#!/usr/local/bin/perl -w
# esumprotgi.pl 
# from ncbi:epost.p; File Description: ePost/eSummary calling example

=item about
 
  collect gi from blastp table output, top 5? hits/query, 
  fetch esummary for all uniq gi's ; merge results w/ query ID, bitscore
  change to 
    1. collect all gi's from all blastp, 
    2. fetch gi summaries to table file,
    3. merge w/ blast results query, ..

set dp=dvir ; 
cat tdprotnear23/$dp-blnear23i-sno.nodup-ncbinr.0?.blastp | grep -v '^#' \
| sort -k1,1 -k12,12nr | perl esumprotgi.pl -blast stdin -nhit 3 \
> $dp-blnear23i-sno.nodup-ncbinr.esum

    
=item blastp2sumtable

 .. add this blastp merge, tophit collect, and merge with query ids and bitscore

  cat tdprotnear/dgri-blnear23i-sno.nodup-ncbinr.0?.blastp | grep -v '^#' |\
   sort -k1,1 -k12,12nr | perl -ne\
   '($g,@v)=split; $b=$v[-1]; if($g eq $lg){ $ng++; $p=($ng<5 and $b > 0.9 * $lb)?1:0;} \
   else { $ng=0; $p=1; }print if($p); ($lb,$lg)=($b,$g);' \
    > dgri-blnear23i-sno.nodup-ncbinr.blastp
  
  cat tdprotnear/dgri-blnear23i-sno.nodup-ncbinr.blastp | \
   perl -ne'print "$1\n" if m/\sgi.(\d+)/;' | sort | uniq\
    > tdprotnear/dgri-blnear23i-sno.nodup-ncbinr.gi

 
=cut

# ---------------------------------------------------------------------------
my $eutils_root  = "http://www.ncbi.nlm.nih.gov/entrez/eutils";
my $ePost_url    = "$eutils_root/epost.fcgi";
my $eSummary_url = "$eutils_root/esummary.fcgi";

my $db_name = "protein"; # fixme option; "PubMed";

# ---------------------------------------------------------------------------
use strict;

use LWP::UserAgent;
use LWP::Simple;
use HTTP::Request;
use HTTP::Headers;
use CGI;
use Getopt::Long;

sub debug { warn @_ if(0); }

my($gilist,$blasttable);
my $nhits= 5;

my $optok= GetOptions(
  "blast=s", \$blasttable, 
  "gilist=s", \$gilist, 
  "nhits=i", \$nhits, 
  "db_name=s", \$db_name, 
);

die "usage: esummary.pl -blast blasttable | -gi gilist 
 options: -nhits=$nhits ; -db protein|nucleotide [$db_name]
 e.g. 
 cat tdprotnear/dgri-blnear23i-sno.nodup-ncbinr.0?.blastp | grep -v '^#' | sort -k1,1 -k12,12nr \\
 perl esumprotgi.pl -blast stdin | sort -k1,1 -k3,3
" 
unless($optok);

# ---------------------------------------------------------------------------
# Read input file into variable $file
# File name - forst argument $ARGV[0]

my $gistring="";
my %blasttab;

if($blasttable) {

# cat tdprotnear/dgri-blnear23i-sno.nodup-ncbinr.0?.blastp | grep -v '^#' |\
#  sort -k1,1 -k12,12nr | perl -ne\
#  '($g,@v)=split; $b=$v[-1]; if($g eq $lg){ $ng++; $p=($ng<5 and $b > 0.9 * $lb)?1:0;} \
#  else { $ng=0; $p=1; }print if($p); ($lb,$lg)=($b,$g);' \
#   > dgri-blnear23i-sno.nodup-ncbinr.blastp
# 
# cat tdprotnear/dgri-blnear23i-sno.nodup-ncbinr.blastp | \
#  perl -ne'print "$1\n" if m/\sgi.(\d+)/;' | sort | uniq\
#   > tdprotnear/dgri-blnear23i-sno.nodup-ncbinr.gi
  
  my($inh, $stdin, $lqid,$lbits,$p,$ng)= ("") x 20;
  if($blasttable =~ /^stdin|-/) { $inh= *STDIN; $stdin=1; } 
  else { open(IN,$blasttable) or die "error for blast in: $blasttable"; $inh=*IN; }

  while(<$inh>) {
    next unless(/^\w/);
    my($qid,$sid,@v)=split; 
    my $bits=$v[-1]; 
    if($qid eq $lqid){ $ng++; $p=($ng < $nhits and $bits > 0.8 * $lbits)?1:0;} 
    else { $ng=0; $p=1; }
    if($p) {
      my($gi)= $sid =~ m/gi\|(\d+)/;
      $blasttab{$gi} .= join("\t",$qid,$bits,"hit$ng")."\n";#? dont need sid
      }
    ($lbits,$lqid)=($bits,$qid);
  }
  close($inh) unless($stdin); #? err  
  $gistring= join ",", sort keys %blasttab;
  
} else {
  $gilist= shift @ARGV unless($gilist);
  
  undef $/;  #for load whole file
  open IF, $gilist || die "Can't open for read: $!\n";
  $gistring = <IF>;
  close IF;
  debug "Loaded file: [$gistring]\n";
  # Prepare file - substitute all separators to comma
  $gistring =~ s/\s+/,/gs;
  # debug "Prepared file: [$gistring]\n";
}


