#!/usr/bin/perl
# geneinfo.perl
# usage: grep match scaffold.tandy.gff | perl -ne geneinfo.perl
## adjust for 1-exon genes? these are mostly parts/altexons inside good genes
## possibly score 1-exon if gene-span is largish

  chomp; m/info=([^;]+)/ or next; 
  $b=$1; @b=split",",$b; %b= map { split ":" }@b;
  next if $b{lowquality}>0;
  $nx= 2; # $b{nexons} == 1 ? 1 : 2;  # not good
  $pa= $b{pctalign} || 100; $goodaln= $pa>33;
  $tandemnew= $goodaln && $b{eqnear} >= $nx && $b{eqsame} < $nx; 
  $tandemknow= $goodaln &&  !$tandemnew && $b{eqnear} >= $nx; #? && $b{fullgenes}>0; 
  $duplnew= $goodaln && ($b{eqnear} >= $nx || $b{eqfar} >= $nx) && $b{eqsame} < $nx; 
  $duplknow= $goodaln && !$duplnew && ($b{eqnear} >= $nx || $b{eqfar} >= $nx ); #? && $b{fullgenes}>0; 
  ($bid)= m/bestids=([^;]+)/; 
  %bid= map { s/\[.*//; s/\d.*$//; $_,1; } split",",$bid; 
  delete $b{"geneids"}; delete $b{"methods"}; delete $b{"repeats"}; 
  @bid= sort keys %bid; 
  foreach $bid ("all",@bid) { 
    $bs{$bid}{n}++; map{ $bs{$bid}{$_}+= $b{$_};} keys %b; 
    $bs{$bid}{xtandemnew} += $tandemnew; $bs{$bid}{xtandemknow} += $tandemknow; 
    $bs{$bid}{zduplnew} += $duplnew; $bs{$bid}{zduplknow} += $duplknow; 
  } 
  END{ 
  @meth=sort keys %bs; 
  @f= grep { $_ ne "n" } sort keys %{$bs{all}}; 
  @fh= @f; foreach (@fh) { s/^(z|x)//; s/gene/gn/; s/pctalign/align/; s/tandem/tand/; s/know/at/; s/s$//; }
  print "# gene info, all-methods\nscores/method\n"; 
  print join"\t","predictor", "count",@fh,"\n"; 
  foreach $m (@meth) { $n=$bs{$m}{n}||1; printf "%-11s\t%3d",$m,$n;  
    foreach $k (@f) { 
      $v= $bs{$m}{$k}; 
      $v= ($k =~ m/repeat|dupl|tandem(new|kno)/) ? $v : sprintf("%.2f",$v/$n); 
      print"\t",$v; 
      } 
    print"\n"; 
    } 
  }