#!/usr/bin/perl -w
##########################
# Transform Personendaten
##########################

# Parst Datumsangaben und erweitert die Datumsfelder um je 8 Felder
# Sowie Artikel bei Orten und Vor/Nanchname

##########################

my %monate=(
  "Januar" => 1,
  "Februar" => 2,
  "März" => 3,
  "April" => 4,
  "Mai" => 5,
  "Juni" => 6,
  "Juli" => 7,
  "August" => 8,
  "September" => 9,
  "Oktober" => 10,
  "November" => 11,
  "Dezember" => 12
);

my %monatsnamen = ();
foreach $monat (keys %monate) { # inverse
  $monatsnamen{$monate{$monat}} = $monat;
}


my $MONAT  = "(" . join("|",keys %monate) . ")";
my $JAHR   = "(\\d{1,4})";
my $ZUSATZ = "([^0-9]+)";


# Bereinigt eine Datumsangabe für den Datumsstandard der Personendaten
sub clean_date {
  my $d = shift;

  # Eckige Klammern entfernen
  $d =~ s/[\[\]]//g;

  # DD.MM.YYYY => DD. MMM YYYY
  if ($d =~ /^0?(\d{1,2})\.(\d{1,2})\.(\d{4})$/) {
    $monat = int $2;
    $d = "$1. ".$monatsnamen{$monat}." $3";
  }

  # Vergessene Leerzeichen hinzufügen
  $d =~ s/([a-z])(\d)/$1 $2/g;
  $d =~ s/(\d)([a-z])/$1 $2/g;
  $d =~ s/\.([^ ])/. $1/g;

  # Doppelte Leerzeichen entfernen
  $d =~ s/\s+/ /g;

  # Leerzeichen am Anfang/Ende entfernen
  $d =~ s/^\s+//;
  $d =~ s/\s+$//;

  # Nur Fragezeichen => "unbekannt"
  $d =~ s/^\?$/unbekannt/;

  # n. Chr. ist implizit (unnötig)
  $d =~ s/n\. ?Chr\.//; # TODO: funtz nicht ganz?? z.B. bei [[Libanios]]

  # Fragezeichen immer klammern und Leerzeichen davor
  $d =~ s/\(\?\)/\?/g;
  $d =~ s/([^ ])\?+/$1 \?/g;
  $d =~ s/\?/(\?)/g;

  # Statt Fragezeichen am Ende "vermutlich" am Anfang
  $d =~ s/^(.+) \(\?\)$/vermutlich $1/;
  $d =~ s/^\(\?\)$/vermutlich/;

  # Slash ohne Leerzeichen
  $d =~ s/ ?\/ ?/\//g;

  # Jänner => Januar
  $d =~ s/Jänner/Januar/; 

  # Abgekürzte Monatsname
  $d =~ s/Jan\./Januar/; 
  $d =~ s/Feb\./Februar/; 
  $d =~ s/Aug\./August/;
  $d =~ s/Okt\./Oktober/;
  $d =~ s/Nov\./November/;
  $d =~ s/Dez\./Dezember/;

  # "Jahrhundert" ausschreiben
  $d =~ s/Jh\./Jahrhundert/g;

  # TOOD
  # "Ende des 5. Jahrhunderts" => "Ende 5. Jahrhundert" (einfacher)
  $d =~ s/Jahrhunderts/Jahrhundert/;
  $d =~ s/ des//;

  $d =~ s/(ungefähr|circa|ca\.?|etwa|gegen|~)/um/; 

  $d =~ s/(wohl|wahrscheinlich)/vermutlich/g;

  # Doppelte Leerzeichen entfernen
  $d =~ s/\s+/ /g;


  #----------Fehlerbereinigung---------
    
  # Sonstige Tippfehler
  $d =~ s/^Um/um/;
  $d =~ s/chr/Chr/;
      
  return $d;
}

sub ilog10 { $x=shift; return int log($x) / log(10); }  

my @fields = ( "day", "month", "year", "century", "decade", "year1", "year2", "note" );

sub parse_date {    
  $date = shift;
  local %d = ();
  
  if (! $date ) {
    $d{"note"} = "";
      
  # Normale Datumsangabe
  } elsif ($date =~ /^$ZUSATZ?(\d+\. )?$MONAT $JAHR( v\. Chr\.)?$/) {        
    $d{"note"} = $1 if defined $1;
    $d{"day"}    = substr($2,0,-2) if defined $2;
    if (defined $3) {
      $d{"month"} = $monate{$3};
      $d{"month"} = 3 if $3 =~ /M.rz/; # Sonderzeichen
    }  
    $d{"year"}   = int $4;
    $d{"year"} = -$d{"year"} if defined $5;  # v. Chr.
        
