########################################################################### # # BasPlug.pm -- base class for all the import plugins # A component of the Greenstone digital library software # from the New Zealand Digital Library Project at the # University of Waikato, New Zealand. # # Copyright (C) 1999-2005 New Zealand Digital Library Project # # This program is free software; you can redistribute it and/or modify # it under the terms of the GNU General Public License as published by # the Free Software Foundation; either version 2 of the License, or # (at your option) any later version. # # This program is distributed in the hope that it will be useful, # but WITHOUT ANY WARRANTY; without even the implied warranty of # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the # GNU General Public License for more details. # # You should have received a copy of the GNU General Public License # along with this program; if not, write to the Free Software # Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. # ########################################################################### package BasPlug; BEGIN { die "GSDLHOME not set\n" unless defined $ENV{'GSDLHOME'}; } eval {require bytes}; # suppress the annoying "subroutine redefined" warning that various # plugins cause under perl 5.6 $SIG{__WARN__} = sub {warn($_[0]) unless ($_[0] =~ /Subroutine\s+\S+\sredefined/)}; use strict; no strict 'subs'; no strict 'refs'; # allow filehandles to be variables and viceversa use File::Basename; use Kea; use multiread; use encodings; use cnseg; use acronym; use textcat; use doc; eval "require diagnostics"; # some perl distros (eg mac) don't have this use DateExtract; use ghtml; use gsprintf 'gsprintf'; use printusage; use parse2; use GISBasPlug; @BasPlug::ISA = ( GISBasPlug ); my $unicode_list = [ { 'name' => "ascii", 'desc' => "{BasPlug.input_encoding.ascii}" }, { 'name' => "utf8", 'desc' => "{BasPlug.input_encoding.utf8}" }, { 'name' => "unicode", 'desc' => "{BasPlug.input_encoding.unicode}" } ]; my $auto_unicode_list = [ { 'name' => "auto", 'desc' => "{BasPlug.input_encoding.auto}" } ]; my $e = $encodings::encodings; foreach my $enc (sort {$e->{$a}->{'name'} cmp $e->{$b}->{'name'}} keys (%$e)) { my $hashEncode = {'name' => $enc, 'desc' => $e->{$enc}->{'name'}}; push(@{$unicode_list},$hashEncode); } push(@{$auto_unicode_list},@{$unicode_list}); my $arguments = [ { 'name' => "process_exp", 'desc' => "{BasPlug.process_exp}", 'type' => "regexp", 'deft' => "", 'reqd' => "no" }, { 'name' => "block_exp", 'desc' => "{BasPlug.block_exp}", 'type' => "regexp", 'deft' => "", 'reqd' => "no" }, { 'name' => "smart_block", 'desc' => "{BasPlug.smart_block}", 'type' => "flag", 'reqd' => "no", 'hiddengli' => "yes" }, { 'name' => "associate_ext", 'desc' => "{BasPlug.associate_ext}", 'type' => "string", 'reqd' => "no" }, { 'name' => "associate_tail_re", 'desc' => "{BasPlug.associate_tail_re}", 'type' => "string", 'reqd' => "no" }, { 'name' => "input_encoding", 'desc' => "{BasPlug.input_encoding}", 'type' => "enum", 'list' => $auto_unicode_list, 'reqd' => "no" , 'deft' => "auto" } , { 'name' => "default_encoding", 'desc' => "{BasPlug.default_encoding}", 'type' => "enum", 'list' => $unicode_list, 'reqd' => "no", 'deft' => "utf8" }, { 'name' => "extract_language", 'desc' => "{BasPlug.extract_language}", 'type' => "flag", 'reqd' => "no" }, { 'name' => "default_language", 'desc' => "{BasPlug.default_language}", 'type' => "string", 'deft' => "en", 'reqd' => "no" }, { 'name' => "extract_acronyms", 'desc' => "{BasPlug.extract_acronyms}", 'type' => "flag", 'reqd' => "no" }, { 'name' => "markup_acronyms", 'desc' => "{BasPlug.markup_acronyms}", 'type' => "flag", 'reqd' => "no" }, { 'name' => "extract_keyphrases", 'desc' => "{BasPlug.extract_keyphrases}", 'type' => "flag", 'reqd' => "no" }, { 'name' => "extract_keyphrases_kea4", 'desc' => "{BasPlug.extract_keyphrases_kea4}", 'type' => "flag", 'reqd' => "no" }, { 'name' => "extract_keyphrase_options", 'desc' => "{BasPlug.extract_keyphrase_options}", 'type' => "string", 'deft' => "", 'reqd' => "no" }, { 'name' => "first", 'desc' => "{BasPlug.first}", 'type' => "string", 'reqd' => "no" }, { 'name' => "extract_email", 'desc' => "{BasPlug.extract_email}", 'type' => "flag", 'reqd' => "no" }, { 'name' => "extract_historical_years", 'desc' => "{BasPlug.extract_historical_years}", 'type' => "flag", 'reqd' => "no" }, { 'name' => "maximum_year", 'desc' => "{BasPlug.maximum_year}", 'type' => "int", 'deft' => (localtime)[5]+1900, 'char_length' => "4", #'range' => "2,100", 'reqd' => "no"}, { 'name' => "maximum_century", 'desc' => "{BasPlug.maximum_century}", 'type' => "string", 'deft' => "-1", 'reqd' => "no" }, { 'name' => "no_bibliography", 'desc' => "{BasPlug.no_bibliography}", 'type' => "flag", 'reqd' => "no"}, { 'name' => "no_cover_image", 'desc' => "{BasPlug.no_cover_image}", 'type' => "flag", 'reqd' => "no" }, { 'name' => "separate_cjk", 'desc' => "{BasPlug.separate_cjk}", 'type' => "flag", 'reqd' => "no", 'hiddengli' => "yes" }, { 'name' => "new_extract_email", 'desc' => "", 'type' => "flag", 'reqd' => "no", 'hiddengli' => "yes" } ]; my $gis_arguments = [ { 'name' => "extract_placenames", 'desc' => "{GISBasPlug.extract_placenames}", 