EBook::Tools::Unpack - Object class for unpacking e-book files into their component parts and metadata


 use EBook::Tools::Unpack;
 my $unpacker = EBook::Tools::Unpack->new(
    'file'     => $filename,
    'dir'      => $dir,
    'encoding' => $encoding,
    'format'   => $format,
    'raw'      => $raw,
    'author'   => $author,
    'title'    => $title,
    'opffile'  => $opffile,
    'tidy'     => $tidy,
    'nosave'   => $nosave,

or, more simply:

 use EBook::Tools::Unpack;
 my $unpacker = EBook::Tools::Unpack->new('file' => 'mybook.prc');



Instantiates a new Ebook::Tools::Unpack object.


  • file

    The file to unpack. Specifying this is mandatory.

  • dir

    The directory to unpack into. If not specified, defaults to the basename of the file.

  • encoding

    If specified, overrides the encoding to use when unpacking. This is normally detected from the file and does not need to be specified.

    Valid values are '1252' (specifying Windows-1252) and '65001' (specifying UTF-8).

  • htmlconvert

    If set to true, an attempt will be made to convert non-HTML output text to HTML where possible.

  • key

    The decryption key to use if necessary (not yet implemented)

  • keyfile

    The file holding the decryption keys to use if necessary (not yet implemented)

  • language

    If specified, overrides the detected language information.

  • opffile

    The name of the file in which the metadata will be stored. If not specified, defaults to content.opf.

  • raw

    If set true, this forces no corrections to be done on any extracted text and a lot of raw, unparsed, unmodified data to be dumped into the directory along with everything else. It's useful for debugging exactly what was in the file being unpacked, and (when combined with nosave) reducing the time needed to extract parsed data from an ebook container without actually unpacking it.

  • author

    Overrides the detected author name.

  • title

    Overrides the detected title.

  • tidy

    If set to true, the unpacker will run tidy on any HTML output files to convert them to valid XHTML. Be warned that this can occasionally change the formatting, as Tidy isn't very forgiving on certain common tricks (such as empty <pre> elements with style elements) that abuse the standard.

  • nosave

    If set to true, the unpacker will run through all of the unpacking steps except those that actually write to the disk. This is useful for testing, but also (particularly when combined with raw) can be used for extracting parsed data from an ebook container without actually unpacking it.


See "new()" for more details on what some of these mean. Note that some values cannot be autodetected until an unpack method executes.





In scalar context, this is the basename of file. In list context, it actually returns the basename, directory, and extension as per fileparse from File::Basename.





This returns the language specified by the user, if any. It remains undefined if the user has not requested that a language code be set even if a language was autodetected.




This returns the title specified by the user, if any. It remains undefined if the user has not requested a title be set even if a title was autodetected.


This returns a hash containing the autodetected metadata, if any.



Attempts to automatically detect the format of the input file and set the internal object attributes $self->{format} and $self->{formatinfo}, where the former is a one-word string used by the dispatcher to select the correct unpacking method and the latter may contain additional detected information (such as a title or version).

Croaks if detection fails.

In scalar context, returns $self->{format}. In list context, returns the two element list ($self->{format},$self->{formatinfo})

This is automatically called by "new()" if the format argument is not specified.


Detects metadata values from the MOBI EXTH headers retrieved via "unpack_mobi_exth()" and places them into the detected attribute.


This generates an OPF file from detected and specified metadata. It does not honor the nosave flag, and will always write its output.

Normally this is called automatically from inside the unpack methods, but can be called manually after an unpack if the nosave flag was set to write an OPF anyway.

Returns the filename of the OPF file.


  • opffile (optional)

    If specified, this overrides the object attribute opffile, and determines the filename to use for the generated OPF file. If not specified, and the object attribute opffile has somehow been cleared (the attribute is set during "new()"), it will be generated by looking at the textfile argument. If no value can be found, the method croaks. If a value was found somewhere other than the object attribute opffile, then the object attribute is updated to match.

  • textfile (optional)

    The file containing the main text of the document. If specified, the method will attempt to split metadata out of the file and add whatever remains to the manifest of the OPF.

  • mediatype (optional)

    The media type (mime type) of the document specified via textfile. If textfile is not specified, this argument is ignored. If textfile is specified, but


This is a dispatcher for the specific unpacking methods needed to unpack a particular format. Unless you feel a need to override the unpacking method specified or detected during object construction, it is probalby better to call this than the specific unpacking methods.


Unpacks Fictionwise/PeanutPress eReader (-er.pdb) files.


Unpacks SoftBook/GEB/REB/eBookWise (.imp) files.


Unpacks Mobipocket (.prc / .mobi) files.


Unpacks Microsoft Reader (.lit) files


Unpacks PalmDoc / AportisDoc (.pdb) files


Unpacks Zip archives (including ePub files).


  • DRM isn't handled. Infrastructure to support this via an external plug-in module may eventually be built, but it will never become part of the main module for legal reasons.

  • Unit tests are incomplete

  • Documentation is incomplete. Accessors in particular could use some cleaning up.

  • Need to implement setter methods for object attributes

  • Import/extraction/unpacking is currently limited to PalmDoc, Mobipocket, and eReader. Extraction from Microsoft Reader (.lit) and ePub is also eventually planned. Other formats may follow from there.


Zed Pobre <>


Copyright 2008 Zed Pobre

Licensed to the public under the terms of the GNU GPL, version 2