  } elsif ($date =~ /^$ZUSATZ?$JAHR( v\. Chr\.)?$/) {
    $d{"note"} = trim($1) if defined $1;
    $d{"year"}   = int $2;    
    $d{"year"} = -$d{"year"} if defined $3;  # v. Chr.        
    
  # Jahrhundert
  #} elsif ($date =~ /^(um |vermutlich |Anfang |Mitte |Ende )?(\d{1,2})\. Jahrhundert( v\. Chr\.)?$/) {
  } elsif ($date =~ /^$ZUSATZ?(\d{1,2})\. Jahrhundert( v\. Chr\.)?$/) {
    $d{"note"} = trim($1) if defined $1;
    $d{"century"} = $2;
    $d{"century"} = -$d{"century"} if defined $3;
  
  # Jahrhzehnt
  } elsif ($date =~ /^(\d{1,4}) ?er( v\. Chr.)?$/) {
    $d{"decade"} = (int $1 / 10) * 10;
    $d{"decade"} = -$d{"decade"} if defined $3;    
    $d{"century"} = int ($d{"decade"} / 10) + 1;
        
  } elsif ($date =~ /^$ZUSATZ?$JAHR oder $JAHR$/) { # TODO: oder != bis, TODO: v. Chr
    $d{"note"} = trim($1) if defined $1;
    $d{"year1"}  = int $2;
    $d{"year2"}  = int $3;
    
  } elsif ($date =~ /^$ZUSATZ?$JAHR\/(\d{1,4})$/) { # TODO: v. Chr
    $d{"note"} = trim($1) if defined $1;
    $d{"year1"}  = int $2;
    $d{"year2"}  = int $3;
    
    # Beispiel: 1632/33 => 1632/1633
    $c = ilog10($d{"year1"}) - ilog10($d{"year2"});
    if ( $c < 0 ) { # Fehler
      $d{"year1"} = "";
      $d{"year2"} = "";
      $d{"note"} = $date;     
    } else {
      $d{"year2"} = substr($d{"year1"},0,$c) . $d{"year2"};
    }   
    
  # Zeitraum von mehreren Jahren  # TODO: nicht getestet!
  } elsif ( $date =~ /^$ZUSATZ?zwischen $JAHR( und )$JAHR( v\. Chr\.)?$/ or
            $date =~ /^$ZUSATZ?$JAHR( bis |-)$JAHR$( v\. Chr\.)?/ )
  {
    $d{"note"} = trim($1) if defined $1;
    
    $d{"year1"}  = int $2;
    $d{"year2"}  = int $4;
        
    if (defined $5) {
      $tmp = $d{"year1"};
      $d{"year1"} = -$d{"year2"};
      $d{"year2"} = -$tmp;
    }   
    
  #} elsif ($date =~ /^$ZUSATZ?$JAHR( bis |-)$JAHR$/) {
  #  $d{"note"} = trim($1) if defined $1;
  #  $d{"year1"}  = int $2;
  #  $d{"year2"}  = int $4;

  } else {    
    $d{"note"} = $date; 
  }

  if (defined $d{"year1"} and defined $d{"year2"}) {
    # ggf. Jahre vertauschen
    #if ($d{"year1"} > $d{"year2"}) {
    #  ($d{"year1"}, $d{"year2"}) = ($d{"year2"}, $d{"year1"});
    #}  

    # ggf. Jahrhundert bestimmen    
    if ( (int $d{"year1"} / 100) eq (int $d{"year2"} / 100)) {
      $d{"century"} = (int $d{"year1"} / 100) + 1;
    }      

    # ggf. Jahrzehnt bestimmen    
    if ( $d{"century"} and (int $d{"year1"} / 10) eq (int $d{"year2"} / 10)) {
      $d{"decade"} = (int $d{"year1"} / 10) * 10;
    }

  }

  # Jahrzehnt und Jahrhundert ableiten
  if (defined $d{"year"}) {      
    $d{"decade"} = (int $d{"year"} / 10) * 10;
  }
  
  # Jahrhundert ableiten
  if (defined $d{"decade"}) {
    $d{"century"} = (int $d{"decade"} / 100) + 1;
  }  
  
  # Nicht vorhandene Felder setzen
  foreach $f (@fields) {
     $d{$f} = '' if not defined $d{$f};
  }

  return %d;
}

sub parse_location {
  $p = trim(shift);
  