'type' => "flag", 'reqd' => "no" }, { 'name' => "gazetteer", 'desc' => "{GISBasPlug.gazetteer}", 'type' => "string", 'reqd' => "no" }, { 'name' => "place_list", 'desc' => "{GISBasPlug.place_list}", 'type' => "flag", 'reqd' => "no" } ]; my $options = { 'name' => "BasPlug", 'desc' => "{BasPlug.desc}", 'abstract' => "yes", 'inherits' => "no", 'args' => $arguments }; sub set_keepold { my $self = shift(@_); my ($keepold) = @_; $self->{'keepold'} = $keepold; } sub get_arguments { my $self = shift(@_); my $optionlistref = $self->{'option_list'}; my @optionlist = @$optionlistref; my $pluginoptions = pop(@$optionlistref); my $pluginarguments = $pluginoptions->{'args'}; return $pluginarguments; } sub print_xml_usage { my $self = shift(@_); # XML output is always in UTF-8 gsprintf::output_strings_in_UTF8; PrintUsage::print_xml_header(); $self->print_xml(); } sub print_xml { my $self = shift(@_); my $optionlistref = $self->{'option_list'}; my @optionlist = @$optionlistref; my $pluginoptions = shift(@$optionlistref); return if (!defined($pluginoptions)); gsprintf(STDERR, "\n"); gsprintf(STDERR, " $pluginoptions->{'name'}\n"); my $desc = gsprintf::lookup_string($pluginoptions->{'desc'}); $desc =~ s//&gt;/g; gsprintf(STDERR, " $desc\n"); gsprintf(STDERR, " $pluginoptions->{'abstract'}\n"); gsprintf(STDERR, " $pluginoptions->{'inherits'}\n"); gsprintf(STDERR, " " . ($pluginoptions->{'explodes'} || "no") . "\n"); gsprintf(STDERR, " \n"); if (defined($pluginoptions->{'args'})) { &PrintUsage::print_options_xml($pluginoptions->{'args'}); } # Recurse up the plugin hierarchy $self->print_xml(); gsprintf(STDERR, " \n"); gsprintf(STDERR, "\n"); } sub print_txt_usage { my $self = shift(@_); # Print the usage message for a plugin (recursively) my $descoffset = $self->determine_description_offset(0); $self->print_plugin_usage($descoffset, 1); } sub determine_description_offset { my $self = shift(@_); my $maxoffset = shift(@_); my $optionlistref = $self->{'option_list'}; my @optionlist = @$optionlistref; my $pluginoptions = shift(@$optionlistref); return $maxoffset if (!defined($pluginoptions)); # Find the length of the longest option string of this plugin my $pluginargs = $pluginoptions->{'args'}; if (defined($pluginargs)) { my $longest = &PrintUsage::find_longest_option_string($pluginargs); if ($longest > $maxoffset) { $maxoffset = $longest; } } # Recurse up the plugin hierarchy $maxoffset = $self->determine_description_offset($maxoffset); $self->{'option_list'} = \@optionlist; return $maxoffset; } sub print_plugin_usage { my $self = shift(@_); my $descoffset = shift(@_); my $isleafclass = shift(@_); my $optionlistref = $self->{'option_list'}; my @optionlist = @$optionlistref; my $pluginoptions = shift(@$optionlistref); return if (!defined($pluginoptions)); my $pluginname = $pluginoptions->{'name'}; my $pluginargs = $pluginoptions->{'args'}; my $plugindesc = $pluginoptions->{'desc'}; # Produce the usage information using the data structure above if ($isleafclass) { if (defined($plugindesc)) { gsprintf(STDERR, "$plugindesc\n\n"); } gsprintf(STDERR, " {common.usage}: plugin $pluginname [{common.options}]\n\n"); } # Display the plugin options, if there are some if (defined($pluginargs)) { # Calculate the column offset of the option descriptions my $optiondescoffset = $descoffset + 2; # 2 spaces between options & descriptions if ($isleafclass) { gsprintf(STDERR, " {common.specific_options}:\n"); } else { gsprintf(STDERR, " {common.general_options}:\n", $pluginname); } # Display the plugin options &PrintUsage::print_options_txt($pluginargs, $optiondescoffset); } # Recurse up the plugin hierarchy $self->print_plugin_usage($descoffset, 0); $self->{'option_list'} = \@optionlist; } sub new { # Set Encodings to the list!! # Start the BasPlug Constructor my $class = shift (@_); my ($pluginlist,$args,$hashArgOptLists) = @_; push(@$pluginlist, $class); my $plugin_name = (defined $pluginlist->[0]) ? $pluginlist->[0] : $class; if(defined $arguments){ push(@{$hashArgOptLists->{"ArgList"}},@{$arguments});} if(defined $options) { push(@{$hashArgOptLists->{"OptList"}},$options)}; if (GISBasPlug::has_mapdata()) { push(@$arguments,@$gis_arguments); } my $self = {}; $self->{'outhandle'} = STDERR; $self->{'option_list'} = $hashArgOptLists->{"OptList"}; $self->{"info_only"} = 0; # Check if gsdlinfo is in the argument list or not - if it is, don't parse # the args, just return the object. foreach my $strArg (@{$args}) { if($strArg eq "-gsdlinfo") { $self->{"info_only"} = 1; return bless $self, $class; } } if(!parse2::parse($args,$hashArgOptLists->{"ArgList"},$self)) { my $classTempClass = bless $self, $class; print STDERR "\n"; &gsprintf(STDERR, "\n{BasPlug.bad_general_option}\n", $plugin_name); $classTempClass->print_txt_usage(""); # Use default resource bundle die "\n"; } delete $self->{"info_only"}; # else parsing was successful. $self->{'plugin_type'} = $plugin_name; #$self->{'outhandle'} = STDERR; $self->{'textcat'} = new textcat(); $self->{'num_processed'} = 0; $self->{'num_not_processed'} = 0; $self->{'num_blocked'} = 0; $self->{'num_archives'} = 0; $self->{'cover_image'} = 1; # cover image is on by default $self->{'cover_image'} = 0 if ($self->{'no_cover_image'}); #$self->{'option_list'} = $hashArgOptLists->{"OptList"}; my $associate_ext = $self->{'associate_ext'}; if ((defined $associate_ext) && ($associate_ext ne "")) { my $associate_tail_re = $self->{'associate_tail_re'}; if ((defined $associate_tail_re) && ($associate_tail_re ne "")) { my $outhandle = $self->{'outhandle'}; print $outhandle "Warning: can only specify 'associate_ext' or 'associate_tail_re'\n"; print $outhandle " defaulting to 'associate_tail_re'\n"; } else { my @exts = split(/,/,$associate_ext); my @exts_bracketed = map { $_ = "(?:\\.