#!/bin/sh -
# This script will send a file to netscape with database IDs linked to SRS.
# To customize this script edit the $DATABASES and $SRS_SERVER variable in the
# perl code below and the NETSCAPE variable at the top of the script.
# To enable this in Artemis, sanger_options must be set to true in the options
# file.
# Default browser is mozilla
# If you are under KDE, you should use Konqueror then
# If you are under MAC OS X, use Safari
if [ `uname` = Darwin ]; then
NETSCAPE=/usr/bin/open
else
for f in /usr/bin/x-www-browser \
/usr/bin/X11/real-netscape \
/usr/bin/firefox \
/usr/bin/iceweasel \
/usr/bin/mozilla; do
if [ -f $f ]; then
NETSCAPE=$f
break
fi
done
fi
if [ ! -f "$NETSCAPE" ]; then
NETSCAPE=`which firefox`
fi
if [ -f "$DIANA_ENVIRONMENT_FILE" ]
then
. $DIANA_ENVIRONMENT_FILE
fi
if [ $# = 0 ]
then
echo no argument given 1>&2
exit 1
fi
file_arg=$1
unique_bit=$$.`hostname`
# sanger hack:
file_arg=`echo $file_arg | sed 's@^/tmp_mnt/nfs/@/nfs/@' | sed 's@^/tmp_mnt/tmp_nfs/@/nfs/@'`
# fix for pcs3:
file_arg=`echo $file_arg | sed 's@^/yeastpub4/@/nfs/disk222/yeastpub4/@'`
# fix for pcs4:
file_arg=`echo $file_arg | sed 's@^/.automount/pcs3/root/nfs/@/nfs/@'`
file_arg=`echo $file_arg | sed 's@^/.automount/evs-users2/root/@/nfs/@'`
if [ -f ./$file_arg ]
then
# the file is in the current directory - we need the full path so netscape
# can find the file
new_file=$PWD/$file_arg.$unique_bit.html
else
new_file=$file_arg.$unique_bit.html
fi
###
# sanger fix for /nfs/repository which isn't mounted on workstations:
REPOSITORY=`echo $new_file | sed -n -e 's|^\/nfs\/repository\/\(.*\)\(\/\)\(.*\)$|\3|p'`
if [ "$REPOSITORY" != "" ]; then
if [ ! -d $HOME/artemis_tmp ]; then
mkdir $HOME/artemis_tmp
fi
new_file="$HOME/artemis_tmp/$REPOSITORY"
fi
#
###
cat <<EOF > $new_file
<HTML>
<HEAD>
<TITLE>
Results for $file_arg
</TITLE>
</HEAD>
<BODY>
<PRE>
EOF
perl -e '
BEGIN {
# change these variable to list the databases to search for the IDs - the
# database names should be separated by spaces
$PROTEIN_DATABASES = "uniprot";
$DNA_DATABASES = "embl";
# change this to point to the wgetz script of your SRS server
#$SRS_SERVER = "www.sanger.ac.uk/srs6bin/cgi-bin/wgetz?-e+";
#$SRS_SERVER = "srs.ebi.ac.uk/srsbin/cgi-bin/wgetz?-e+";
$SRS_SERVER = "www.bioinformatics.nl/srsbin/cgi-bin/wgetz?-e+";
$PROTEIN_DATABASES =~ s/ /%20/g;
$DNA_DATABASES =~ s/ /%20/g;
%GENEDB_PATTERNS = (tryp => q!Tb\d+\.\d+\.\d+|TRYP_\S+!,
leish => q!LmjF\d+.\d+!);
$GENEDB_PATTERN = join ("|", values %GENEDB_PATTERNS);
$BLAST_START_LINE = "Sequences producing High-scoring Segment Pairs|" .
