—

              
#------------------------------------------------------------------------------
# File:         Import.pm
#
# Description:  Import CSV and JSON database files
#
# Revisions:    2011-03-05 - P. Harvey Created
#------------------------------------------------------------------------------
package Image::ExifTool::Import;
use strict;
require Exporter;
use vars qw($VERSION @ISA @EXPORT_OK);
$VERSION = '1.14';
@ISA = qw(Exporter);
@EXPORT_OK = qw(ReadCSV ReadJSON);
sub ReadJSONObject($;$);
my %unescapeJSON = ( 't'=>"\t", 'n'=>"\n", 'r'=>"\r", 'b' => "\b", 'f' => "\f" );
my $charset;
#------------------------------------------------------------------------------
# Read CSV file
# Inputs: 0) CSV file name, file ref or RAF ref, 1) database hash ref,
#         2) missing tag value, 3) delimiter if other than ','
# Returns: undef on success, or error string
# Notes: There are various flavours of CSV, but here we assume that only
#        double quotes are escaped, and they are escaped by doubling them
sub ReadCSV($$;$$)
{
    local ($_, $/);
    my ($file, $database, $missingValue, $delim) = @_;
    my ($buff, @tags, $found, $err, $raf, $openedFile);
    if (UNIVERSAL::isa($file, 'File::RandomAccess')) {
        $raf = $file;
        $file = 'CSV file';
    } elsif (ref $file eq 'GLOB') {
        $raf = File::RandomAccess->new($file);
        $file = 'CSV file';
    } else {
        open CSVFILE, $file or return "Error opening CSV file '${file}'";
        binmode CSVFILE;
        $openedFile = 1;
        $raf = File::RandomAccess->new(\*CSVFILE);
    }
    $delim = ',' unless defined $delim;
    # set input record separator by first newline found in the file
    # (safe because first line should contain only tag names)
    while ($raf->Read($buff, 65536)) {
        $buff =~ /(\x0d\x0a|\x0d|\x0a)/ and $/ = $1, last;
    }
    $raf->Seek(0,0);
    while ($raf->ReadLine($buff)) {
        my (@vals, $v, $i, %fileInfo);
        my @toks = split /\Q$delim/, $buff;
        while (@toks) {
            ($v = shift @toks) =~ s/^ +//;  # remove leading spaces
            if ($v =~ s/^"//) {
                # quoted value must end in an odd number of quotes
                while ($v !~ /("+)\s*$/ or not length($1) & 1) {
                    if (@toks) {
                        $v .= $delim . shift @toks;
                    } else {
                        # read another line from the file
                        $raf->ReadLine($buff) or last;
                        @toks = split /\Q$delim/, $buff;
                        last unless @toks;
                        $v .= shift @toks;
                    }
                }
                $v =~ s/"\s*$//;    # remove trailing quote and whitespace
                $v =~ s/""/"/g;     # un-escape quotes
            } else {
                $v =~ s/[ \n\r]+$//;# remove trailing spaces/newlines
            }
            push @vals, $v;
        }
        if (@tags) {
            # save values for each tag
            $fileInfo{_ordered_keys_} = [ ];
            for ($i=0; $i<@vals and $i<@tags; ++$i) {
                # ignore empty entries unless missingValue is empty too
                next unless length $vals[$i] or defined $missingValue and $missingValue eq '';
                # delete tag (set value to undef) if value is same as missing tag
                $fileInfo{$tags[$i]} =
                    (defined $missingValue and $vals[$i] eq $missingValue) ? undef : $vals[$i];
                push @{$fileInfo{_ordered_keys_}}, $tags[$i];
            }
            # figure out the file name to use
            if ($fileInfo{SourceFile}) {
                $$database{$fileInfo{SourceFile}} = \%fileInfo;
                $found = 1;
            }
        } else {
            # the first row should be the tag names
            foreach (@vals) {
                # terminate at first blank tag name (eg. extra comma at end of line)
                last unless length $_;
                @tags or s/^\xef\xbb\xbf//; # remove UTF-8 BOM if it exists
                /^([-_0-9A-Z]+:)*[-_0-9A-Z]+#?$/i or $err = "Invalid tag name '${_}'", last;
                push(@tags, $_);
            }
            last if $err;
            @tags or $err = 'No tags found', last;
            # fix "SourceFile" case if necessary
            $tags[0] = 'SourceFile' if lc $tags[0] eq 'sourcefile';
        }
    }
    close CSVFILE if $openedFile;
    undef $raf;
    $err = 'No SourceFile column' unless $found or $err;
    return $err ? "$err in $file" : undef;