$_)" } @exts; my $associate_tail_re = join("|",@exts_bracketed); $self->{'associate_tail_re'} = $associate_tail_re; } delete $self->{'associate_ext'}; } $self->{'shared_fileroot'} = {}; $self->{'file_blocks'} = {}; if ($self->{'extract_placenames'}) { my $outhandle = $self->{'outhandle'}; my $places_ref = GISBasPlug::loadGISDatabase($outhandle,$self->{'gazetteer'}); if (!defined $places_ref) { print $outhandle "Warning: Error loading mapdata gazetteer \"$self->{'gazetteer'}\"\n"; print $outhandle " No placename extraction will take place.\n"; $self->{'extract_placenames'} = undef; } else { $self->{'places'} = $places_ref; } } return bless $self, $class; } # initialize BasPlug options # if init() is overridden in a sub-class, remember to call BasPlug::init() sub init { my $self = shift (@_); my ($verbosity, $outhandle, $failhandle) = @_; # verbosity is passed through from the processor $self->{'verbosity'} = $verbosity; # as are the outhandle and failhandle $self->{'outhandle'} = $outhandle if defined $outhandle; $self->{'failhandle'} = $failhandle; # set process_exp and block_exp to defaults unless they were # explicitly set if ((!$self->is_recursive()) and (!defined $self->{'process_exp'}) || ($self->{'process_exp'} eq "")) { $self->{'process_exp'} = $self->get_default_process_exp (); if ($self->{'process_exp'} eq "") { warn ref($self) . " Warning: Non-recursive plugin has no process_exp\n"; } } if ((!defined $self->{'block_exp'}) || ($self->{'block_exp'} eq "")) { $self->{'block_exp'} = $self->get_default_block_exp (); } } sub begin { my $self = shift (@_); my ($pluginfo, $base_dir, $processor, $maxdocs) = @_; $self->initialise_extractors(); } sub end { # potentially called at the end of each plugin pass # import.pl only has one plugin pass, but buildcol.pl has multiple ones my ($self) = @_; $self->finalise_extractors(); } sub deinit { # called only once, after all plugin passes have been done my ($self) = @_; } # this function should be overridden to return 1 # in recursive plugins sub is_recursive { my $self = shift (@_); return 0; } sub get_default_block_exp { my $self = shift (@_); return ""; } sub get_default_process_exp { my $self = shift (@_); return ""; } # default implementation is to do nothing. sub store_block_files { my $self =shift (@_); my ($filename) = @_; return; } #default implementation is to block a file with same name as this, but extension jpg or JPG, if cover_images is on. sub block_cover_image { my $self =shift; my $filename = shift; if ($self->{'cover_image'}) { my $coverfile = $filename; $coverfile =~ s/\.[^\\\/\.]+$/\.jpg/; if (!-e $coverfile) { $coverfile =~ s/jpg$/JPG/; } if (-e $coverfile) { $self->{'file_blocks'}->{$coverfile} = 1; } } return; } sub root_ext_split { my $self = shift (@_); my ($filename,$tail_re) = @_; my ($file_prefix,$file_ext) = ($filename =~ m/^(.*?)($tail_re)$/); if ((!defined $file_prefix) || (!defined $file_ext)) { ($file_prefix,$file_ext) = ($filename =~ m/^(.*)(\..*?)$/); } return ($file_prefix,$file_ext); } sub metadata_read { my $self = shift (@_); my ($pluginfo, $base_dir, $file, $metadata, $extrametakeys, $extrametadata, $processor, $maxdocs, $gli) = @_; # Keep track of filenames with same root but different extensions # Used to support -associate_ext and the more generalised # -associate_tail_re my $associate_tail_re = $self->{'associate_tail_re'}; if ((defined $associate_tail_re) && ($associate_tail_re ne "")) { my ($file_prefix,$file_ext) = $self->root_ext_split($file,$associate_tail_re); if ((defined $file_prefix) && (defined $file_ext)) { my $shared_fileroot = $self->{'shared_fileroot'}; if (!defined $shared_fileroot->{$file_prefix}) { my $file_prefix_rec = { 'tie_to' => undef, 'exts' => {} }; $shared_fileroot->{$file_prefix} = $file_prefix_rec; } my $file_prefix_rec = $shared_fileroot->{$file_prefix}; my $process_exp = $self->{'process_exp'}; if ($file =~ m/$process_exp/) { # This is the document the others should be tied to $file_prefix_rec->{'tie_to'} = $file_ext; } else { if ($file_ext =~ m/$associate_tail_re$/) { $file_prefix_rec->{'exts'}->{$file_ext} = 1; } } } } # now check whether we are actually processing this my $filename = $file; $filename = &util::filename_cat ($base_dir, $file) if $base_dir =~ /\w/; if ($self->{'process_exp'} eq "" || $filename !~ /$self->{'process_exp'}/ || !