"Sequences producing significant alignments:";
$FASTA_START_LINE = "The best scores are";
# the list of IDs we have seen so far
@ids = ();
# the list of IDs we have made anchors for so far
@anchored_ids = ();
$db_type = "unknown";
}
sub hyperlink_to_anchor
{
$id = shift;
qq(<a href="#$id">$id</a>);
}
sub hyperlink_id
{
$id = shift;
if ($db_type eq "dna") {
$r = qq#<a href="http://$SRS_SERVER\[\{$DNA_DATABASES\}-ID:$id*]|[\{$DNA_DATABASES\}-AccNumber:$id*]">$id</a>#;
} else {
for my $org (keys %GENEDB_PATTERNS) {
my $pattern = $GENEDB_PATTERNS{$org};
if ($id =~ /$pattern/) {
$r = qq#<a href="http://www.genedb.org/genedb/Search?organism=$org&name=$id">$id</a>#;
return $r
}
}
# Text Entry
$r = qq#<a href="http://www.uniprot.org/uniprot/$id">$id</a>#;
#$r = qq#<a href="http://$SRS_SERVER-id+1+\[\{$PROTEIN_DATABASES\}-ID:$id*]|[\{$PROTEIN_DATABASES\}-AccNumber:$id*]+-vn+2">$id</a>#;
#$r = qq#<a href="http://$SRS_SERVER-id+1+\[\{$PROTEIN_DATABASES\}-ID:$id*]|[\{$PROTEIN_DATABASES\}-AccNumber:$id*]">$id</a>#;
}
return $r
}
$file_name = $ARGV[0];
if ($file_name =~ /\.gz$/) {
open IN_FILE, "gzip -d < $file_name |" or die "failed to open $file_name\n";
} else {
open IN_FILE, "$file_name" or die "failed to open $file_name\n";
}
while (<IN_FILE>) {
if ($. == 1) {
if (/^\s*([^\s]+)/) {
if (lc $1 eq "blastn" or lc $1 eq "tblastn" or lc $1 eq "tblastx") {
$db_type = "dna";
} else {
if (lc $1 eq "fasta" or lc $1 eq "blastp" or lc $1 eq "blastx") {
$db_type = "protein";
} else {
print "\nWARNING: could not identify file type: $1\n";
}
}
} else {
print "\nWARNING: could not identify file type\n";
}
}
# ignore header lines
if (1..m/$BLAST_START_LINE|$FASTA_START_LINE/) {
print;
next;
}
if (@ids && /^\s*$/) {
$summary_finished = 1;
}
if (/^>?\w+\|\w+\|(\w+)|^(?:(?:>?>?(?:[A-Z]+:)?)(\w+)|($GENEDB_PATTERN))|^(?:(?:>?>?(?:[A-Z]+:)?)(\w+)\\.\d+\s+) /) {
$id = $1;
if (!defined $id) {
$id = $2;
}
if ($summary_finished) {
if ((grep {$_ eq $id} @ids) && (!grep {$_ eq $id} @anchored_ids)) {
# not anchored yet so make it an anchor
if (s/\b$id\b/"<a name=\"$id\">" . (hyperlink_id($id)) . "<\/a>"/e) {
push @anchored_ids, $id;
}
}
} else {
if (!grep {$_ eq $id} @ids) {
push @ids, $id;
}
s/$id/hyperlink_to_anchor($id)/ei;
# if (!s/ $id/" " . hyperlink_id($id)/ei) {
# if the id occurs once in the line put a link at end of line
s/$/" LINK:" . hyperlink_id($id)/e;
# }
}
}
print;
}
' $file_arg >> $new_file
cat <<EOF >> $new_file;
</PRE>
</BODY>
</HTML>
EOF
# delete it at some point
echo "rm -f $new_file > /dev/null 2>&1" | at now + 12 hours
# MAC OS X => Safari browser
# The command then is the following
if [ -f "/usr/bin/open" -a `uname` = Darwin ]
then
/usr/bin/open $new_file
exit 0
elif [ -f "/Applications/Safari.app/Contents/MacOS/Safari" ]
then
if $NETSCAPE $new_file
then
exit 0
fi
# For Netscape or mozilla
# Use openURL($new_file, new-tab) to get new-tab
elif $NETSCAPE -remote "openURL($new_file)"
then
sleep 1m
rm -f $new_file
exit 0
else
echo starting new netscape 2>&1
# netscape isn't running - so start it
($NETSCAPE &)&
# now send the URL. we do things this way so that the script doesn't exit
# until netscape has successfully shown the URL
sleep 1
# don't exit the script until the file is successfully displayed or until
# 40 seconds is up
for i in 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
do
if $NETSCAPE -remote "openURL($new_file)" 2> /dev/null
then
sleep 1m
rm -f $new_file
exit 0
else
sleep 2
fi
done
exit 1
fi