NAME

Data::Sofu - Perl extension for Sofu data

Synopsis

        use Data::Sofu;
        %hash=readSofu("file.sofu");
        ...
        writeSofu("file.sofu",\%hash);
        
Or a litte more complex:
        use Data::Sofu qw/packSofu unpackSofu/;
        %hash=readSofu("file.sofu");
        $comments=getSofucomments;
        open fh,">:UTF16-LE","file.sofu";
        writeSofu(\*fh,\$hash,$comments);
        close fh;
        $texta=packSofu($arrayref);
        $texth=packSofu($hashref);
        $arrayref=unpackSofu($texta);
        $arrayhash=unpackSofu($texth);

Synopsis - oo-style

        require Data::Sofu;
        my $sofu=new Sofu;
        %hash=$sofu->read("file.sofu");
        $comments=$sofu->comments;
        $sofu->write("file.sofu",$hashref);
        open fh,">:UTF16-LE",file.sofu";
        $sofu->write(\*fh,$hashref,$comments);
        close fh;
        $texta=$sofu->pack($arrayref);
        $texth=$sofu->pack($hashref);
        $arrayref=$sofu->unpack($texta);
        $arrayhash=$sofu->unpack($texth);

DESCRIPTION

This Module provides the ability to read and write sofu files of the versions 0.1 and 0.2. Visit http://sofu.sf.net for a description about sofu.

It can also read not-so-wellformed sofu files and correct their errors.

Additionally it provides the ability to pack HASHes and ARRAYs to sofu strings and unpack those.

The comments in a sofu file can be preserved if they're saved with $sofu->comment or getSofucomments or if loadFile/load is used.

It also provides a compatibility layer for sofud via Data::Sofu::Object and Data::Sofu->loadFile();

Data::Sofu::Binary provides an experimental interface to Binary Sofu (.bsofu) files and streams.

SYNTAX

This module can either be called using object-orientated notation or using the funtional interface.

Some features are only avaiable when using OO.

FUNCTIONS

getSofucomments()

Gets the comments of the last file read

writeSofu(FILE,DATA,[COMMENTS])

Writes a sofu file with the name FILE.

FILE can be:

A reference to a filehandle with the right encoding set or

a filename or

a reference to a scalar (Data will be read from a scalar)

An existing file of this name will be overwritten.

DATA can be a scalar, a hashref or an arrayref.

The top element of sofu files must be a hash, so any other datatype is converted to {Value=>DATA}. 


        @a=(1,2,3);
        $sofu->write("Test.sofu",\@a);
        %data=$sofu->read("Test.sofu");
        @a=@{$data->{Value}}; # (1,2,3)

COMMENTS is a reference to hash with comments like the one retuned by comments()

readSofu(FILE)

Reads the sofu file FILE and returns a hash with the data.

FILE can be:

A reference to a filehandle with the right encoding set or

a filename or

a reference to a scalar (Data will be read from a scalar)

These methods are not exported by default:

loadSofu(FILE)

Reads a .sofu file and converts it to Sofud compatible objects

FILE can be:

A reference to a filehandle with the right encoding set or

a filename or

a reference to a scalar (Data will be read from a scalar)

Returns a Data::Sofu::Object

packSofu(DATA,[COMMENTS])

Packs DATA to a sofu string.

DATA can be a scalar, a hashref or an arrayref.

This is different from a normal write(), because the lines are NOT indented and there will be placed brackets around the topmost element. (Which is not Sofu 0.2 conform, please use write(\$scalar,$data) instead).

COMMENTS is a reference to hash with comments like the one retuned by comments().

unpackSofu(SOFU STRING)

This function unpacks SOFU STRING and returns a scalar, which can be either a string or a reference to a hash or a reference to an array.

Can read Sofu and SofuML files but not binary Sofu files

Note you can also read packed Data with readSofu(\<packed Data string>):

        my $packed = packSofu($tree,$comments);
        my $tree2 = unpackSofu($packed);
        my $tree3 = readSofu(\$packed); 
        # $tree2 has the same data as $tree3 (and $tree of course)

writeSofuBinary(FILE, DATA, [Comments, [Encoding, [ByteOrder, [SofuMark]]]])

Writes the Data as a binary file.