}
#------------------------------------------------------------------------------
# Convert unicode code point to UTF-8
# Inputs: 0) integer Unicode character
# Returns: UTF-8 bytes
sub ToUTF8($)
{
    require Image::ExifTool::Charset;
    return Image::ExifTool::Charset::Recompose(undef, [$_[0]], $charset);
}
#------------------------------------------------------------------------------
# Read JSON object from file
# Inputs: 0) RAF reference or undef, 1) optional scalar reference for data
#            to read before reading from file (ie. the file read buffer)
# Returns: JSON object (scalar, hash ref, or array ref), or undef on EOF or
#          empty object or array (and sets $$buffPt to empty string on EOF)
# Notes: position in buffer is significant
sub ReadJSONObject($;$)
{
    my ($raf, $buffPt) = @_;
    # initialize buffer if necessary
    my ($pos, $readMore, $rtnVal, $tok, $key, $didBOM);
    if ($buffPt) {
        $pos = pos $$buffPt;
        $pos = pos($$buffPt) = 0 unless defined $pos;
    } else {
        my $buff = '';
        $buffPt = \$buff;
        $pos = 0;
    }
Tok: for (;;) {
        # (didn't spend the time to understand how $pos could be undef, but
        #  put a test here to be safe because one user reported this problem)
        last unless defined $pos;
        if ($pos >= length $$buffPt or $readMore) {
            last unless defined $raf;
            # read another 64kB and add to unparsed data
            my $offset = length($$buffPt) - $pos;
            if ($offset) {
                my $buff;
                $raf->Read($buff, 65536) or $$buffPt = '', last;
                $$buffPt = substr($$buffPt, $pos) . $buff;
            } else {
                $raf->Read($$buffPt, 65536) or $$buffPt = '', last;
            }
            unless ($didBOM) {
                $$buffPt =~ s/^\xef\xbb\xbf//;  # remove UTF-8 BOM if it exists
                $didBOM = 1;
            }
            $pos = pos($$buffPt) = 0;
            $readMore = 0;
        }
        unless ($tok) {
            # skip white space and find next character
            $$buffPt =~ /(\S)/g or $pos = length($$buffPt), next;
            $tok = $1;
            $pos = pos $$buffPt;
        }
        # see what type of object this is
        if ($tok eq '{') {      # object (hash)
            $rtnVal = { _ordered_keys_ => [ ] } unless defined $rtnVal;
            for (;;) {
                # read "KEY":"VALUE" pairs
                unless (defined $key) {
                    $key = ReadJSONObject($raf, $buffPt);
                    $pos = pos $$buffPt;
                }
                # ($key may be undef for empty JSON object)
                if (defined $key) {
                    # scan to delimiting ':'
                    $$buffPt =~ /(\S)/g or $readMore = 1, next Tok;
                    $1 eq ':' or return undef;  # error if not a colon
                    my $val = ReadJSONObject($raf, $buffPt);
                    $pos = pos $$buffPt;
                    return undef unless defined $val;
                    $$rtnVal{$key} = $val;
                    push @{$$rtnVal{_ordered_keys_}}, $key;
                    undef $key;
                }
                # scan to delimiting ',' or bounding '}'
                $$buffPt =~ /(\S)/g or $readMore = 1, next Tok;
                last if $1 eq '}';          # check for end of object
                $1 eq ',' or return undef;  # error if not a comma
            }
        } elsif ($tok eq '[') { # array
            $rtnVal = [ ] unless defined $rtnVal;
            for (;;) {
                my $item = ReadJSONObject($raf, $buffPt);
                $pos = pos $$buffPt;
                # ($item may be undef for empty array)
                push @$rtnVal, $item if defined $item;
                # scan to delimiting ',' or bounding ']'
                $$buffPt =~ /(\S)/g or $readMore = 1, next Tok;
                last if $1 eq ']';          # check for end of array
                $1 eq ',' or return undef;  # error if not a comma
            }
        } elsif ($tok eq '"') { # quoted string
            for (;;) {
                $$buffPt =~ /(\\*)"/g or $readMore = 1, next Tok;
                last unless length($1) & 1; # check for escaped quote
            }
            $rtnVal = substr($$buffPt, $pos, pos($$buffPt)-$pos-1);
            # unescape characters
            $rtnVal =~ s/\\u([0-9a-f]{4})/ToUTF8(hex $1)/ige;
            $rtnVal =~ s/\\(.)/$unescapeJSON{$1}||$1/sge;
            # decode base64 (binary data) values
