein.pdf
+http://archive.numdam.org/article/CM_1980__41_2_245_0.pdf
+http://www.pnas.org/content/102/26/9099.full.pdf
+http://iopscience.iop.org/1126-6708/2004/06/043/pdf/1126-6708_2004_06_043.pdf
+http://scitation.aip.org/getpdf/servlet/GetPDFServlet?filetype=pdf&id=JMAPAQ000042000007002896000001&idtype=cvips&doi=10.1063/1.1372177&prog=normal
+http://msp.org/gt/2009/13-03/gt-2009-13-033s.pdf
+http://msp.org/apde/2012/5-2/apde-v5-n2-p07-s.pdf
+http://www.ems-ph.org/journals/show_pdf.php?issn=1661-7207&vol=5&iss=2&rank=6
+http://www.ems-ph.org/journals/show_pdf.php?issn=0010-2571&vol=88&iss=1&rank=3
+http://prd.aps.org/pdf/PRD/v13/i12/p3491_1