class HexaPDF::Type::FileSpecification


Represents a file specification dictionary.

File specifications are used to refer to other files or URLs from within a PDF file. Simple file specifications are just strings. However, the are automatically converted on access to a full file specification to provide a unified interface.

Working with File Specifications

A file specification may refer to a file or an URL. This can easily be checked with url?. Independent of whether the file specification referes to an URL or a file, the path method returns the “best” useable path for it.

Modifying a file specification should be done via the path= and url= methods as they ensure that no obsolescent entries are used and the file specification is consistent.

Finally, since embedded files in a PDF document are always linked to a file specification it is useful to provide embedding/unembedding operations in this class, see embed and unembed.

See: PDF2.0 s7.11

Field Definitions

NameType/Allowed ValuesRequiredDefault Value
FHexaPDF::DictionaryFields::PDFByteString or Stringfalsenil
DOSHexaPDF::DictionaryFields::PDFByteString or Stringfalsenil
MacHexaPDF::DictionaryFields::PDFByteString or Stringfalsenil
UnixHexaPDF::DictionaryFields::PDFByteString or Stringfalsenil
IDHexaPDF::PDFArray or Arrayfalsenil
VTrueClass or FalseClassfalsenil
EFHexaPDF::Type::FileSpecification::EFDictionary or Hashfalsenil
RFHexaPDF::Dictionary or Hashfalsenil
CIHexaPDF::Dictionary or Hashfalsenil

Public Instance Methods

embed(filename, name: File.basename(filename), mime_type: nil, register: true) → ef_stream
embed(io, name:, mime_type: nil, register: true) → ef_stream

Embeds the given file or IO stream into the PDF file, sets the path and MIME type accordingly and returns the created stream object.

If a file is given, the name option defaults to the basename of the file. However, if an IO object is given, the name argument is mandatory.

If there already was a file embedded for this file specification, it is unembedded first.

The embedded file stream automatically uses the FlateEncode filter for compressing the embedded file.



The name that should be used as path value and when registering.


Optionally specifies the MIME type of the file.


Specifies whether the embedded file will be added to the EmbeddedFiles name tree under the name. If the name is already taken, it’s value is overwritten.

The file has to be available until the PDF document gets written because reading and writing is done lazily.


Returns true if this file specification contains an embedded file.

See: embedded_file_stream


Returns the embedded file associated with this file specification, or nil if this file specification references no embedded file.

If there are multiple possible embedded files, the /EF fields are searched in the following order and the first one with a value is used: /UF, /F, /Unix, /Mac, /DOS.


Returns the path for the referenced file or URL. An empty string is returned if no file specification string is set.

If multiple file specification strings are available, the fields are search in the following order and the first one with a value is used: /UF, /F, /Unix, /Mac, /DOS.

The encoding of the returned path string is either UTF-8 (for /UF) or BINARY (for /F /Unix, /Mac and /DOS).


Sets the file specification string to the given filename.

Since the /Unix, /Mac and /DOS fields are obsolescent, only the /F and /UF fields are set.


Deletes any embedded file streams associated with this file specification. A possible entry in the EmbeddedFiles name tree is also deleted.


Sets the file specification string to the given URL and updates the file system entry appropriately.

The provided URL needs to be in an RFC1738 compliant string representation. If not, an error is raised.


Returns true if this file specification references an URL and not a file.