-f $filename) { return undef; # can't recognise } # do smart blocking if appropriate if ($self->{'smart_block'}) { $self->store_block_files($filename); } # block the cover image if there is one if ($self->{'cover_image'}) { $self->block_cover_image($filename); } return 1; } sub tie_to_filename { my $self = shift (@_); my ($file_ext,$file_prefix_rec) = @_; if (defined $file_prefix_rec) { my $tie_to = $file_prefix_rec->{'tie_to'}; if (defined $tie_to) { if ($tie_to eq $file_ext) { return 1; } } } return 0; } sub tie_to_assoc_file { my $self = shift (@_); my ($file_ext,$file_prefix_rec) = @_; if (defined $file_prefix_rec) { my $tie_to = $file_prefix_rec->{'tie_to'}; if (defined $tie_to) { my $exts = $file_prefix_rec->{'exts'}; my $has_file_ext = $exts->{$file_ext}; if ($has_file_ext) { return 1; } } } return 0; } sub associate_with { my $self = shift (@_); my ($file, $filename, $metadata) = @_; my $associate_tail_re = $self->{'associate_tail_re'}; return 0 if (!$associate_tail_re); # If file, see if matches with "tie_to" doc or is one of the # associated filename extensions. my ($file_prefix,$file_ext) = $self->root_ext_split($file,$associate_tail_re); if ((defined $file_prefix) && (defined $file_ext)) { my $file_prefix_rec = $self->{'shared_fileroot'}->{$file_prefix}; if ($self->tie_to_filename($file_ext,$file_prefix_rec)) { # Set up gsdlassocfile_tobe my $exts = $file_prefix_rec->{'exts'}; if (!defined $metadata->{'gsdlassocfile_tobe'}) { $metadata->{'gsdlassocfile_tobe'} = []; } my $assoc_tobe = $metadata->{'gsdlassocfile_tobe'}; my ($full_prefix) = ($filename =~ m/^(.*)\..*?$/); foreach my $e (keys %$exts) { my $assoc_file = "$full_prefix$e"; print STDERR " $self->{'plugin_type'}: Associating $file_prefix$e with $file_prefix_rec->{'tie_to'} version\n"; my $mime_type = ""; # let system auto detect this push(@$assoc_tobe,"$assoc_file:$mime_type:"); } } elsif ($self->tie_to_assoc_file($file_ext,$file_prefix_rec)) { # a form of smart block return 1; } } return 0; } sub read_block { my $self = shift (@_); my ($pluginfo, $base_dir, $file, $metadata, $processor, $maxdocs, $total_count, $gli) = @_; my $filename = $file; $filename = &util::filename_cat ($base_dir, $file) if $base_dir =~ /\w/; if ($self->associate_with($file,$filename,$metadata)) { # a form of smart block $self->{'num_blocked'} ++; return (0,undef); # blocked } my $smart_block = $self->{'smart_block'}; my $smart_block_BN = $self->{'smart_block_BN'}; if ($smart_block || $smart_block_BN) { if (defined $self->{'file_blocks'}->{$filename} && $self->{'file_blocks'}->{$filename} == 1){ $self->{'num_blocked'} ++; return (0,undef); # blocked } } else { if ($self->{'block_exp'} ne "" && $filename =~ /$self->{'block_exp'}/) { $self->{'num_blocked'} ++; return (0,undef); # blocked } if ($self->{'cover_image'}) { if (defined $self->{'file_blocks'}->{$filename} && $self->{'file_blocks'}->{$filename} == 1){ $self->{'num_blocked'} ++; return (0,undef); # blocked } } } if ($filename !~ /$self->{'process_exp'}/ || !-f $filename) { return (undef,undef); # can't recognise } return (1,$filename); } sub read_tidy_file { my $self = shift (@_); my ($file) = @_; $file =~ s/^[\/\\]+//; # $file often begins with / so we'll tidy it up return $file; } # The BasPlug read_into_doc_obj() function. This function does all the # right things to make general options work for a given plugin. It reads in # a file and sets up a slew of metadata all saved in doc_obj, which # it then returns as part of a tuple (process_status,doc_obj) # # Much of this functionality used to reside in read, but it was broken # down into a supporting routine to make the code more flexible. # # recursive plugins (e.g. RecPlug) and specialized plugins like those # capable of processing many documents within a single file (e.g. # GMLPlug) will normally want to implement their own version of # read_into_doc_obj() # # Note that $base_dir might be "" and that $file might # include directories sub read_into_doc_obj { my $self = shift (@_); my ($pluginfo, $base_dir, $file, $metadata, $processor, $maxdocs, $total_count, $gli) = @_; if ($self->is_recursive()) { gsprintf(STDERR, "{BasPlug.read_must_be_implemented}") && die "\n"; } my $outhandle = $self->{'outhandle'}; my ($block_status,$filename) = $self->read_block(@_); return $block_status if ((!defined $block_status) || ($block_status==0)); $file = $self->read_tidy_file($file); # Do encoding stuff my ($language, $encoding) = $self->textcat_get_language_encoding ($filename); # create a new document my $doc_obj = new doc ($filename, "indexed_doc"); $doc_obj->set_OIDtype ($processor->{'OIDtype'}); $doc_obj->add_utf8_metadata($doc_obj->get_top_section(), "Language", $language); $doc_obj->add_utf8_metadata($doc_obj->get_top_section(), "Encoding", $encoding); $doc_obj->add_utf8_metadata($doc_obj->get_top_section(), "Plugin", "$self->{'plugin_type'}"); $doc_obj->add_utf8_metadata($doc_obj->get_top_section(), "FileSize", (-s $filename)); my ($filemeta) = $file =~ /([^\\\/]+)$/; # how do we know what encoding the filename is in? $doc_obj->add_metadata($doc_obj->get_top_section(), "Source", &ghtml::dmsafe($filemeta)); if ($self->{'cover_image'}) { $self->associate_cover_image($doc_obj, $filename); } # read in file ($text will be in utf8) my $text = ""; $self->read_file ($filename, $encoding, $language, \$text); if (!length ($text)) { my $plugin_name = ref ($self); if ($gli) { print STDERR "\n"; } gsprintf($outhandle, "$plugin_name: {BasPlug.file_has_no_text}\n", $filename) if $self->{'verbosity'}; my $failhandle = $self->{'failhandle'}; gsprintf($failhandle, "$file: " . ref($self) . ": {BasPlug.empty_file}\n"); # print $failhandle "$file: " . ref($self) . ": file contains no text\n"; $self->{'num_not_processed'} ++; return (0,undef); # what should we return here?? error but don't want to pass it on } # include any metadata passed in from previous plugins # note that this metadata is associated with the top level section my $associate_tail_re = $self->{'associate_tail_re'}; $self->extra_metadata ($doc_obj, $doc_obj->get_top_section(), $metadata); # do plugin specific processing of doc_obj unless (defined ($self->process (\$text, $pluginfo, $base_dir, $file, $metadata, $doc_obj, $gli))) { $text = ''; undef $text; print STDERR "\n" if ($gli); return (-1,undef); } $text=''; undef $text; # do any automatic metadata extraction $self->auto_extract_metadata ($doc_obj); # add an OID # see if there is a plugin-specific set_OID function... if (defined ($self->can('set_OID'))) { # it will need $doc_obj to set the Identifier metadata... $self->set_OID($doc_obj); } else { # use the default set_OID() in doc.pm $doc_obj->set_OID(); } return (1,$doc_obj); } # The BasPlug read() function. This function calls read_into_doc_obj() # to ensure all the right things to make general options work for a # given plugin are done. It then calls the process() function which # does all the work specific to a plugin (like the old read functions # used to do). Most plugins should define their own process() function # and let this read() function keep control. # # recursive plugins (e.g. RecPlug) and specialized plugins like those # capable of processing many documents within a single file (e.g. # GMLPlug) might want to implement their own version of read(), but # more likely need to implement their own version of read_into_doc_obj() # # Return number of files processed, undef if can't recognise, -1 if can't # process sub read { my $self = shift (@_); my ($pluginfo, $base_dir, $file, $metadata, $processor, $maxdocs, $total_count, $gli) = @_; my ($process_status,$doc_obj) = $self->read_into_doc_obj(@_); if ((defined $process_status) && ($process_status == 1)) { # process the document $processor->process($doc_obj); if(defined($self->{'places_filename'})){ &util::rm($self->{'places_filename'}); $self->{'places_filename'} = undef; } $self->{'num_processed'} ++; undef $doc_obj; } # if process_status == 1, then the file has been processed. return $process_status; } # returns undef if file is rejected by the plugin sub process { my $self = shift (@_); my ($textref, $pluginfo, $base_dir, $file, $metadata, $doc_obj, $gli) = @_; gsprintf(STDERR, "BasPlug::process {common.must_be_implemented}\n") && die "\n"; # die "Basplug::process function must be implemented in sub-class\n"; return undef; # never gets here } # uses the multiread package to read in the entire file pointed to # by filename and loads the resulting text into $$textref. Input text # may be in any of the encodings handled by multiread, output text # will be in utf8 sub read_file { my $self = shift (@_); my ($filename, $encoding, $language, $textref) = @_; if (!-r $filename) { my $outhandle = $self->{'outhandle'}; gsprintf($outhandle, "{BasPlug.read_denied}\n", $filename) if $self->{'verbosity'}; # print $outhandle "Read permission denied for $filename\n" if $self->{'verbosity'}; return; } $$textref = ""; if (!open (FILE, $filename)) { gsprintf(STDERR, "BasPlug::read_file {BasPlug.could_not_open_for_reading} ($!)\n", $filename); die "\n"; } if ($encoding eq "ascii") { undef $/; $$textref = ; $/ = "\n"; } else { my $reader = new multiread(); $reader->set_handle ('BasPlug::FILE'); $reader->set_encoding ($encoding); $reader->read_file ($textref); #Now segments chinese if the separate_cjk option is set if ($self->{'separate_cjk'}) { # segment the Chinese words $$textref = &cnseg::segment($$textref); } } close FILE; } # write_file -- used by ConvertToPlug, for example in post processing # sub utf8_write_file { my $self = shift (@_); my ($textref, $filename) = @_; if (!open (FILE, ">$filename")) { gsprintf(STDERR, "ConvertToPlug::write_file {ConvertToPlug.could_not_open_for_writing} ($!)\n", $filename); die "\n"; } print FILE $$textref; close FILE; } sub filename_based_title { my $self = shift (@_); my ($file) = @_; my $file_derived_title = $file; $file_derived_title =~ s/_/ /g; $file_derived_title =~ s/\..*?$//; return $file_derived_title; } sub title_fallback { my $self = shift (@_); my ($doc_obj,$section,$file) = @_; if (!defined $doc_obj->get_metadata_element ($section, "Title")) { my $file_derived_title = $self->filename_based_title($file); $doc_obj->add_metadata ($section, "Title", $file_derived_title); } } sub textcat_get_language_encoding { my $self = shift (@_); my ($filename) = @_; my ($language, $encoding, $extracted_encoding); if ($self->{'input_encoding'} eq "auto") { # use textcat to automatically work out the input encoding and language ($language, $encoding) = $self->get_language_encoding ($filename); } elsif ($self->{'extract_language'}) { # use textcat to get language metadata ($language, $extracted_encoding) = $self->get_language_encoding ($filename); $encoding = $self->{'input_encoding'}; # don't print this message for english... english in utf8 is identical # to english in iso-8859-1 (except for some punctuation). We don't have # a language model for en_utf8, so textcat always says iso-8859-1! if ($extracted_encoding ne $encoding && $language ne "en" && $self->{'verbosity'}) { my $plugin_name = ref ($self); my $outhandle = $self->{'outhandle'}; gsprintf($outhandle, "$plugin_name: {BasPlug.wrong_encoding}\n", $filename, $encoding, $extracted_encoding); # print $outhandle "$plugin_name: WARNING: $filename was read using $encoding encoding but "; # print $outhandle "appears to be encoded as $extracted_encoding.