add get and parse uniprot results

git-svn-id: file:///srv/svn/function2gene/trunk@33 a0738b58-4706-0410-8799-fb830574a030

diff --git a/bin/get_uniprot_results b/bin/get_uniprot_results
new file mode 100755 (executable)
index 0000000..680e850
--- /dev/null
+++ b/bin/get_uniprot_results
@@ -0,0 +1,147 @@
+#! /usr/bin/perl
+
+# get_uniprot_results retreives files of search results from uniprot,
+# and is released under the terms of the GPL version 2, or any later
+# version, at your option. See the file README and COPYING for more
+# information.
+
+# Copyright 2008 by Don Armstrong <don@donarmstrong.com>.
+
+
+use warnings;
+use strict;
+
+
+use Getopt::Long;
+use Pod::Usage;
+
+=head1 NAME
+
+  get_uniprot_results [options]
+
+=head1 SYNOPSIS
+
+
+ Options:
+  --dir, -D directory to stick results into [default .]
+  --name, -n file naming scheme [default ${search}_results.$format]
+  --terms, -t file of search terms [default -]
+  --debug, -d debugging level [default 0]
+  --help, -h display this help
+  --man, -m display manual
+
+=head1 OPTIONS
+
+=over
+
+=item B<--debug, -d>
+
+Debug verbosity. (Default 0)
+
+=item B<--help, -h>
+
+Display brief useage information.
+
+=item B<--man, -m>
+
+Display this manual.
+
+=back
+
+=head1 EXAMPLES
+
+  get_uniprot_results -D ./uniprot_results/ -n '${search}_name.html' < search_parameters
+
+Will pretty much do what you want
+
+=cut
+
+
+
+use vars qw($DEBUG $REVISION);
+
+BEGIN{
+     ($REVISION) = q$LastChangedRevision: 1$ =~ /LastChangedRevision:\s+([^\s]+)/;
+     $DEBUG = 0 unless defined $DEBUG;
+}
+
+use IO::File;
+use URI;
+use WWW::Mechanize;
+use Time::HiRes qw(usleep);
+
+# XXX parse config file
+
+my %options = (debug    => 0,
+              help     => 0,
+              man      => 0,
+              dir      => '.',
+              name     => '${search}_results_uniprot',
+              terms    => '-',
+              uniprot_site => 'http://ca.expasy.org',
+              uniprot_search_url  => '/cgi-bin/sprot-search-ful?S=on&T=on',
+             );
+
+GetOptions(\%options,'name|n=s',
+          'terms|t=s','dir|D=s','debug|d+','help|h|?','man|m');
+
+pod2usage() if $options{help};
+pod2usage({verbose=>2}) if $options{man};
+
+$DEBUG = $options{debug};
+
+if (not -d $options{dir}) {
+     die "$options{dir} does not exist or is not a directory";
+}
+
+#open search terms file
+my $terms;
+if ($options{terms} eq '-') {
+     $terms = \*STDIN;
+}
+else {
+     $terms = new IO::File $options{terms}, 'r' or die "Unable to open file $options{terms}: $!";
+}
+
+#For every term
+while (<$terms>) {
+     # Get uids to retrieve
+     chomp;
+     s/\r$//g;
+     my $search = $_;
+     my $dir_name = eval qq("$options{name}") or die $@;
+     if (not -d "$options{dir}/$dir_name") {
+         mkdir("$options{dir}/$dir_name") or die "Unable to make directory $options{dir}/$dir_name $!";
+     }
+     my $uri = URI->new($options{uniprot_site}.$options{uniprot_search_url});
+     $uri->query_form($uri->query_form(),
+                     SEARCH => $search,
+                    );
+     my $url = $uri->as_string;
+     my $mech = WWW::Mechanize->new(agent=>"DA_get_uniprot_results/$REVISION");
+     print "retreived $url\n";
+     $mech->get($url);
+     my @links = $mech->find_all_links(text_regex=>qr/_HUMAN/);
