package PubMed::Index;

use strict;
use warnings;

use Carp qw(confess);
use Data::Dumper;
use File::Compare;
use XML::LibXML;

#use PubMed::Api;
use PubMed::Common;
use PubMed::Paths;
use PubMed::Search;
use PubMed::Strings;

my $barf_dir=$PubMed::Paths::barf_dir;

sub parse_stalefi_line {
  my $line=shift // confess "I need a line here.";
  my @comps=split(' ',$line);
  my $start=$comps[1];
  my $length=$comps[2];
  my $infi=$comps[3];
  return ($start,$length,$infi);
}

sub read_stalefi {
  my ($start,$length,$infi,$fhs)=@_;
  if(not defined($infi)) {
    confess "I need an infi.";
  }
  my $rawx_dir="$PubMed::Paths::rawx_dir";
  my $rawx_file="$rawx_dir/pubmed$infi.xml";
  if(not defined($fhs->{$infi})) {
    open($fhs->{$infi},"< $rawx_file")
      or confess "I can not open the file $rawx_file";
  }
  seek($fhs->{$infi},$start,0);
  my $text;
  read($fhs->{$infi},$text,$length);
  return $text;
}

sub get_latest_loc {
  my $line=shift;
  my @parts=split(' ',$line);
  my $loc=join(' ',@parts[0..3]);
  return $loc;
}

sub get_purd_by_pmid {
  my $pmid=shift;
  my $loc=&check_purd_by_pmid($pmid) or return '';
  if($loc=~m|^\d+$|) {
    confess "loc $loc is a pmid.";
  }
  my @stalefi=&PubMed::Index::parse_stalefi_line($loc);
  if(not scalar(@stalefi) == 3) {
    die Dumper @stalefi;
  }
  return &PubMed::Index::read_stalefi(@stalefi);
}

sub check_purd_by_pmid {
  my $pmid=shift // confess "I need a pmid here.";
  my $pemi=&PubMed::Common::pmid_to_pemi($pmid);
  my $pemi_file="$barf_dir/$pemi.sox";
  if(not -f $pemi_file) {
    confess "I don't see $pemi_file";
  }
  #my $s="$PubMed::Paths::look -b $pmid $pemi_file";
  #my $loc=`$s`;
  #chomp $loc;
  my $loc=&PubMed::Search::search($pemi_file,$pmid);
  if(not $loc) {
    return '';
  }
  return $loc;
}

#sub get_purd_by_line {
#  my $line=shift;
  ## fapi case
  #if($line=~m|^(\d+)$|) {
  #  my $pmid=$1;
  #  return &PubMed::Api::get_purd_by_pmid($pmid) // '';
  #}
  #if($line=~m-^\d+ \d+ \d+ \d{2}n\d{4}(\s|$)-) {
  #  my @stalefi=&PubMed::Index::parse_stalefi_line($line);
  #  return &PubMed::Index::read_stalefi(@stalefi);
  #}
  ## in the case of pmid + number, seed 2016-01-08
  #if($line=~m|^(\d+) |) {
  #  my $pmid=$1;
  #  warn "You gave me a bad line '$line', but I try pmid $pmid";
  #  return &PubMed::Index::get_purd_by_pmid($pmid) // '';
  #}
  #confess "You gave me a very bad line '$line'";
#}


sub get_pmid_from_loc {
  my $loc=shift;
  my $index=index($loc,' ');
  my $pmid=substr($loc,0,$index) // '';
  if(not $pmid) {
    confess "I can't get the pmid from '$loc'";
  }
  return $pmid;
}

sub get_earliest_infi {
  my $loc=shift;
  my @parts=split(/ /,$loc);
  my $infi=$parts[$#parts];
  if(not $infi=~m|^\d{2}n\d{4}$|) {
    confess "I found a bad infi $infi";
  }
  return $infi;
}



1;