\n"; } } else { $language = $self->{'default_language'}; $encoding = $self->{'input_encoding'}; } return ($language, $encoding); } # Uses textcat to work out the encoding and language of the text in # $filename. All html tags are removed before processing. # returns an array containing "language" and "encoding" sub get_language_encoding { my $self = shift (@_); my ($filename) = @_; my $outhandle = $self->{'outhandle'}; my $unicode_format = ""; # read in file open (FILE, $filename) || (gsprintf(STDERR, "BasPlug::get_language_encoding {BasPlug.could_not_open_for_reading} ($!)\n", $filename) && die "\n"); # die "BasPlug::get_language_encoding could not open $filename for reading ($!)\n"; undef $/; my $text = ; $/ = "\n"; close FILE; # check if first few bytes have a Byte Order Marker my $bom=substr($text,0,2); # check 16bit unicode if ($bom eq "\xff\xfe") { # little endian 16bit unicode $unicode_format="unicode"; } elsif ($bom eq "\xfe\xff") { # big endian 16bit unicode $unicode_format="unicode"; } else { $bom=substr($text,0,3); # check utf-8 if ($bom eq "\xef\xbb\xbf") { # utf-8 coded FEFF bom $unicode_format="utf8"; # } elsif ($bom eq "\xef\xbf\xbe") { # utf-8 coded FFFE bom. Error!? # $unicode_format="utf8"; } } # VB scripting generated Word to HTML file if ($text =~ /charset=(windows.*)[\"]/ig){ my $vbhtml_encoding = $1; $vbhtml_encoding =~ s/-+/_/g; $self->{'input_encoding'} = $vbhtml_encoding; } # remove stuff -- as titles tend often to be in English # for foreign language documents $text =~ s/(.|\n)*?<\/title>//i; # remove all HTML tags # XXX this doesn't match plugins derived from HTMLPlug (except ConvertTo) if (ref($self) eq 'HTMLPlug' || (exists $self->{'converted_to'} && $self->{'converted_to'} eq 'HTML')){ $text =~ s/<[^>]*>//sg; } # get the language/encoding my $results = $self->{'textcat'}->classify(\$text); # if textcat returns 3 or less possibilities we'll use the # first one in the list - otherwise use the defaults if (scalar @$results > 3) { my $best_encoding=""; if ($unicode_format) { # in case the first had a BOM $best_encoding=$unicode_format; } else { my %guessed_encodings = (); foreach my $result (@$results) { $result =~ /([^\-]+)$/; my $enc=$1; if (!defined($guessed_encodings{$enc})) { $guessed_encodings{$enc}=0; } $guessed_encodings{$enc}++; } $guessed_encodings{""}=-1; # for default best_encoding of "" foreach my $enc (keys %guessed_encodings) { if ($guessed_encodings{$enc} > $guessed_encodings{$best_encoding}){ $best_encoding=$enc; } } } if ($self->{'input_encoding'} ne 'auto') { if ($self->{'extract_language'} && ($self->{'verbosity'}>2)) { gsprintf($outhandle, "BasPlug: {BasPlug.could_not_extract_language}\n", $filename, $self->{'default_language'}); } return ($self->{'default_language'}, $self->{'input_encoding'}); } else { if ($self->{'verbosity'}>2) { gsprintf($outhandle, "BasPlug: {BasPlug.could_not_extract_language}\n", $filename, $self->{'default_language'}); } return ($self->{'default_language'}, $best_encoding); } } # format language/encoding my ($language, $encoding) = $results->[0] =~ /^([^-]*)(?:-(.*))?$/; if (!defined $language) { if ($self->{'verbosity'}>2) { gsprintf($outhandle, "BasPlug: {BasPlug.could_not_extract_language}\n", $filename, $self->{'default_language'}); } $language = $self->{'default_language'}; } if (!defined $encoding) { if ($self->{'verbosity'}>2) { gsprintf($outhandle, "BasPlug: {BasPlug.could_not_extract_encoding}\n", $filename, $self->{'default_encoding'}); } $encoding = $self->{'default_encoding'}; } # check for equivalents where textcat doesn't have some encodings... # eg MS versions of standard encodings if ($encoding =~ /^iso_8859_(\d+)/) { my $iso = $1; # which variant of the iso standard? # iso-8859 sets don't use chars 0x80-0x9f, windows codepages do if ($text =~ /[\x80-\x9f]/) { # Western Europe if ($iso == 1 or $iso == 15) { $encoding = 'windows_1252' } elsif ($iso == 2) { $encoding = 'windows_1250' } # Central Europe elsif ($iso == 5) { $encoding = 'windows_1251' } # Cyrillic elsif ($iso == 6) { $encoding = 'windows_1256' } # Arabic elsif ($iso == 7) { $encoding = 'windows_1253' } # Greek elsif ($iso == 8) { $encoding = 'windows_1255' } # Hebrew elsif ($iso == 9) { $encoding = 'windows_1254' } # Turkish } } if ($encoding !