FILE can be:

A reference to a filehandle with raw encoding set or

a filename or

a reference to a scalar (Data will be read from a scalar)

DATA has to be a reference to a Hash or Data::Sofu::Object

COMMENTS is a reference to hash with comments like the one retuned by comments

More info on the other parameters in Data::Sofu::Binary

To write other Datastructures use this:

        writeSofuBinary("1.sofu",{Value=>$data});

writeSofuML(FILE, DATA, [COMMENTS,[HEADER]])

Writes the Data as an XML file (for postprocessing with XSLT or CSS)

FILE can be:

A reference to a filehandle with some encoding set or

a filename or

a reference to a scalar (Data will be read from a scalar)

DATA has to be a reference to a Hash or Data::Sofu::Object

COMMENTS is a reference to hash with comments like the one retuned by comments, only used when DATA is not a Data::Sofu::Object

HEADER can be an costum file header, (defaults to qq(<?xml version="1.0" encoding="UTF-8" standalone="no"?>\n<!DOCTYPE Sofu SYSTEM "http://sofu.sf.net/Sofu.dtd">\n) );

Default output (when given a filename) is UTF-8.

CLASS-METHODS

loadFile(FILE)

Reads a .sofu file and converts it to Sofud compatible objects.

FILE can be:

A reference to a filehandle with the right encoding set or

a filename or

a reference to a scalar (Data will be read from a scalar)

Returns a Data::Sofu::Object

        my $tree=Data::Sofu->loadFile("1.sofu");
        print $tree->list("Foo")->value(5);
        $tree->list("Foo")->appendElement(new Data::Sofu::Value(8));
        $tree->write("2.sofu");

METHODS (OO)

new()

Creates a new Data::Sofu object.

setIndent(INDENT)

Sets the indent to INDENT. Default indent is ``\t''.

setWarnings( 1/0 )

Enables/Disables sofu syntax warnings.

comments()

Gets/sets the comments of the last file read

write(FILE,DATA,[COMMENTS])

Writes a sofu file with the name FILE.

FILE can be:

A reference to a filehandle with the right encoding set or

a filename or

a reference to a scalar (Data will be read from a scalar)

An existing file of this name will be overwritten.

DATA can be a scalar, a hashref or an arrayref.

The top element of sofu files must be a hash, so any other datatype is converted to {Value=>DATA}. 


        @a=(1,2,3);
        $sofu->write("Test.sofu",\@a);
        %data=$sofu->read("Test.sofu");
        @a=@{$data->{Value}}; # (1,2,3)

COMMENTS is a reference to hash with comments like the one retuned by comments()

read(FILE)

Reads the sofu file FILE and returns a hash with the data.

FILE can be:

A reference to a filehandle with the right encoding set or

a filename or

a reference to a scalar (Data will be read from a scalar)

pack(DATA,[COMMENTS])

Packs DATA to a sofu string.

DATA can be a scalar, a hashref or an arrayref.

COMMENTS is a reference to hash with comments like the one retuned by comments

This is different from a normal write(), because the lines are NOT indented and there will be placed brackets around the topmost element. (Which is not Sofu 0.2 conform, please use write(\$scalar,$data) instead).

unpack(SOFU STRING)

This function unpacks SOFU STRING and returns a scalar, which can be either a string or a reference to a hash or a reference to an array.

load(FILE)

Reads a .sofu file and converts it to Sofud compatible objects

FILE can be:

A reference to a filehandle with the right encoding set or

a filename or

a reference to a scalar (Data will be read from a scalar)

Returns a Data::Sofu::Object

toObjects(DATA, [COMMENTS])

Builds a Sofu Object Tree from a perl data structure

DATA can be a scalar, a hashref or an arrayref.

COMMENTS is a reference to hash with comments like the one retuned by comments

Returns a Data::Sofu::Object

writeBinary(FILE, DATA, [Comments, [Encoding, [ByteOrder, [SofuMark]]]])

Writes the Data as a binary file.

