#!/usr/local/bin/perl # # Copyright (c) 1998-1999, Shigeru Ishida (ishida@isl.intec.co.jp) # # Permission to use, copy, modify, distribute, and its documentation, # in whole or in part, for any purpose, is hereby granted, provided that: # # 1. This copyright and permission notice appear in all copies of the # software and its documentation. Notices of copyright or attribution # which appear at the beginning of any file must remain unchanged. # # 2. The name of Shigeru Ishida may not be used to endorse or # promote products derived from this software without specific # prior written permission. # # THIS SOFTWARE IS PROVIDED "AS IS" AND WITHOUT WARRANTY OF ANY KIND, # EXPRESS, IMPLIED, OR OTHERWISE, INCLUDING WITHOUT LIMITATION, ANY # WARRANTY OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. # IN NO EVENT SHALL SHIGERU ISHIDA BE LIABLE FOR ANY SPECIAL, INCIDENTAL, # INDIRECT OR CONSEQUENTIAL DAMAGES OF ANY KIND, OR ANY DAMAGES # WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER OR # NOT ADVISED OF THE POSSIBILITY OF DAMAGE, AND ON ANY THEORY OF # LIABILITY, ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE # OF THIS SOFTWARE. # # -- # PDF2TXT is a simple tool to extract text from PDF files. # # History: # 0.81 1999.11.21 # - Add '-i' option for extracting document information. # Thanks to Masaaki Kondo . # # 0.80 1999.10.13 # - Check font size to decide whether inserts a space code between words # or not. # # 0.79 1999.10.7 # - Modify dealing with text positioning operators. # # 0.78 1999.10.6 # - Check Type3 font encoding. # # 0.77 1999.10.5 # - Modify dealing with text positioning operators. # # 0.76 1999.10.4 # - Fix bug in reading trailer information. # # 0.75 1999.9.22 # - Fix bug in reading cross-reference table. # - Fix bug in extracting string data from the stream decoded. # # 0.74 1999.4.26 # - Correspond to Ext-H CMap. # # 0.73 1999.4.24 # - Modify cid2code(). # # 0.72 1999.4.23 # - Correspond to CMaps for Katakana and Hankaku. # Thanks to Rei FURUKAWA . # # 0.71 1999.4.22 # - Set start offset (in byte) of PDF content. In case of MacBinary II format, # it's offset is 128 bytes. Thanks to Rei FURUKAWA . # # 0.70 1999.4.14 # - Modify cid2code(). # - Fix bug in dealing with predefined font encodings. # # 0.66 1999.4.14 # - Fix bug in read_encodings(). # # 0.65 1999.4.13 # - Add $txtfh FileHandle variable and delete TXTFH FileHandle variable. # - Correspond to modify &CMap::cid2code(). # - Modify add_space_in_line() and add_space_in_word(). # # 0.64 1999.4.9 # - Change FileHandle from STDOUT to STDERR to output a message about loading # CMap files. # - Correspond to modify &CMap::cid2code(). # # 0.63 1999.4.8 # - Add two options to indicate a range of pages from which text is extracted. # -f : first page # -l : last page # # 0.62 1999.4.7 # - Fix bug in dealing with predefined font encodings. # # 0.61 1999.4.7 # - Fix bug in dealing with predefined font encodings. # # 0.60 1999.4.7 # - Correspond to predefined font encodings below: # StandardEncoding # SymbolEncoding # ZapfDingbatsEncoding # MacRomanEncoding # WinAnsiEncoding # - Remove '-o' option for omitting character code \ddd (octal). # # 0.57 1999.4.3 # - Correspond to Type 0 fonts and use CMap.pl to convert CID binary strings # into character codes. # # 0.56 1999.4.2 # * Read all names of encoding methods of all subtypes as well as Type0 fonts. # # 0.55 1999.4.2 # * Modify dealing with unknown decode-filters. # # 0.54 1999.4.2 # * Fix bug in extracting string data from the stream decoded. # # 0.53 1999.3.31 # * Add PDFLZW.pl. # # 0.52 1999.3.31 # * Add Base85.pl. # * Remove '-f' option for base85-decoding if length of base85 data not # a multiple of 5. # * Remove '-h' option to delete a hyphen as a word separator just before # newline. # # 0.51 1999.3.30 # * Read names of encoding methods of Type0 fonts. # # 0.50 1999.3.27 # * Prepared version for corresponding to Type 0 fonts. Notice that this # version may not yet deal with these fonts. # # 0.27 1999.1.22 # - Add the option to delete a hyphen as a word separator just before newline. # # 0.26 1999.1.21 # - Fix bug in extracting stream data. # - Fix bug in extracting string data from the stream decoded. # # 0.25 1998.12.17 # - Use a temporary file when uncompress is called to decompress LZW data. # And so IPC::Open2 is not used any longer. # # 0.24 1998.11.28 # - Refer the operators of drawing text data. And so remove '-n' option not # to put newline '\n' with the text extracted from PDF files. # # 0.23 1998.11.6 # - Fix bug in extracting stream data. # # 0.22 1998.10.31 # - Fix bug in converting "\(" into "(", "\)" into ")" and "\\" into "\". # # 0.21 1998.10.29 # - Fix bug about $lzw_stream variable. # # 0.20 1998.10.29 # - Convert PDF-format LZW data to compress-format LZW data and then deccode # the data by use of UNIX uncompress command. The standard UNIX compress utility # uses LZW, but with a slightly different file format. This conversion does *not* # decompress the data. It simply converts it to a different file format. # pdf2txt then calls uncompress to actually decompress the data. # For Unisys's LZW patent, see . # # 0.12 1998.10.27 # - Add the option for omitting character code \ddd (octal). # - Convert "\(" into "(", "\)" into ")" and "\\" into "\". # - Fix bug in extracting string data from the stream decoded. # # 0.11 1998.10.15 # - Add the option for the output on stdout. # # 0.10 1998.10.12 # - Use cross-reference table. # # 0.04 1998.10.10 # - Fix bug in dealing with unknown decode-filters. # - Add the option for base85-decoding if length of base85 data not a multiple of 5. # # 0.03 1998.10.9 # - Fix bug in extracting stream data. # # 0.02 1998.10.6 # - Fix bug in processing postscript after decode-filtering. # # 0.01 1998.10.5 # - Initial release. # # This tool uses these softwares below. # # . zlib http://www.cdrom.com/pub/infozip/zlib/ # . Compress::Zlib CPAN # . uncompress UNIX uncompress command # . Jcode http://openlab.ring.gr.jp/Jcode/ # . Base85.pl ftp://paprika.noc.intec.co.jp/pub/person/ishida/freeware/pdf2txt/ # . PDFLZW.pl ftp://paprika.noc.intec.co.jp/pub/person/ishida/freeware/pdf2txt/ # . PDFEncoding.pl ftp://paprika.noc.intec.co.jp/pub/person/ishida/freeware/pdf2txt/ # . makeCMap.pl ftp://paprika.noc.intec.co.jp/pub/person/ishida/freeware/pdf2txt/ # # Supported decoding methods: # # /ASCIIHexDecode # /ASCII85Decode ASCII85Decode filter, # Portable Document Format Reference Manual Version 1.2, p.51, # 27-Nov-1996. Adobe Systems Incorporated. # /FlateDecode RFC1951 # /LZWDecode See History: 0.20 above. # use Getopt::Std; getopts('f:l:si'); if ($opt_f >= 1) { $opt_f--; } else { $opt_f = 0; } if ($opt_l >= 1) { $opt_l--; } else { $opt_l = 100000; } if ($#ARGV < 0) { print < : first page to extract -l : last page to extract -s : output on stdout -i : document information (version 0.81 copyright (c) 1998-1999, Shigeru Ishida) USAGE exit(1); } use Compress::Zlib; use Jcode; require "Base85.pl"; require "PDFLZW.pl"; require "PDFEncoding.pl"; print STDERR "loading CMap...\n"; require "CMap.pl"; print STDERR "done.\n"; sleep 1; # Global variables: # PDFFH my($txtfh) = ""; local($pdffile); my($pdf_start_offset); my(%xref); my(%trailer); my(%info); my(@pages); my(%parents); my(%encodings); my($cur_encoding); my($cur_font); my($cur_cmap); %name2esc = ( "78-H" => "\x1b\x24\x40", "78-V" => "\x1b\x24\x40", "H" => "\x1b\x24\x42", "V" => "\x1b\x24\x42", "Hojo-H" => "\x1b\x24\x28\x44", "Hojo-V" => "\x1b\x24\x28\x44", "Katakana" => "\x1b\x28\x49", "Ext-H" => "\x1b\x24\x42", "Ext-V" => "\x1b\x24\x42", ); %char2print = ( Lslash => "L", OE => "OE", Scaron => "S", Ydieresis => "Y", Zcaron => "Z", ampersand => "\&", asciicircum => "\^", asciitilde => "\~", asterisk => "\*", at => "\@", backslash => "\\", bar => "\|", braceleft => "\{", braceright => "\}", bracketleft => "\[", bracketright => "\]", bullet => "\*", colon => "\:", comma => "\,", degree => "\-", dollar => "\$", dotlessi => "i", ellipsis => "\.\.\.", emdash => "\-", endash => "\-", equal => "\=", exclam => "\!", fi => "fi", fl => "fl", greater => "\>", hyphen => "\-", less => "\<", lslash => "l", numbersign => "\#", oe => "oe", parenleft => "\(", parenright => "\)", percent => "\%", period => "\.", plus => "\+", question => "\?", quotedbl => "\"", quotedblbase => "\"", quotedblleft => "\"", quotedblright => "\"", quoteleft => "\`", quoteright => "\'", quotesinglbase => "\'", quotesingle => "\'", scaron => "s", semicolon => "\;", slash => "/", space => " ", trademark => "TM", underscore => "_", yen => "\\", zcaron => "z", ); foreach $pdffile (@ARGV) { my($txtfile); if (!open(PDFFH, "$pdffile")) { warn "Can't open $pdffile: $!\n"; next; } binmode PDFFH; if ($opt_s == 1) { $txtfh = STDOUT; print STDERR "convert: $pdffile ==> stdout\n"; } else { if ($pdffile =~ /(.*)\.pdf$/i) { $txtfile = "$1.txt"; } else { $txtfile = "$pdffile.txt"; } if (!open($txtfh, ">$txtfile")) { warn "Can't open $txtfile: $!