  # [[a]] oder [[b]] <- sowas gibt's! 
    
  # [[...]]
  # [[...]], ...
  # [[...]]/ ... <- so besser nicht
  if ($p =~ /^\[\[([^\]]+)\]\](,.+)?$/) {
    $a = $1;    
    $a =~ s/\|.*$//;        
    #print "$p|$a\n";
    return $a;
  } else {
    #print "!$p\n";
    return "";     
  }  
}


sub parse_pnd {
  $p = trim(shift);
    
  $pnd_nr = "";
  $pnd_date = "";

  if ($p =~ /^([0-9])([0-9])([0-9])([0-9])([0-9])([0-9])([0-9])([0-9])([0-9X])/) {    
    $check = $9; 
    $check = 10 if($check eq "X");
    # pruefziffer    
    if ( ((2*$1+3*$2+4*$3+5*$4+6*$5+7*$6+8*$7+9*$8) % 11) eq $check ) {
      # TODO: Nummernbereich 10000000 bis 14999999
      if ($1 == "1" && $2>=0 && $2<=4) {
        $pnd_nr = "$1$2$3$4$5$6$7$8$9";
      }  
    }  
    
  } 
  
  if ($p =~ /(\d){1,2}\.(\d{1,2})\.(\d\d\d\d)/) {
    $pnd_date = "$3-$2-$1"; 
  } 
  
  return ($pnd_nr, $pnd_date);
}


sub unbracket {
  $p = shift;

  # Eckige Klammern entfernen
  $p =~ s/[\[\]]//g;
  
  return $p;
}  
   
sub trim {
  $p = shift;

  if ($p) {
    # Leerzeichen am Anfang/Ende entfernen
    $p =~ s/^\s+//;
    $p =~ s/\s+$//;
  }  

  return $p;
}
  
##########################

my @pd = ();

while(<>) {
  $line = $_;
  @pd = split("\t",$line);
  chop($pd[-1]); # remove end-of-line
  
  $pd_transformed[0] = trim($pd[0]); # pd_id
  $pd_transformed[1] = trim($pd[1]); # pd_article
  $pd_transformed[2] = trim($pd[2]); # pd_name
  $pd_transformed[3] = trim($pd[3]); # pd_alternative
  $pd_transformed[4] = trim($pd[4]); # pd_description
  $pd_transformed[5] = unbracket(trim($pd[5])); # pd_born  
  $pd_transformed[6] = trim($pd[6]); # pd_born_in
  $pd_transformed[7] = unbracket(trim($pd[7])); # pd_died
  $pd_transformed[8] = trim($pd[8]); # pd_died_in
  $pd_transformed[9] = trim($pd[9]); # pd_pnd
  
  # extract checked pnd-nr and additional date
  ($pd_transformed[10], $pd_transformed[11]) = parse_pnd($pd[9]); # pnr_nr, pnd_date
      
  if ( trim($pd[2]) =~ /^([^,]+),([^,]+)$/ ) {
    $pd_transformed[12] = trim($2); # n_given
    $pd_transformed[13] = trim($1); # n_surname
  } else {
    $pd_transformed[12] = ''; # n_given
    $pd_transformed[13] = ''; # n_surname
  }  

  $pd_transformed[14] = parse_location($pd[6]); # b_place
  $pd_transformed[15] = parse_location($pd[8]); # d_place
    
  %born = parse_date(clean_date($pd[5]));
  
  $pd_transformed[16] = $born{"day"};     # b_day
  $pd_transformed[17] = $born{"month"};   # b_month
  $pd_transformed[18] = $born{"year"};    # b_year
  $pd_transformed[19] = $born{"decade"};  # b_decade
  $pd_transformed[20] = $born{"century"}; # b_century
  $pd_transformed[21] = $born{"year1"};   # b_year1
  $pd_transformed[22] = $born{"year2"};   # b_year2
  $pd_transformed[23] = $born{"note"};    # b_note      
        
  %died = parse_date(clean_date($pd[7]));

  $pd_transformed[24] = $died{"day"};     # d_day
  $pd_transformed[25] = $died{"month"};   # d_month
  $pd_transformed[26] = $died{"year"};    # d_year
  $pd_transformed[27] = $died{"decade"};  # d_decade  
  $pd_transformed[28] = $died{"century"}; # d_century
  $pd_transformed[29] = $died{"year1"};   # d_year1
  $pd_transformed[30] = $died{"year2"};   # d_year2
  $pd_transformed[31] = $died{"note"};    # d_note

  print join("\t",@pd_transformed) . "\n";
}