~ /^(ascii|utf8|unicode)$/ && !defined $encodings::encodings->{$encoding}) { if ($self->{'verbosity'}) { gsprintf($outhandle, "BasPlug: {BasPlug.unsupported_encoding}\n", $filename, $encoding, $self->{'default_encoding'}); } $encoding = $self->{'default_encoding'}; } return ($language, $encoding); } # add any extra metadata that's been passed around from one # plugin to another. # extra_metadata uses add_utf8_metadata so it expects metadata values # to already be in utf8 sub extra_metadata { my $self = shift (@_); my ($doc_obj, $cursection, $metadata) = @_; my $associate_tail_re = $self->{'associate_tail_re'}; foreach my $field (keys(%$metadata)) { # $metadata->{$field} may be an array reference if ($field eq "gsdlassocfile_tobe") { # 'gsdlassocfile_tobe' is artificially introduced metadata # that is used to signal that certain additional files should # be tied to this document. Useful in situations where a # metadata pass in the plugin pipeline works out some files # need to be associated with a document, but the document hasn't # been formed yet. my $equiv_form = ""; foreach my $gaf (@{$metadata->{$field}}) { my ($full_filename,$mimetype) = ($gaf =~ m/^(.*):(.*):$/); my ($tail_filename) = ($full_filename =~ /^.*[\/\\](.+?)$/); my $filename = $full_filename; $doc_obj->associate_file($full_filename,$tail_filename,$mimetype); # work out extended tail extension (i.e. matching tail re) my ($file_prefix,$file_extended_ext) = $self->root_ext_split($tail_filename,$associate_tail_re); my ($pre_doc_ext) = ($file_extended_ext =~ m/^(.*)\..*$/); my ($doc_ext) = ($tail_filename =~ m/^.*\.(.*)$/); my $start_doclink = "<a href=\"_httpcollection_/index/assoc/{Or}{[parent(Top):archivedir],[archivedir]}/$tail_filename\">"; my $srcicon = "_icon".$doc_ext."_"; my $end_doclink = "</a>"; my $assoc_form = "$start_doclink\{If\}{$srcicon,$srcicon,$doc_ext\}$end_doclink"; if (defined $pre_doc_ext) { # for metadata such as [mp3._edited] [mp3._full] ... $doc_obj->add_utf8_metadata ($cursection, "$doc_ext.$pre_doc_ext", $assoc_form); } # for multiple metadata such as [mp3.assoclink] $doc_obj->add_utf8_metadata ($cursection, "$doc_ext.assoclink", $assoc_form); $equiv_form .= " $assoc_form"; } $doc_obj->add_utf8_metadata ($cursection, "equivlink", $equiv_form); } elsif (ref ($metadata->{$field}) eq "ARRAY") { map { $doc_obj->add_utf8_metadata ($cursection, $field, $_); } @{$metadata->{$field}}; } else { $doc_obj->add_utf8_metadata ($cursection, $field, $metadata->{$field}); } } } # initialise metadata extractors sub initialise_extractors { my $self = shift (@_); if ($self->{'extract_acronyms'} || $self->{'markup_acronyms'}) { &acronym::initialise_acronyms(); } } # finalise metadata extractors sub finalise_extractors { my $self = shift (@_); if ($self->{'extract_acronyms'} || $self->{'markup_acronyms'}) { &acronym::finalise_acronyms(); } } # FIRSTNNN: extract the first NNN characters as metadata sub extract_first_NNNN_characters { my $self = shift (@_); my ($textref, $doc_obj, $thissection) = @_; foreach my $size (split /,/, $self->{'first'}) { my $tmptext = $$textref; $tmptext =~ s/^\s+//; $tmptext =~ s/\s+$//; $tmptext =~ s/\s+/ /gs; $tmptext = substr ($tmptext, 0, $size); $tmptext =~ s/\s\S*$/…/; $doc_obj->add_utf8_metadata ($thissection, "First$size", $tmptext); } } sub extract_email { my $self = shift (@_); my ($textref, $doc_obj, $thissection) = @_; my $outhandle = $self->{'outhandle'}; gsprintf($outhandle, " {BasPlug.extracting_emails}...\n") if ($self->{'verbosity'} > 2); my @email = ($$textref =~ m/([-a-z0-9\.@+_=]+@(?:[-a-z0-9]+\.)+(?:com|org|edu|mil|int|net|[a-z][a-z]))/g); @email = sort @email; # if($self->{"new_extract_email"} == 0) # { # my @email2 = (); # foreach my $address (@email) # { # if (!(join(" ",@email2) =~ m/(^| )$address( |$)/ )) # { # push @email2, $address; # $doc_obj->add_utf8_metadata ($thissection, "emailAddress", $address); # # print $outhandle " extracting $address\n" # &gsprintf($outhandle, " {BasPlug.extracting} $address\n") # if ($self->{'verbosity'} > 3); # } # } # } # else # { my $hashExistMail = {}; foreach my $address (@email) { if (!(defined $hashExistMail->{$address})) { $hashExistMail->{$address} = 1; $doc_obj->add_utf8_metadata ($thissection, "emailAddress", $address); gsprintf($outhandle, " {BasPlug.extracting} $address\n") if ($self->{'verbosity'} > 3); } } gsprintf($outhandle, " {BasPlug.done_email_extract}\n") if ($self->{'verbosity'} > 2); } # extract metadata sub auto_extract_metadata { my $self = shift (@_); my ($doc_obj) = @_; if ($self->{'extract_email'}) { my $thissection = $doc_obj->get_top_section(); while (defined $thissection) { my $text = $doc_obj->get_text($thissection); $self->extract_email (\$text, $doc_obj, $thissection) if $text =~ /./; $thissection = $doc_obj->get_next_section ($thissection); } } if ($self->{'extract_placenames'}) { my $thissection = $doc_obj->get_top_section(); while (defined $thissection) { my $text = $doc_obj->get_text($thissection); $self->extract_placenames (\$text, $doc_obj, $thissection) if $text =~ /./; $thissection = $doc_obj->get_next_section ($thissection); } } if ($self->{'extract_keyphrases'} || $self->{'extract_keyphrases_kea4'}) { $self->extract_keyphrases($doc_obj); } if ($self->{'first'}) { my $thissection = $doc_obj->get_top_section(); while (defined $thissection) { my $text = $doc_obj->get_text($thissection); $self->extract_first_NNNN_characters (\$text, $doc_obj, $thissection) if $text =~ /./; $thissection = $doc_obj->get_next_section ($thissection); } } if ($self->{'extract_acronyms'}) { my $thissection = $doc_obj->get_top_section(); while (defined $thissection) { my $text = $doc_obj->get_text($thissection); $self->extract_acronyms (\$text, $doc_obj, $thissection) if $text =~ /./; $thissection = $doc_obj->get_next_section ($thissection); } } if ($self->{'markup_acronyms'}) { my $thissection = $doc_obj->get_top_section(); while (defined $thissection) { my $text = $doc_obj->get_text($thissection); $text = $self->markup_acronyms ($text, $doc_obj, $thissection); $doc_obj->delete_text($thissection); $doc_obj->add_text($thissection, $text); $thissection = $doc_obj->get_next_section ($thissection); } } if($self->{'extract_historical_years'}) { my $thissection = $doc_obj->get_top_section(); while (defined $thissection) { my $text = $doc_obj->get_text($thissection); &DateExtract::get_date_metadata($text, $doc_obj, $thissection, $self->{'no_bibliography'}, $self->{'maximum_year'}, $self->{'maximum_century'}); $thissection = $doc_obj->get_next_section ($thissection); } } } #adding kea keyphrases sub extract_keyphrases { my $self = shift(@_); my $doc_obj = shift(@_); # Use Kea 3.0 unless 4.0 has been specified my $kea_version = "3.0"; if ($self->{'extract_keyphrases_kea4'}) { $kea_version = "4.0"; } # Check that Kea exists, and tell the user where to get it if not my $keahome = &Kea::get_Kea_directory($kea_version); if (!-e $keahome) { gsprintf(STDERR, "{BasPlug.missing_kea}\n", $keahome, $kea_version); return; } my $thissection = $doc_obj->get_top_section(); my $text = ""; my $list; #loop through sections to gather whole doc while (defined $thissection) { my $sectiontext = $doc_obj->get_text($thissection); $text = $text.$sectiontext; $thissection = $doc_obj->get_next_section ($thissection); } if($self->{'extract_keyphrase_options'}) { #if kea options flag is set, call Kea with specified options $list = &Kea::extract_KeyPhrases ($kea_version, $text, $self->{'extract_keyphrase_options'}); } else { #otherwise call Kea with no options $list = &Kea::extract_KeyPhrases ($kea_version, $text); } if ($list){ # if a list of kea keyphrases was returned (ie not empty) if ($self->{'verbosity'}) { gsprintf(STDERR, "{BasPlug.keyphrases}: $list\n"); } #add metadata to top section $thissection = $doc_obj->get_top_section(); # add all key phrases as one metadata $doc_obj->add_metadata($thissection, "Keyphrases", $list); # add individual key phrases as multiple metadata foreach my $keyphrase (split(',', $list)) { $keyphrase =~ s/^\s+|\s+$//g; $doc_obj->add_metadata($thissection, "Keyphrase", $keyphrase); } } } # extract acronyms from a section in a document. progress is # reported to outhandle based on the verbosity. both the Acronym # and the AcronymKWIC metadata items are created. sub extract_acronyms { my $self = shift (@_); my ($textref, $doc_obj, $thissection) = @_; my $outhandle = $self->{'outhandle'}; # print $outhandle " extracting acronyms ...\n" gsprintf($outhandle, " {BasPlug.extracting_acronyms}...\n") if ($self->{'verbosity'} > 2); my $acro_array = &acronym::acronyms($textref); foreach my $acro (@$acro_array) { #check that this is the first time ... my $seen_before = "false"; my $previous_data = $doc_obj->get_metadata($thissection, "Acronym"); foreach my $thisAcro (@$previous_data) { if ($thisAcro eq $acro->to_string()) { $seen_before = "true"; if ($self->{'verbosity'} >= 4) { gsprintf($outhandle, " {BasPlug.already_seen} " . $acro->to_string() . "\n"); } } } if ($seen_before eq "false") { #write it to the file ... $acro->write_to_file(); #do the normal acronym $doc_obj->add_utf8_metadata($thissection, "Acronym", $acro->to_string()); gsprintf($outhandle, " {BasPlug.adding} ".$acro->to_string()."\n") if ($self->{'verbosity'} > 3); } } gsprintf($outhandle, " {BasPlug.done_acronym_extract}\n") if ($self->{'verbosity'} > 2); } sub markup_acronyms { my $self = shift (@_); my ($text, $doc_obj, $thissection) = @_; my $outhandle = $self->{'outhandle'}; gsprintf($outhandle, " {BasPlug.marking_up_acronyms}...\n") if ($self->{'verbosity'} > 2); #self is passed in to check for verbosity ... $text = &acronym::markup_acronyms($text, $self); gsprintf($outhandle, " {BasPlug.done_acronym_markup}\n") if ($self->{'verbosity'} > 2); return $text; } sub compile_stats { my $self = shift(@_); my ($stats) = @_; $stats->{'num_processed'} += $self->{'num_processed'}; $stats->{'num_not_processed'} += $self->{'num_not_processed'}; $stats->{'num_archives'} += $self->{'num_archives'}; } sub associate_cover_image { my $self = shift; my ($doc_obj, $filename) = @_; $filename =~ s/\.[^\\\/\.]+$/\.jpg/; if (exists $self->{'covers_missing_cache'}->{$filename}) { # don't stat() for existence eg for multiple document input files # (eg SplitPlug) return; } my $top_section=$doc_obj->get_top_section(); if (-e $filename) { $doc_obj->associate_file($filename, "cover.jpg", "image/jpeg"); $doc_obj->add_utf8_metadata($top_section, "hascover", 1); } else { my $upper_filename = $filename; $upper_filename =~ s/jpg$/JPG/; if (-e $upper_filename) { $doc_obj->associate_file($upper_filename, "cover.jpg", "image/jpeg"); $doc_obj->add_utf8_metadata($top_section, "hascover", 1); } else { # file doesn't exist, so record the fact that it's missing so # we don't stat() again (stat is slow) $self->{'covers_missing_cache'}->{$filename} = 1; } } } 1;