########################################################################### # # CSVPlugin.pm -- A plugin for files in comma-separated value format # # A component of the Greenstone digital library software # from the New Zealand Digital Library Project at the # University of Waikato, New Zealand. # # Copyright 2006 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 CSVPlugin; use extrametautil; use ReadTextFile; use SplitTextFile; # for a couple routines, but we not inheriting use MetadataRead; use CSVFieldSeparator; use strict; no strict 'refs'; # allow filehandles to be variables and viceversa use Text::CSV; sub BEGIN { @CSVPlugin::ISA = ('MetadataRead', 'ReadTextFile', 'CSVFieldSeparator'); binmode(STDERR, ":utf8"); } my $arguments = [ { 'name' => "process_exp", 'desc' => "{BaseImporter.process_exp}", 'type' => "regexp", 'reqd' => "no", 'deft' => &get_default_process_exp() }, { 'name' => "filename_field", 'desc' => "{CSVPlugin.filename_field}", 'type' => "string", 'reqd' => "no", 'deft' => "Filename" }, { 'name' => "no_document_if_source_unspecified", 'desc' => "{CSVPlugin.no_document_if_source_unspecified}", 'type' => "flag", 'reqd' => "no"}, { 'name' => "no_document_if_source_missing", 'desc' => "{CSVPlugin.no_document_if_source_missing}", 'type' => "flag", 'reqd' => "no"}, { 'name' => "use_namespace_for_field_names", 'desc' => "{CSVPlugin.use_namespace_for_field_names}", 'type' => "string", 'reqd' => "no"}, { 'name' => "store_field_values_as_document_text", 'desc' => "{CSVPlugin.store_field_values_as_document_text}", 'type' => "flag", 'reqd' => "no"}, ]; my $options = { 'name' => "CSVPlugin", 'desc' => "{CSVPlugin.desc}", 'abstract' => "no", 'inherits' => "yes", 'explodes' => "yes", 'args' => $arguments }; # This plugin processes files with the suffix ".csv" sub get_default_process_exp { return q^(?i)(\.csv)$^; } sub new { my ($class) = shift (@_); my ($pluginlist,$inputargs,$hashArgOptLists) = @_; push(@$pluginlist, $class); push(@{$hashArgOptLists->{"ArgList"}}, @{$arguments}); push(@{$hashArgOptLists->{"OptList"}}, $options); new CSVFieldSeparator($pluginlist, $inputargs, $hashArgOptLists); my $self = new ReadTextFile($pluginlist, $inputargs, $hashArgOptLists); $self->{'textcat_store'} = {}; $self->{'metapass_srcdoc'} = {}; # which segments have valid metadata_srcdoc if ($self->{'use_namespace_for_field_names'} =~ /^\s*$/) { $self->{'use_namespace_for_field_names'} = undef; } return bless $self, $class; } # mark the file as a metadata file sub file_block_read { my $self = shift (@_); my ($pluginfo, $base_dir, $file, $block_hash, $metadata, $gli) = @_; my $filename_full_path = &FileUtils::filenameConcatenate($base_dir, $file); return undef unless $self->can_process_this_file($filename_full_path); if (($ENV{'GSDLOS'} =~ m/^windows$/) && ($^O ne "cygwin")) { # convert to full name - paths stored in block hash are long filenames $filename_full_path = &util::upgrade_if_dos_filename($filename_full_path); } # kjdon - upgrade method converts everyhting to lower case drive letter. # so would we need the following stuff??? # my $lower_drive = $filename_full_path; # $lower_drive =~ s/^([A-Z]):/\l$1:/i; # my $upper_drive = $filename_full_path; # $upper_drive =~ s/^([A-Z]):/\u$1:/i; # $block_hash->{'metadata_files'}->{$lower_drive} = 1; # $block_hash->{'metadata_files'}->{$upper_drive} = 1; # } # else { ### $block_hash->{'metadata_files'}->{$filename_full_path} = 1; # } $block_hash->{'metadata_files'}->{$filename_full_path} = 1; return undef; #1 } sub metadata_read { my $self = shift (@_); my ($pluginfo, $base_dir, $file, $block_hash, $extrametakeys, $extrametadata, $extrametafile, $processor, $gli, $aux) = @_; # can we process this file?? my ($filename_full_path, $filename_no_path) = &util::get_full_filenames($base_dir, $file); return undef unless $self->can_process_this_file_for_metadata($filename_full_path); print STDERR "\n\n" if ($gli); print STDERR "CSVPlugin: processing $file\n" if ($self->{'verbosity'}) > 1; my $outhandle = $self->{'outhandle'}; my $failhandle = $self->{'failhandle'}; my $verbosity = $self->{'verbosity'}; # don't add to block list, as we may do some processing in read. # Do encoding stuff my ($language, $content_encoding) = $self->textcat_get_language_encoding ($filename_full_path); if ($self->{'verbosity'} > 2) { print $outhandle "CSVPlugin: reading $file as ($content_encoding,$language)\n"; } # store these values for read my $le_rec = { 'language' => $language, 'encoding' => $content_encoding }; $self->{'textcat_store'}->{$file} = $le_rec; my $metadata_store = {}; $self->{'metadata_store'}->{$file} = $metadata_store; # used to record metadata for segments with no src doc my $CSV_FILE; open($CSV_FILE, "<:encoding($content_encoding)", "$filename_full_path"); my $separate_char = $self->{'csv_field_separator'}; my $md_val_sep = $self->{'metadata_value_separator'}; undef $md_val_sep if ($md_val_sep eq ""); my $csv_file_field_line; if ($separate_char =~ m/^auto$/i) { $csv_file_field_line = <$CSV_FILE>; $separate_char = $self->resolve_auto($csv_file_field_line,$self->{'plugin_type'}); seek $CSV_FILE, 0, 0; # move pointer back to start of file, as we want to read in the fields using csv. } my $csv = Text::CSV->new(); $csv->sep_char($separate_char); $csv->binary(1); my @csv_file_fields = undef; my $first_row = $csv->getline ($CSV_FILE); if (defined $first_row) { @csv_file_fields = @$first_row; } else { $self->print_error($outhandle, $failhandle, $gli, $filename_full_path, "Error: Badly formatted CSV header line: $csv_file_field_line"); return -1; } my $found_filename_field = 0; my $filename_field = $self->{'filename_field'}; for (my $i = 0; $i < scalar(@csv_file_fields); $i++) { # Remove any spaces from the field names, and surrounding quotes too $csv_file_fields[$i] =~ s/ //g; $csv_file_fields[$i] =~ s/^"//; $csv_file_fields[$i] =~ s/"$//; if ($self->{'use_namespace_for_field_names'}) { $csv_file_fields[$i] = $self->{'use_namespace_for_field_names'}. "." . $csv_file_fields[$i]; } if ($csv_file_fields[$i] eq $filename_field) { $found_filename_field = 1; } } if (!$found_filename_field) { $self->print_warning($outhandle, $failhandle, $gli, $filename_full_path, "No $filename_field field in CSV file, metadata cannot be assigned to documents, will use metadata only dummy documents"); } my $count = 0; while (my $csv_line = $csv->getline($CSV_FILE)) { my @md_vals = @$csv_line; $count++; # Build a hash of metadata name to metadata value for this line my %csv_line_metadata; my $md_vals_len = scalar(@md_vals); for (my $i=0; $i<$md_vals_len; $i++) { my $md_val = $md_vals[$i]; # Only bother with non-empty values if ($md_val ne "" && defined($csv_file_fields[$i])) { my $md_name = $csv_file_fields[$i]; $csv_line_metadata{$md_name} = []; if (defined $md_val_sep) { my @within_md_vals = split(/${md_val_sep}/,$md_val); # protect square brackets in metadata values by hex entity encoding them # As unescaped square bracket chars in metadata # have special meaning in GS' Java runtime code my @escaped_within_md_vals = (); for my $meta_value (@within_md_vals) { $meta_value =~ s/\[/&\#091;/g; $meta_value =~ s/\]/&\#093;/g; push(@escaped_within_md_vals, $meta_value); } push (@{$csv_line_metadata{$md_name}}, @escaped_within_md_vals); } else { # protect square brackets in metadata values by hex entity encoding them my $escaped_metadata_value = $md_val; $escaped_metadata_value =~ s/\[/&\#091;/g; $escaped_metadata_value =~ s/\]/&\#093;/g; push (@{$csv_line_metadata{$md_name}}, $escaped_metadata_value); } } } my $csv_line_section_array = $csv_line_metadata{"Section"}; my $section_suffix = ""; if (defined $csv_line_section_array) { my $section_value = shift(@$csv_line_section_array); if ($section_value =~ /[\d.]+/m){ my $section_suffix = "///Section/" . $section_value; foreach my $metaname (keys %csv_line_metadata) { my $new_name = $metaname . $section_suffix; $csv_line_metadata{$new_name} = delete $csv_line_metadata{$metaname}; } } else{ unshift(@$csv_line_section_array, $section_value); } } # do we have filename field? # We can't associate any metadata without knowing the file to associate it with my $has_srcdoc = 0; my $missing_srcdoc = 0; my $csv_line_filename="";; if ($found_filename_field) { # is there a srcdoc mentioned? my $csv_line_filename_array = $csv_line_metadata{$filename_field}; if (!defined $csv_line_filename_array) { $self->print_warning($outhandle, $failhandle, $gli, $filename_full_path, "No $filename_field metadata in CSV line num $count"); } else { $csv_line_filename = shift(@$csv_line_filename_array); # TODO - have an option for whether we do this or not if (&FileUtils::fileExists(&FileUtils::filenameConcatenate($base_dir, $csv_line_filename))) { $has_srcdoc = 1; delete $csv_line_metadata{$filename_field}; } else { $self->print_warning($outhandle, $failhandle, $gli, $filename_full_path, "$csv_line_filename in $filename_field metadata in CSV line num $count is not found"); $missing_srcdoc = 1; # there was one mentioned but its not found } } } if ($has_srcdoc) { print $outhandle "Storing metadata, segment $count, for document $csv_line_filename\n" if ($verbosity > 2); $self->store_meta_in_extrametadata($csv_line_filename, \%csv_line_metadata, $file, $filename_full_path, $extrametakeys, $extrametadata, $extrametafile); } else { my $store_for_dummy = 1; if ($missing_srcdoc && $self->{'no_document_if_source_missing'}) { $self->print_warning($outhandle, $failhandle, $gli, $filename_full_path,"Not storing metadata for line $count as source doc is missing"); $store_for_dummy = 0; } elsif(!