+     for my $gene_link (@links) {
+         my $gene_url = $gene_link->url_abs->as_string;
+         print STDERR "saving $gene_url\n";
+         $mech->get($gene_url);
+         $mech->follow_link(text_regex => qr/View\s*entry\s*in\s*raw\s*text\s*format/);
+         my $cleaned_url = $gene_url;
+         $cleaned_url =~ s{http://}{}g;
+         $cleaned_url =~ s/[^\w]//g;
+         eval {
+              $mech->save_content($options{dir}.'/'.$dir_name.'/'.$cleaned_url);
+         };
+         if ($@) {
+              warn $@;
+         }
+     }
+}
+
+
+
+
+
+
+__END__

diff --git a/bin/parse_uniprot_results b/bin/parse_uniprot_results
new file mode 100755 (executable)
index 0000000..6b90eb7
--- /dev/null
+++ b/bin/parse_uniprot_results
@@ -0,0 +1,214 @@
+#! /usr/bin/perl
+
+# parse_uniprot_results retreives files of search results from ncbi,
+# and is released under the terms of the GPL version 2, or any later
+# version, at your option. See the file README and COPYING for more
+# information.
+
+# Copyright 2004 by Don Armstrong <don@donarmstrong.com>.
+
+# $Id: ss,v 1.1 2004/06/29 05:26:35 don Exp $
+
+
+use warnings;
+use strict;
+
+
+use Getopt::Long;
+use Pod::Usage;
+
+=head1 NAME
+
+  parse_uniprot_results [options]
+
+=head1 SYNOPSIS
+
+
+ Options:
+  --dir, -D directory to stick results into [default .]
+  --name, -n file naming scheme [default ${search}_results.$format]
+  --terms, -t file of search terms [default -]
+  --debug, -d debugging level [default 0]
+  --help, -h display this help
+  --man, -m display manual
+
+=head1 OPTIONS
+
+=over
+
+=item B<--debug, -d>
+
+Debug verbosity. (Default 0)
+
+=item B<--help, -h>
+
+Display brief useage information.
+
+=item B<--man, -m>
+
+Display this manual.
+
+=back
+
+=head1 EXAMPLES
+
+  parse_uniprot_results -D ./uniprot_results/ -n '${search}_name.html' < search_parameters
+
+Will pretty much do what you want
+
+=cut
+
+
+
+use vars qw($DEBUG $REVISION);
+
+BEGIN{
+     ($REVISION) = q$LastChangedRevision: 1$ =~ /LastChangedRevision:\s+([^\s]+)/;
+     $DEBUG = 0 unless defined $DEBUG;
+}
+
+use IO::File;
+use IO::Dir;
+
+use HTML::TreeBuilder;
+use HTML::ElementTable;
+
+my %options = (debug    => 0,
+              help     => 0,
+              man      => 0,
+              dir      => '.',
+              keyword  => undef,
+              keywords => 0,
+             );
+
+GetOptions(\%options,'keyword|k=s','dir|D=s','debug|d+','help|h|?','man|m',
+          'keywords',
+         );
+
+
+pod2usage() if $options{help};
+pod2usage({verbose=>2}) if $options{man};
+
+$DEBUG = $options{debug};
+
+# CSV columns
+use constant {NAME        => 0,
+             REFSEQ      => 1,
+             LOCATION    => 2,
+             ALIAS       => 3,
+             FUNCTION    => 4,
+             DESCRIPTION => 5,
+             KEYWORD     => 6,
+             DBNAME      => 7,
+             FILENAME    => 8,
+            };
+
+if ($options{keywords}) {
+     if (@ARGV != 1) {
+         pod2usage("If the --keywords option is used, exactly one argument (the keyword) must be passed");
+     }
+     $options{dir} = "$ARGV[0]_results_uniprot";
+}
+
+if (not -d $options{dir}) {
+     die "$options{dir} does not exist or is not a directory";
+}
+
+my $dir = new IO::Dir $options{dir} or die "Unable to open dir $options{dir}: $!";