FILE can be:

A reference to a filehandle with raw encoding set or

a filename or

a reference to a scalar (Data will be read from a scalar)

DATA has to be a reference to a Hash or Data::Sofu::Object

COMMENTS is a reference to hash with comments like the one retuned by comments

More info on the other parameters in Data::Sofu::Binary

To write other Datastructures use this:

        $sofu->writeBinary("1.sofu",{Value=>$data});

writeML(FILE, DATA, [COMMENTS,[HEADER]])

Writes the Data as an XML file (for postprocessing with XSLT or CSS)

FILE can be:

A reference to a filehandle with some encoding set or

a filename or

a reference to a scalar (Data will be read from a scalar)

DATA has to be a reference to a Hash or Data::Sofu::Object

COMMENTS is a reference to hash with comments like the one retuned by comments, only used when DATA is not a Data::Sofu::Object

HEADER can be an costum file header, (defaults to qq(<?xml version="1.0" encoding="UTF-8" standalone="no"?>\n<!DOCTYPE Sofu SYSTEM "http://sofu.sf.net/Sofu.dtd">\n) );

Default output (when given a filename) is UTF-8.

packML (DATA, COMMENTS,[HEADER])

Returns DATA as an XML file (for postprocessing with XSLT or CSS) with no Indentation

DATA has to be a reference to a Hash or Data::Sofu::Object

COMMENTS is a reference to hash with comments like the one retuned by comments, only used when DATA is not a Data::Sofu::Object

HEADER can be an costum file header, (defaults to qq(<?xml version="1.0" encoding="UTF-8" standalone="no"?>\n<!DOCTYPE Sofu SYSTEM "http://sofu.sf.net/Sofu.dtd">\n) );

Those are not (quite) the same:

        $string = $sofu->packML($tree,$comments) #Will not indent.
        $sofu->writeML(\$string,$tree,$comments)# Will indent.

CHANGES

Kyes are now automatically escaped according to the new sofu specification.

Double used references will now be converted to Sofu-References.

read, load, readSofu, loadSofu and Data::Sofu::loaFile now detect binary sofu (and load Data::Sofu::Binary)

read, load, readSofu, loadSofu, Data::Sofu::loaFile, unpackSofu and unpack detect SofuML (and load Data::Sofu::SofuML)

BUGS

Comments written after an object will be rewritten at the top of an object:

        foo = { # Comment1
                Bar = "Baz"
        } # Comment2

will get to:

        foo = { # Comment1
        # Comment 2
                Bar = "Baz"
        }

NOTE on Unicode

Sofu File are normally written in a Unicode format. Data::Sofu is trying to guess which format to read (usually works, thanks to Encode::Guess).

On the other hand the output defaults to UTF-16 (UNIX) (like SofuD). If you need other encoding you will have to prepare the filehandle yourself and give it to the write() funktions...

        open my $fh,">:encoding(latin1)","out.sofu";
        writeSofu($fh,$data);

Warning: UTF32 BE is not supported without BOM (looks too much like Binary);

Notes:

As for Encodings under Windows you should always have a :raw a first layer, but to make them compatible with Windows programs you will have to access special tricks:

        open my $fh,">:raw:encoding(UTF-16):crlf:utf8","out.sofu" #Write Windows UTF-16 Files
        open my $fh,">:raw:encoding(UTF-16)","out.sofu" #Write Unix UTF-16 Files
        #Same goes for UTF32
        
        #UTF-8: Don't use :utf8 or :raw:utf8 alone here, 
        #Perl has a different understanding of utf8 and UTF-8 (utf8 allows some errors).
        open my $fh,">:raw:encoding(UTF-8)","out.sofu" #Unix style UTF-8 
        open my $fh,">:raw:encoding(UTF-8):crlf:utf8","out.sofu" #Windows style UTF-8
        #And right after open():
        print $fh chr(65279); #Print UTF-8 Byte Order Mark (Some programs want it, some programs die on it...)
        
One last thing:
        open my $out,">:raw:encoding(UTF-16BE):crlf:utf8","out.sofu";
        print $out chr(65279); #Byte Order Mark
        #Now you can write out UTF16 with BOM in BigEndian (even if you machine in Little Endian)

SEE ALSO

perl(1),http://sofu.sf.net

For Sofud compatible Object Notation: the Data::Sofu::Object manpage

For Sofu Binary: the Data::Sofu::Binary manpage

For SofuML the Data::Sofu::SofuML manpage