\n"; close PDFFH; next; } print STDERR "convert: $pdffile ==> $txtfile\n"; } # set start offset (in byte) of PDF content &set_start_offset(); # read tables of xref and trailer if (&read_xref_trailer("\r") == 0) { &read_xref_trailer("\n"); } # read encrypt information if exists if (defined $trailer{'Encrypt'}) { warn " Encrypted. (not process)\n"; next; } # read info dictionary &read_info(); if ($opt_i == 1) { my($t); foreach $t (sort keys(%info)) { print STDERR " $t: $info{$t}\n"; } } # read table of contents &read_pages(); # read name of encording method of subtype &read_encodings(); # proccess contents of pages for ($i = 0; $i <= $#pages; $i++) { next if $i < $opt_f || $i > $opt_l; my(@contents) = (); @contents = &read_contents($pages[$i]); &set_encoding($pages[$i]); for ($j = 0; $j <= $#contents; $j++) { my($obj, $gen) = split /\s+/, $contents[$j]; my($data, $filter); local($stream); my(@decodes, $line); $data = &get_object_by_ref($contents[$j]); if ($data =~ /^\d+\s+\d+\s+obj\s+\<\<\s+(.*)\s+\>\>\s+stream\r?\n?(.*)\r?\n?endstream\s+/s) { $filter = $1; $stream = $2; } else { next; } @decodes = &decode_mechod($filter); next if $decodes[0] eq "NO_APPLY"; next if &decode_stream(*stream, \@decodes) == 0; &output_stream(*stream); } print $txtfh "\n\f\n"; } close $txtfh if $opt_s != 1; close PDFFH; print STDERR " done.\n"; } &PDFLZW::unlink_tmpfile(); sub set_start_offset { my($fnsize); my($fn); my($ftype); $pdf_start_offset = 0; # check if MacBinary II or not seek(PDFFH, 1, 0); read(PDFFH, $fnsize, 1); $fnsize = unpack('C', $fnsize); return if $fnsize < 1 || $fnsize > 63; seek(PDFFH, 2, 0); read(PDFFH, $fn, $fnsize); return unless $pdffile =~ /^$fn$/i; seek(PDFFH, 65, 0); read(PDFFH, $ftype, 4); if ($ftype =~ /^pdf/i) { $pdf_start_offset = 128; return; } } sub get_object_by_ref { my($ref) = @_; my($sep, $line); seek(PDFFH, $xref{$ref} + $pdf_start_offset, 0); $sep = $/; $/ = "endobj"; $line = ; $/ = $sep; $line; } sub get_object_in_seek { my($sep, $line); $sep = $/; $/ = "endobj"; $line = ; $/ = $sep; $line; } sub parse_object { my($line, $list) = @_; my($pline, $t); if ($line =~ /^\d+\s+\d+\s+obj\s+\<\<\s*(.+)\s*\>\>\s+endobj\s*$/s) { $line = $1; } $line =~ s/\r([\d\]]+)/$1/g; $line = "\r" . $line; foreach $pline (split /[\r\n]\//, $line) { if ($pline =~ /^(\S+)\s+(\d+)\s+(\d+)\s+R\s*/s) { $$list{$1} = "$2 $3"; } elsif ($pline =~ /^(\S+)\s+(\d+)\s*/s) { $$list{$1} = $2; } elsif ($pline =~ /^(\S+)\s+\/(\S+)\s*/s) { $$list{$1} = $2; } elsif ($pline =~ /^(\S+)\s*\[\s*(.+)\s*\]\s*/s) { $$list{$1} = $2; $t = \$$list{$1}; $$t =~ s/^\s*//; $$t =~ s/\s*$//; } elsif ($pline =~ /^(\S+)\s*(\<(.+)\>)\s*/s) { $$list{$1} = $2; } elsif ($pline =~ /^(\S+)\s*(\(.+\))\s*/s) { $$list{$1} = $2; } } } sub read_xref_trailer { my($sep) = @_; %xref = (); %trailer = (); seek(PDFFH, 0, 0); $/ = $sep; while () { chop; next if !/^\s*xref\s*$/; while () { chop; if (/\s*(\d+)\s+(\d+)\s*$/) { my($obj_num, $obj_cnt) = ($1, $2); my($i, $key); for ($i = 0; $i < $obj_cnt; $i++) { $_ = ; if (/\s*(\d+)\s+(\d+)\s+n\s*$/) { $key = sprintf "%d", $2; $key = $obj_num+$i . " $key"; $xref{$key} = $1; } } next; } else { last; } } if (/^\s*trailer(\s*$|\s+\<\<)/) { my($line) = get_object_in_seek(); &parse_object($line, \%trailer); if ($trailer{'ID'} =~ /\s*\<([\da-fA-F]+)\>/) { $trailer{'ID'} = $1; $trailer{'ID'} =~ s/([\dA-F]{2})/pack('C',hex($1))/eig; } } } seek(PDFFH, 0, 0); scalar keys(%xref); } sub unescape_string { my($string) = @_; $$string =~ s/\\n/\n/g; $$string =~ s/\\r/\r/g; $$string =~ s/\\t/\t/g; $$string =~ s/\\b/\b/g; $$string =~ s/\\f/\f/g; $$string =~ s/\\\\/\\/g; $$string =~ s/\\\(/\(/g; $$string =~ s/\\\)/\)/g; } sub add_pages { my($ref) = @_; my($line); my(%list) = (); my($pline); my(@pages) = (); $line = &get_object_by_ref($ref); &parse_object($line, \%list); if ($list{'Type'} eq "Page") { $parents{$ref} = $list{'Parent'} if defined $list{'Parent'}; push @pages, $ref; } else { $parents{$ref} = $list{'Parent'} if defined $list{'Parent'}; foreach $pline (split /R/, $list{'Kids'}) { push @pages, &add_pages("$1 $2") if $pline =~ /^\s*(\d+)\s+(\d+)\s*$/; } } @pages; } sub read_info { my($obj, $gen) = split /\s+/, $trailer{'Info'}; my($line, $t); %info = (); $line = &get_object_by_ref($trailer{'Info'}); &parse_object($line, \%info); foreach $t (keys %info) { if ($info{$t} =~ /^\<(.+)\>$/) { $info{$t} = $1; $info{$t} =~ s/([\dA-F]{2})/pack('C',hex($1))/eig; } elsif ($info{$t} =~ /^\((.+)\)$/) { $info{$t} = $1; $info{$t} =~ s/\\\r//g; } $info{$t} =~ s/\\(\d{3})/pack('C',oct($1))/eig; &unescape_string(\$info{$t}); if (($t eq "CreationDate") || ($t eq "ModDate")) { $info{$t} = $1 if $info{$t} =~ /^D:\s*(.*)\s*$/; } if (($info{$t} =~ /^\xfe\xff(.*)$/) || ($info{$t} =~ /^\xff\xfe(.*)$/)) { $info{$t} = $1; if ($info{$t} =~ /^\x00\x1b(.{2,4})\x00\x1b(.*)$/) { my($lc) = $1; $info{$t} = $2; next if !($lc =~ /^ja/); } $info{$t} = Jcode::convert($info{$t}, euc, ucs2); } } } sub read_pages { my($line); my(%list) = (); @pages = (); %parents = (); $line = &get_object_by_ref($trailer{'Root'}); &parse_object($line, \%list); @pages = &add_pages($list{'Pages'}); } sub read_encodings { my($i, $line, $pline); %encodings = (); for ($i = 0; $i <= $#pages; $i++) { my(@f) = (); my(%fn) = (); my($fref); my($encoding_p) = {}; $line = &get_object_by_ref($pages[$i]); $line =~ s/[\r\n]/ /g; if ($line =~ /\/Resources\s+(\d+)\s+(\d+)\s+R/) { $pline = &get_object_by_ref("$1 $2"); $pline =~ s/[\r\n]/ /g; } elsif ($line =~ /\/Resources\s+\<\<\s+(.+)/) { $pline = $1; } else { next; } if ($pline =~ /\/Font\s+(\d+)\s+(\d+)\s+R/) { $pline = &get_object_by_ref("$1 $2"); $pline =~ s/[\r\n]/ /g; if ($pline =~ /\s+\<\<\s+([\da-zA-Z\/\s]+)\s+\>\>/) { my($pline2) = $1; my($t); foreach $t (split /R/, $pline2) { if ($t =~ /\/(\S+)\s+(\d+)\s+(\d+)\s*$/) { push @f, "$2 $3"; $fn{"$2 $3"} = $1; } } } else { next; } } elsif ($pline =~ /\/Font\s+\<\<\s+([\da-zA-Z\-\/\s]+)\s+\>\>/) { my($pline2) = $1; my($t); foreach $t (split /R/, $pline2) { if ($t =~ /\/(\S+)\s+(\d+)\s+(\d+)\s*$/) { push @f, "$2 $3"; $fn{"$2 $3"} = $1; } } } else { next; } foreach $fref (@f) { my($fline); my($subtype, $enc); my($basefont); my($eline); my($differences) = ""; my(@enc_array) = (); $fline = &get_object_by_ref($fref); $fline =~ s/[\r\n]/ /g; if ($fline =~ /\/Subtype\s+\/(\S+)\s/) { $subtype = $1; } else { next; } if ($fline =~ /\/BaseFont\s+\/(\S+)\s/) { $basefont = $1; } else { $basefont = ""; } if ($fline =~ /\/Encoding\s+\/(\S+)\s/) { $enc = $1; } else { if ($fline =~ /\/Encoding\s+(\d+)\s+(\d+)\s+R/) { $eline = &get_object_by_ref("$1 $2"); $eline =~ s/[\r\n]/ /g; } elsif ($fline =~ /\/Encoding\s+\<\<\s+([\da-zA-Z\/\s]+)\s+\>\>/) { $eline = $1; } else { next; } if ($eline =~ /\BaseEncoding\s+\/(\S+)\s/) { $enc = $1; } else { if (&PDFEncoding::check_encoding_by_basefont($basefont) == 1) { $enc = &PDFEncoding::get_name_by_basefont($basefont); } else { $enc = ""; } } } if ($eline =~ /\/Differences\s*\[\s*([\da-zA-Z\/\s]+)\s*\]/) { $differences = $1; $differences =~ s/[\/\r\n\s]+/ /g; } if (&PDFEncoding::check_encoding($enc) == 1) { my($code, $var) = (0, ""); @enc_array = &PDFEncoding::get_encoding($enc); foreach $var (split /\s+/, $differences) { if ($var =~ /^(\d+)$/) { $code = $var; } else { $enc_array[$code++] = $var; } } } if (defined $fn{$fref}) { $$encoding_p{$fn{$fref}} = {subtype => $subtype, encoding => $enc, encoding_array => [@enc_array]}; } elsif ($fline =~ /\/Name\s+\/(\S+)\s/) { $$encoding_p{$1} = {subtype => $subtype, encoding => $enc, encoding_array => [@enc_array]}; } } $encodings{$pages[$i]} = $encoding_p; } } sub add_contents { my($ref) = @_; my($line, $tline, $pline); my(%list) = (); my(@contents) = (); $line = &get_object_by_ref($ref); if ($line =~ /^\d+\s+\d+\s+obj\s+\[\s*(.+)\s*\]\s+endobj\s*$/s) { $tline = $1; foreach $pline (split /R/, $tline) { if ($pline =~ /^\s*(\d+)\s+(\d+)\s*$/) { push @contents, "$1 $2"; } } } else { push @contents, $ref; } @contents; } sub read_contents { my($ref) = @_; my($line); my(%list) = (); my(@contents) = (); $line = &get_object_by_ref($ref); &parse_object($line, \%list); foreach $pline (split /R/, $list{'Contents'}) { if ($pline =~ /^\s*(\d+)\s+(\d+)\s*$/) { push @contents, &add_contents("$1 $2"); } } @contents; } sub decode_mechod { local($line) = @_; my($filter) = ("NO_APPLY"); if ($line =~ /\s*\/Filter\s+\[\s*(.*)\s*\]\s/) { $filter = $1; } elsif ($line =~ /\s*\/Filter\s+(\S+)/) { $filter = $1; } else { return (); } if ($line =~ /\s*\/Type\s+\/(.*)\s+/) { if ($1 =~ /XObject/) { $filter = "NO_APPLY"; } } if ($line =~ /\s*\/Subtype\s+\/(.*)\s+/) { if ($1 =~ /Image/) { $filter = "NO_APPLY"; } } if ($line =~ /\s*\/Width\s+/) { $filter = "NO_APPLY"; } if ($line =~ /\s*\/Height\s+/) { $filter = "NO_APPLY"; } if ($line =~ /\s*\/BitsPerComponent\s+/) { $filter = "NO_APPLY"; } if ($line =~ /\s*\/ColorSpace\s+/) { $filter = "NO_APPLY"; } split(/\s+/, $filter); } sub inflate { local($input) = @_; local($x, $output, $status) = ('', '', ''); if (!($x = inflateInit())) { warn "Cannot create a inflation stream\n"; return ""; } ($output, $status) = $x->inflate(\$input); if ($status != Z_OK && $status != Z_STREAM_END) { warn "inflation failed\n"; return ""; } $output; } sub decode_stream { local(*stream, $decodes) = @_; my($unknown, $decode_cnt) = (0, 0); foreach $filter (@$decodes) { if ($filter eq "/ASCIIHexDecode") { $decode_cnt++; $stream =~ s/([\dA-F]{2})/pack('C',hex($1))/eig; last if length($stream) == 0; } elsif ($filter eq "/ASCII85Decode") { last if $#$decodes+1 == 1; $decode_cnt++; $stream = &Base85::decode($stream); last if length($stream) == 0; } elsif ($filter eq "/FlateDecode") { $decode_cnt++; $stream = &inflate($stream); last if length($stream) == 0; } elsif ($filter eq "/LZWDecode") { $decode_cnt++; $stream = &PDFLZW::uncompress($stream); last if length($stream) == 0; } else { if ($decode_cnt == 0) { $unknown = 1; last; } } } if ($unknown == 0) { return 1; } else { return 0; } } sub output_stream { local(*stream) = @_; my(@text_contents) = (); my($line); my($bt) = 0; local($ly) = 0; local(@tm) = (1, 0, 0, 1, 0, 0); my($fontsize) = 1; my($x_prev, $y_prev) = (0, 0); my($t); foreach $line (split(/[\r\n]/, $stream)) { if ($bt == 0) { if ($line =~ /^\s*BT\s*$/) { $bt = 1; next; } } if ($line =~ /^\s*ET\s*$/) { $bt = 0; next; } next if $line =~ /^\s*$/; if ($line =~ /\s*\/(\S+)\s+([\d\.]+)\s+Tf\s*/) { &set_cmap($1); $fontsize = $2; } &set_matrix($line, \$ly, \@tm); $line = &extract_text($line); next if length($line) == 0; push @text_contents, {x => $tm[4], y => $tm[5], fs => $fontsize, d => $line}; } foreach $t (@text_contents) { if (abs($t->{y}-$y_prev) >= 1) { print $txtfh "\n"; } elsif (($t->{x}-$x_prev)/$t->{fs} >= 1) { print $txtfh " "; } print $txtfh $t->{d}; $x_prev = $t->{x}; $y_prev = $t->{y}; } } sub set_matrix { my($line, $ly, $tm) = @_; my($x, $y); if ($line =~ /\s*T\*\s*/) { $x = $$tm[2]*$$ly+$$tm[4]; $y = $$tm[3]*$$ly+$$tm[5]; $$tm[4] = $x; $$tm[5] = $y; } elsif ($line =~ /\s*\'\s*$/) { $x = $$tm[2]*$$ly+$$tm[4]; $y = $$tm[3]*$$ly+$$tm[5]; $$tm[4] = $x; $$tm[5] = $y; } elsif ($line =~ /\s*\"\s*$/) { $x = $$tm[2]*$$ly+$$tm[4]; $y = $$tm[3]*$$ly+$$tm[5]; $$tm[4] = $x; $$tm[5] = $y; } elsif ($line =~ /\s*([\d\-\+\.]+)\s+TL\s*/) { $$ly = $1; } elsif ($line =~ /\s*([\d\-\+\.]+)\s+([\d\-\+\.]+)\s+Td\s*/) { $x = $$tm[0]*$1+$$tm[2]*$2+$$tm[4]; $y = $$tm[1]*$1+$$tm[3]*$2+$$tm[5]; $$tm[4] = $x; $$tm[5] = $y; } elsif ($line =~ /\s*([\d\-\+\.]+)\s+([\d\-\+\.]+)\s+TD\s*/) { $$ly = -$2; $x = $$tm[0]*$1+$$tm[2]*$2+$$tm[4]; $y = $$tm[1]*$1+$$tm[3]*$2+$$tm[5]; $$tm[4] = $x; $$tm[5] = $y; } elsif ($line =~ /\s*([\d\-\+\.]+)\s+([\d\-\+\.]+)\s+([\d\-\+\.]+)\s+([\d\-\+\.]+)\s+([\d\-\+\.]+)\s+([\d\-\+\.]+)\s+Tm\s*/) { $$tm[0] = $1; $$tm[1] = $2; $$tm[2] = $3; $$tm[3] = $4; $$tm[4] = $5; $$tm[5] = $6; } } sub extract_text { my($line) = @_; local($data) = ""; if ($line =~ /\[(.*)\]\s*TJ/) { $line = $1; $line =~ s/^\s*//; $line =~ s/\s*$//; $line =~ s/^[\d\-\.]+\s*//; $line =~ s/\>\s*([\d\-\.]+)\s*\ $1 \\s*Tj/) { $line = "<$1>"; } elsif ($line =~ /\((.*)\)\s*Tj/) { $line = "($1)"; } else { return ""; } if ($line =~ /^\s*(\<.*\>)\s*$/) { my($pline) = $1; my($t); foreach $t (split /\s+/, $pline) { if ($t =~ /^\<(.*)\>$/) { $t = $1; $t =~ s/([\dA-F]{2})/pack('C',hex($1))/eig; &cid2code(\$t); } elsif ($t =~ /^[\d\-\.]+$/) { $t = &add_space_in_word($t); } else { next; } $data .= $t; } } elsif ($line =~ /\((\\[\d\\]+)\)$/) { $data = $1; $data =~ s/\\(\d{3})/pack('C',oct($1))/eig; } elsif ($line =~ /\((.