+
+print join(",", map {qq("$_");} qw(Name RefSeq Location Alias Function Description Keyword DBName Filename)),qq(\n);
+
+my ($keyword) = $options{keyword} || $options{dir} =~ m#(?:^|/)([^\/]+)_results_uniprot#;
+
+FILE: while ($_ = $dir->read) {
+     my $file_name = $_;
+     next if $file_name =~ /^\./;
+     next unless -f "$options{dir}/$file_name" and -r "$options{dir}/$file_name";
+
+     my $file = new IO::File "$options{dir}/$file_name", 'r' or die "Unable to open file $file_name";
+
+     my @fields;
+     my $current_field = undef;
+ LINE: while (<$file>) {
+         my ($type,$data) = $_ =~ /^(\w{2})\ {3}(.+)/s;
+         $type ||= 'SQ';
+         next LINE if not defined $data;
+         if (not defined $current_field and $type ne 'ID') {
+              next FILE;
+         }
+         if (not defined $current_field or $current_field->{type} ne $type) {
+              if (defined $current_field) {
+                   push @fields,$current_field;
+              }
+              $current_field = {type => $type,
+                                data  => '',
+                               };
+         }
+         $current_field->{data} .= $data;
+     }
+     if (defined $current_field) {
+         push @fields,$current_field;
+     }
+
+     my @results;
+
+     for my $field (@fields) {
+         my $type = $field->{type};
+         if ($type eq 'GN') {
+              # Find gene name
+              ($results[NAME]) = $field->{data} =~ m{Name=(.+?);}xis;
+         }
+         elsif ($type eq 'DR') {
+              # Find REF SEQ number
+              ($results[REFSEQ]) = $field->{data} =~ m{RefSeq;\s*([^;]+)}xis;
+              if (not defined $results[REFSEQ]) {
+                   ($results[REFSEQ]) = $field->{data} =~ m{Ensembl;\s*([^;]+)}xis;
+              }
+         }
+         elsif ($type eq 'DE') {
+              # Find gene aliases; these are odd bits separated by ()
+              # and such.
+              my $alias = $field->{data};
+              $alias =~ s/\n/ /g;
+              my @aliases = split /(?:\)\s+\(|\s+\(|\)(?:\s+|\.\s*$))/, $alias;
+              $results[ALIAS] = join('; ',map {s/(\s)\s+/$1/g; $_;} @aliases);
+         }
+         elsif ($type eq 'CC') {
+              my ($function) = $field->{data} =~ m{-!-\s*FUNCTION:\s*(.+?)(?:-!-|$)}xs;
+              if (defined $function) {
+                   $function =~ s/\n//g;
+                   $function =~ s/(\s)\s+/$1/g;
+                   $results[FUNCTION] = $function;
+              }
+              my $description = $field->{data};
+              $description =~ s/\n/ /g;
+              $description =~ s/-!-//g;
+              $description =~ s/(\s)\s+/$1/g;
+              $description =~ s/-----{5,}.+$//;
+              $results[DESCRIPTION] = $description;
+         }
+     }
+
+     $results[NAME] ||= 'NO NAME';
+     $results[REFSEQ] ||= 'NO REFSEQ';
+
+     $results[LOCATION] ||= 'NO LOCATION';
+     $results[ALIAS] ||= 'NO ALIASES';
+     $results[FUNCTION] ||= 'NO FUNCTION';
+
+     # Figure out the keyword used
+     $results[KEYWORD] ||= $keyword || 'NO KEYWORD';
+
+     # Database searched
+     $results[DBNAME] = 'uniprot';
+     $results[FILENAME] = $file_name;
+
+     print join(',',map {qq("$_")} @results),qq(\n);
+}
+
+
+
+
+
+
+__END__