#Create CGI param line
my $form_data = "db=$db_name&id=$gistring";

my $ngi= $gistring =~ tr/,/,/; $ngi++ if($gistring);
warn "ncbi esummary db=$db_name, ngi=$ngi\n";

my $result= callEutil($db_name,$form_data);

#------------ process result -------------------------
my %doc;
my %gidocs;
my @tags= qw(Gi Caption TaxId Title  ); # Extra < dont need
@doc{@tags}="";
 
foreach (split "\n", $result) {
  $doc{Id}=$1 if(m,<Id>(\w+)</Id>,);
  if(m,<Item,) {
  foreach my $k (@tags) { $doc{$k}=$1 if(m,<Item Name="$k"[^>]*>([^<]+),);}
  }
  if(m,</DocSum>,) { printdoc();  } # savedoc(); 
}

if(%blasttab) {
  print "#".join("\t",qw(query bits hit ncbigi acc taxid title)),"\n";
  foreach my $gi (sort keys %blasttab) {
    my $doc  = $gidocs{$gi} || "nodoc";
    my $blast= $blasttab{$gi} || "noblast";
    my @blast= split "\n", $blast;
    map{ print "$_\t$doc\n"; } @blast;
    ## qid, bits, hitnum, Gi, Caption, TaxId, Title
  }
}


sub callEutil {
  my($db_name,$form_data)=@_;

# Create HTTP request
  my $headers = new HTTP::Headers(
    Accept		=> "text/html, text/plain",
    Content_Type	=> "application/x-www-form-urlencoded"
  );
  
  my $request = new HTTP::Request("POST", $ePost_url, $headers );
  $request->content($form_data);
  
  # Create the user agent object
  my $ua = new LWP::UserAgent;
  $ua->agent("eugenes.org");
  
  # send file to ePost by HTTP
  my $response = $ua->request($request);
  debug "Responce status message: [" . $response->message . "]\n";
  debug "Responce content: [" .        $response->content . "]\n";
  
  # Parse response->content and extract QueryKey & WebEnv
  $response->content =~ 
    m|<QueryKey>(\d+)</QueryKey>.*<WebEnv>(\S+)</WebEnv>|s;
  my $QueryKey = $1;
  my $WebEnv   = $2;
  
  debug "\nEXTRACTED:\nQueryKey = $QueryKey;\nWebEnv = $WebEnv\n\n";
  
  # ---------------------------------------------------------------------------
  # Retrieve DocSum from eSummary by simple::get method and print it
  
  # debug "eSummary result: \n";
  my $result= get("$eSummary_url?db=$db_name&query_key=$QueryKey&WebEnv=$WebEnv");
  # debug  "]\n";
  # print $result;
  return $result;
}

sub savedoc {
  if($doc{Id}) { $gidocs{ $doc{Id} }= join("\t", @doc{@tags}); }
  @doc{@tags}="";
}

sub printdoc {
  if(%blasttab) { savedoc(); }
  else { print join("\t", @doc{@tags}),"\n" if(%doc); }
  @doc{@tags}="";
}



=item url example

curl 'http://eutils.ncbi.nlm.nih.gov/entrez/eutils/esummary.fcgi?db=protein&id=28800982,28628843'

perl ~/desktop/epost.pl test.gi | perl
^^ keep these fields as table: Gi/Id, Caption, Title, Extra, TaxId
.. see below

=cut


=item eSummaryResult

<eSummaryResult>

<DocSum>
<Id>28800982</Id>
<Item Name="Caption" Type="String">AAO47091</Item>
<Item Name="Title" Type="String">hemochromatosis [Homo sapiens]</Item>
<Item Name="Extra" Type="String">gi|28800982|gb|AAO47091.1|[28800982]</Item>
<Item Name="Gi" Type="Integer">28800982</Item>
<Item Name="CreateDate" Type="String">2003/03/03</Item>
<Item Name="UpdateDate" Type="String">2003/03/03</Item>
<Item Name="Flags" Type="Integer">0</Item>
<Item Name="TaxId" Type="Integer">9606</Item>
<Item Name="Length" Type="Integer">268</Item>
<Item Name="Status" Type="String">live</Item>
<Item Name="ReplacedBy" Type="String"></Item>
<Item Name="Comment" Type="String"><![CDATA[  ]]></Item>
</DocSum>


<DocSum>
<Id>28628843</Id>
<Item Name="Caption" Type="String">AAO49381</Item>
<Item Name="Title" Type="String">erythroid associated factor [Homo sapiens]</Item>
<Item Name="Extra" Type="String">gi|28628843|gb|AAO49381.1|AF485325_1[28628843]</Item>
<Item Name="Gi" Type="Integer">28628843</Item>
<Item Name="CreateDate" Type="String">2003/03/02</Item>
<Item Name="UpdateDate" Type="String">2003/03/02</Item>
<Item Name="Flags" Type="Integer">0</Item>
<Item Name="TaxId" Type="Integer">9606</Item>
<Item Name="Length" Type="Integer">102</Item>
<Item Name="Status" Type="String">live</Item>
<Item Name="ReplacedBy" Type="String"></Item>
<Item Name="Comment" Type="String"><![CDATA[  ]]></Item>
</DocSum>

</eSummaryResult>

=cut