            if ($rtnVal =~ /^base64:[A-Za-z0-9+\/]*={0,2}$/ and length($rtnVal) % 4 == 3) {
                require Image::ExifTool::XMP;
                $rtnVal = ${Image::ExifTool::XMP::DecodeBase64(substr($rtnVal,7))};
            }
        } elsif ($tok eq ']' or $tok eq '}' or $tok eq ',') {
            # return undef for empty object, array, or list item
            # (empty list item actually not valid JSON)
            pos($$buffPt) = pos($$buffPt) - 1;
        } else {                # number, 'true', 'false', 'null'
            $$buffPt =~ /([\s:,\}\]])/g or $readMore = 1, next;
            pos($$buffPt) = pos($$buffPt) - 1;
            $rtnVal = $tok . substr($$buffPt, $pos, pos($$buffPt)-$pos);
        }
        last;
    }
    return $rtnVal;
}
#------------------------------------------------------------------------------
# Read JSON file
# Inputs: 0) JSON file name, file ref, RAF ref or SCALAR ref, 1) database hash ref,
#         2) flag to delete "-" tags, 3) character set
# Returns: undef on success, or error string
sub ReadJSON($$;$$)
{
    local $_;
    my ($file, $database, $missingValue, $chset) = @_;
    my ($raf, $openedFile);
    # initialize character set for converting "\uHHHH" chars
    $charset = $chset || 'UTF8';
    if (UNIVERSAL::isa($file, 'File::RandomAccess')) {
        $raf = $file;
        $file = 'JSON file';
    } elsif (ref $file eq 'GLOB') {
        $raf = File::RandomAccess->new($file);
        $file = 'JSON file';
    } elsif (ref $file eq 'SCALAR') {
        $raf = File::RandomAccess->new($file);
        $file = 'in memory';
    } else {
        open JSONFILE, $file or return "Error opening JSON file '${file}'";
        binmode JSONFILE;
        $openedFile = 1;
        $raf = File::RandomAccess->new(\*JSONFILE);
    }
    my $obj = ReadJSONObject($raf);
    close JSONFILE if $openedFile;
    unless (ref $obj eq 'ARRAY') {
        ref $obj eq 'HASH' or return "Format error in JSON file '${file}'";
        $obj = [ $obj ];
    }
    my ($info, $found);
    foreach $info (@$obj) {
        next unless ref $info eq 'HASH';
        # fix "SourceFile" case, or assume '*' if SourceFile not specified
        unless (defined $$info{SourceFile}) {
            my ($key) = grep /^SourceFile$/i, keys %$info;
            if ($key) {
                $$info{SourceFile} = $$info{$key};
                delete $$info{$key};
            } else {
                $$info{SourceFile} = '*';
            }
        }
        if (defined $missingValue) {
            $$info{$_} eq $missingValue and $$info{$_} = undef foreach keys %$info;
        }
        $$database{$$info{SourceFile}} = $info;
        $found = 1;
    }
    return $found ? undef : "No valid JSON objects in '${file}'";
}
1; # end
__END__
HideShow 67 lines of Pod
=head1 NAME
Image::ExifTool::Import - Import CSV and JSON database files
=head1 SYNOPSIS
    use Image::ExifTool::Import qw(ReadCSV ReadJSON);
    $err = ReadCSV($csvFile, \%database);
    $err = ReadJSON($jsonfile, \%database);
=head1 DESCRIPTION
This module contains routines for importing tag information from CSV (Comma
Separated Value) and JSON (JavaScript Object Notation) database files.
=head1 EXPORTS
Exports nothing by default, but ReadCSV and ReadJSON may be exported.
=head1 METHODS
=head2 ReadCSV / ReadJSON
Read CSV or JSON file into a database hash.
=over 4
=item Inputs:
0) CSV file name or file reference.
1) Hash reference for database object.
2) Optional string used to represent an undefined (missing) tag value. 
(Used for deleting tags.)
3) For ReadCSV this gives the delimiter for CSV entries, with a default of
",".  For ReadJSON this is the character set for converting Unicode escape
sequences in strings, with a default of "UTF8".  See the ExifTool Charset
option for a list of valid character sets.
=item Return Value:
These functions return an error string, or undef on success and populate the
database hash with entries from the CSV or JSON file.  Entries are keyed
based on the SourceFile column of the CSV or JSON information, and are
stored as hash lookups of tag name/value for each SourceFile.  The order
of the keys (CSV column order or order in a JSON object) is stored as an
ARRAY reference in a special "_ordered_keys_" element of this hash.
=back
=head1 AUTHOR
Copyright 2003-2025, Phil Harvey (philharvey66 at gmail.com)
This library is free software; you can redistribute it and/or modify it
under the same terms as Perl itself.
=head1 SEE ALSO
L<Image::ExifTool(3pm)|Image::ExifTool>
=cut