#!/usr/bin/perl

use strict;

# Here we pull in our helper libraries
use FileHandle;     # This will let us keep the output up-to-date
use Census;
use Data::Dumper;

# Autoflush (output) for immediate feedback
$| = 1;


# ************************
# *** TSV Manuipuation ***
# ************************

# Given a fresh file, grab the first row which we will assume are our headers
sub getHeaders {
  my ($file) = @_;
  my $line;
  if($line = <$file>) {
    chomp $line;
    $line =~ tr/\n\r//d;
    $line =~ s/\t+$//;
    my (@header) = split(/\t/, $line);
    return @header;
  } else {
    die "Header not found!";
  }
}

# Given a file and our headers, grab the next line and split it into
# name->value pairs
sub getRowHash {
  my ($file, @headers) = @_;
  my $line;
  if($line = <$file>) {
    $line =~ tr/\n\r//d;
    $line =~ s/\t+$//;
    my %row;
    my (@entries) = split(/\t/, $line);
    for(my $i = 0; $i < $#headers + 1; $i++) {
      $row{$headers[$i]} = $entries[$i];
    }
    return %row;
  } else {
    print "X\n";
    # We didn't find a line, so we'll return nothingness
    return ();
  }
}

sub printRow {
  my ($file, @data) = @_;
  print $file (join "\t", @data) . "\n";
}


# ******************
# *** Main Logic ***
# ******************

sub main {
  my ($filename) = @_;
  $filename =~ s/\.dat$//; # Chop off .dat

  my %mem;
  my $geo;
  open($geo, "data/azgeo.csv");
  while(<$geo>) {
    my ($tract, $block_group, $block, $id) = split(/,/,$_);
    chomp $id;
    $mem{$tract+0}{$block_group+0}{$block+0} = $id;
    #print "$tract $block_group $block -> $mem{$tract}{$block_group}{$block}\n";
    print "$id\n"
  }

  my ($in, $cur, $out);

  open($in, "$filename.dat")
    or die "Error opening input: $!\n";

  if(-e ($filename . "_out.dat")) {
    system('mv ' . $filename . '_out.dat ' . $filename . '_old.dat');
    open($cur, $filename . "_old.dat")
      or die "Error opening old output: $!\n";
  }

  open($out, ">" . $filename . "_out.dat")
    or die "Error opening output: $!\n";

  $out->autoflush(1); # Make sure that as soon as we output something we save it

#  my (@newstuff) = qw( tract block_group block district );
  my (@newstuff) = qw( LOGRECNO );

  my $linenum = 0;
  my (@headers) = getHeaders($in);
  getHeaders($cur) if $cur;
  @headers = (@headers, @newstuff);
  printRow($out, @headers);

  my %row;
  while((%row) = getRowHash($in, @headers)) {
    $linenum++;

    if($cur) {
      my %cur_row = getRowHash($cur, @headers);
      undef $cur unless %cur_row;
      foreach (@newstuff) {
        $row{$_} = $cur_row{$_};
      }
    }
    
    #if($row{LOGRECNO}) {
    if(0) {
   # } && $row{district} && $row{block_group} && $row{block}) {
      print "$linenum - Found cached version, skipping.\n";
      my (@out) = map { $row{$_} } @headers;
      printRow($out, @out);
    } else {


      # This is what I want
   #   my (%census_data) = Census::get($street);
     my $grp = $row{block_group};
     my $tract = $row{tract};
     $tract *= 100;
      print "  --> $tract $grp $row{block}\n";
      my (%census_data) = (LOGRECNO => $mem{$tract}{$grp}{$row{block}});
      print "$linenum - " . join(',',%census_data) . "\n";
      %row = ( %row, %census_data );
      my (@out) = map { $row{$_} } @headers;
      printRow($out, @out);
      #print "Sleeping for 3 seconds...\r";
      #sleep 3;
    }

  }
}

main(@ARGV);

=pod

# Here we go through, check to see that everything lines up, and also try the
# rows that didn't work again (getting them from the census website)
sub verify_and_reget {
  my ($in, $out, $new_out, %row);
  open($in, "input.dat") or die "Error opening input file!\n";
  open($out, "output.csv");
  open($new_out, ">new_out.csv");

  my $linenum = 0;
  my (@headers) = getHeaders($in);
  while((%row) = getRowHash($in, @headers)) {
    last if(exists($row{''}));
    $linenum++;
    print "Working on line $linenum\r";
    print "\n" unless $linenum % 1000;
    my $street = join(',',$row{StreetNo},$row{StreetDir},
      $row{StreetName},$row{StreetType});

    my $outline = <$out>;
    if($outline =~ /^$street,(.*)$/) {
      my $data = $1;
      if($data =~ /,,,/) {
        print "Found blank for $street\n";

        my $str2 = $street;
        $str2 =~ s/,/ /g;
        my $content = get_content($str2);
        my ($tract, $block_group, $block, $district) = parse_content($str2,$content);

        print "$linenum\t";
        print "Tract: $tract\t";
        print "Block Group: $block_group\t";
        print "Block: $block\t";
        print "District: $district\n";
        print $new_out "$street,$tract,$block_group,$block,$district\n";
  
        # eh... who needs sleep? you're never gunna get it.
        # sleep 2;

      } else {
        print $new_out $outline;
      }
    } else {
      if(%row) {
      use Data::Dumper;
      print "DUMP: " . Dumper(\%row) . "\n";
      die "Error on line $linenum -- No match! '$street', $outline";
      }
    }
  }
  print "\nNo problemo!\n";
}


# Now we call one of the main routines. For now we just switch by hand.

#initial_get();

verify_and_reget();

=cut