$missing_srcdoc && $self->{'no_document_if_source_unspecified'}) { $self->print_warning($outhandle, $failhandle, $gli, $filename_full_path,"Not storing metadata for line $count as source doc is unspecified"); $store_for_dummy = 0; } if ($store_for_dummy) { print $outhandle "Storing metadata for dummy document, segment $count\n" if ($verbosity > 2); $metadata_store->{$count} = \%csv_line_metadata; } } } # while csv_line = csv->getline close ($CSV_FILE); } #adapted from read in splittextfile sub read { my $self = shift (@_); my ($pluginfo, $base_dir, $file, $block_hash, $metadata, $processor, $maxdocs, $total_count, $gli) = @_; my $outhandle = $self->{'outhandle'}; my $verbosity = $self->{'verbosity'}; # can we process this file?? my ($filename_full_path, $filename_no_path) = &util::get_full_filenames($base_dir, $file); return undef unless $self->can_process_this_file($filename_full_path); $file =~ s/^[\/\\]+//; # $file often begins with / so we'll tidy it up my $le_rec = $self->{'textcat_store'}->{$file}; if (!defined $le_rec) { # means no text was found; return 0; # not processed but no point in passing it on } print STDERR "\n" if ($gli); print $outhandle "$self->{'plugin_type'} processing $file\n" if $self->{'verbosity'} > 1; my $language = $le_rec->{'language'}; my $encoding = $le_rec->{'encoding'}; $self->{'textcat_store'}->{$file} = undef; my $metadata_store = $self->{'metadata_store'}->{$file}; # a hash of seg num to metadata hash # Process each segment in turn my $segment = 0; #which segment/record number we have my $count = 0; # num doc objs produced my ($filemeta) = $file =~ /([^\\\/]+)$/; #why? my $plugin_filename_encoding = $self->{'filename_encoding'}; my $filename_encoding = $self->deduce_filename_encoding($file,$metadata,$plugin_filename_encoding); my $id; foreach $segment (sort { $a <=> $b } keys (%$metadata_store)) { print $outhandle "processing segment $segment as its own document\n" if $self->{'verbosity'} > 1; $count++; # create a new document my $doc_obj = new doc ($filename_full_path, "indexed_doc", $self->{'file_rename_method'}); my $cursection = $doc_obj->get_top_section(); $doc_obj->add_utf8_metadata($cursection, "Language", $language); $doc_obj->add_utf8_metadata($cursection, "Encoding", $encoding); $self->set_Source_metadata($doc_obj, $filename_full_path, $filename_encoding); $doc_obj->add_utf8_metadata($cursection, "SourceSegment", "$segment"); if ($self->{'cover_image'}) { $self->associate_cover_image($doc_obj, $filename_full_path); } $doc_obj->add_utf8_metadata($cursection, "Plugin", "$self->{'plugin_type'}"); # include any metadata passed in from previous plugins # note that this metadata is associated with the top level section $self->extra_metadata ($doc_obj, $cursection, $metadata); # add our stored metadata from metadata_read pass my $segment_metadata = $metadata_store->{$segment}; $self->extra_metadata($doc_obj, $cursection, $segment_metadata); if ($self->{'store_field_values_as_document_text'}) { my $new_text = ""; foreach my $f (keys %$segment_metadata) { my $values = $segment_metadata->{$f}; $new_text .= join (", ", @$values).", "; } $doc_obj->add_utf8_text($cursection, $new_text); } # do any automatic metadata extraction - does this make sense?? #$self->auto_extract_metadata ($doc_obj); # Calculate a "base" document ID. if (!defined $id) { $id = &SplitTextFile::get_base_OID($self,$doc_obj); } # add an OID &SplitTextFile::add_segment_OID($self, $doc_obj, $id, $segment); # process the document $processor->process($doc_obj); $self->{'num_processed'} ++; if ($maxdocs != -1 && $self->{'num_processed'} >= $maxdocs) { last; } } delete $self->{'metadata_store'}->{$file}; # Return number of document objects produced return $count; } sub print_warning { my $self = shift(@_); my ($outhandle, $failhandle, $gli, $file, $error) = @_; print $outhandle "CSVPlugin Warning: $file: $error\n"; print $failhandle "CSVPlugin Warning: $file: $error\n"; print STDERR "\n" if ($gli); } sub print_error { my $self = shift(@_); my ($outhandle, $failhandle, $gli, $file, $error) = @_; print $outhandle "CSVPlugin Error: $file: $error\n"; print $failhandle "CSVPlugin Error: $file: $error\n"; print STDERR "\n" if ($gli); } 1;