*)\)$/) { $data = $1; &unescape_string(\$data); $data =~ s/([\x00-\x1f\x7f-\xff])/&decode_charcode(unpack('C', $1))/eg; $data =~ s/\\(\d{3})/&decode_charcode(oct($1))/eg; } else { return ""; } if (&check_type("Type3") == 1) { $data =~ s/[\x00-\x1f]/ /g; } $data; } sub check_type { my($type) = @_; my($p) = $$cur_encoding{$cur_font}; return -1 unless defined $p; if ($p->{subtype} eq $type) { return 1; } else { return 0; } } sub add_space_in_word { if (-1 >= int($_[0]*13/2000)) { return " "; } else { return ""; } } sub set_encoding { my($ref) = @_; my($encoding_p) = ""; $cur_encoding = ""; $cur_cmap = ""; while (1) { $encoding_p = $encodings{$ref}; if (defined $encoding_p) { $cur_encoding = $encoding_p; return 1; } else { $ref = $parents{$ref}; return 0 unless defined $ref; next; } } } sub set_cmap { ($cur_font) = @_; my($p) = $$cur_encoding{$cur_font}; $cur_encoding_array = $p->{encoding_array}; $cur_cmap = $p->{encoding}; $cur_cmap =~ s/^Identity-(\S+)\s*/$1/; if ($p->{subtype} eq "Type0") { return &CMap::set_cmap($cur_cmap); } else { return -1; } } sub cid2code { my($string) = @_; my($p) = $$cur_encoding{$cur_font}; my($data) = $$string; return -1 unless defined $p; return -2 if $p->{subtype} ne "Type0"; if (defined $name2esc{$cur_cmap}) { $$string = ""; while (length($data) >= 2) { if ($data =~ /^(([\x01-\xff][\x00-\xff])+)/) { my($pdata) = $1; my($len) = 0; substr($data, 0, length($pdata)) = ""; while (length($pdata) > 0){ if (&_cid2code($string, \$pdata, \$len, $cur_cmap) < 0 && &_cid2code($string, \$pdata, \$len, "Katakana") < 0 && &_cid2code($string, \$pdata, \$len, "Hankaku") < 0 && &_cid2code($string, \$pdata, \$len, "Ext-H") < 0) { if (length($pdata) >= 2) { $$string .= substr($pdata, 0, 2); substr($pdata, 0, 2) = ""; next; } else { $$string .= $pdata; last; } } } }elsif ($data =~ /^((\x00[\x00-\xff])+)/) { my($pdata) = $1; my($len) = 0; substr($data, 0, length($pdata)) = ""; $pdata =~ s/\x00//g; while (length($pdata) > 0){ if (($pdata =~ /^[\x80-\xff]/ || &_cid2code($string, \$pdata, \$len, "UniJIS-UCS2-H") < 0) && &_cid2code($string, \$pdata, \$len, "Katakana") < 0 && &_cid2code($string, \$pdata, \$len, "Hankaku") < 0 && &_cid2code($string, \$pdata, \$len, "Ext-H") < 0) { if (length($pdata) >= 2) { $$string .= substr($pdata, 0, 2); substr($pdata, 0, 2) = ""; next; } else { $$string .= $pdata; last; } } } } } $$string .= $data; return 0; } else { my($len) = 0; my($ret); $ret = &CMap::cid2code(\$data, \$len); if ($ret == 1 || $ret == 0) { $$string = $data; return 1; } else { return $ret; } } } sub decode_charcode { my($code) = @_; my($print) = $char2print{$$cur_encoding_array[$code]}; if (defined $print) { return $print; } else { return pack('C', $code); } } sub _cid2code { my($string, $_pdata, $_len, $cmap) = @_; my $esc = $name2esc{$cmap}; my $esc_ascii = $esc ? "\x1b\x28\x4a" : ''; &CMap::set_cmap($cmap); my $ret = &CMap::cid2code($_pdata, $_len); my $conv = substr($$_pdata, 0, $$_len); substr($$_pdata, 0, $$_len) = ""; $conv =~ s/\x00//g; if ($ret == 1){ $$string .= $esc . $conv . $esc_ascii; return 1; }elsif ($ret == 0 && $$_len > 0){ $$string .= $esc . $conv . $esc_ascii; return 0; } -1; }