serialization¶

New in version 1.1.0.

Data serialization

Attributes¶

smisk.serialization.serializers¶

The serializer registry.

Type:	`Registry`

Classes¶

class smisk.serialization.data(smisk.serialization.plistlib_.Data)¶

Represent arbitrary bytes.

data¶

Actual storage of the bytes.

Type:	str

__init__(source):: source can be a str or a file-like object with a read method returning str.

encode() → str¶: Instance method to encode data into a printable, base-64 encoded, string.

decode(string)¶: Class method for creating a data object from a base-64 encoded string.

class smisk.serialization.Registry(object)¶

A serializers registry.

Can be accessed like a dictionary:

>>> reg = Registry()
>>> # (register some serializers here)
>>> reg[3]
<class MySerializer ...
>>> MySerializer in reg
True
>>> for serializer in reg:
...   print serializer
<class Serializer1 ...
<class Serializer2 ...

first_in¶

First registered serializer.

Type:	`Serializer`
Value:	None

media_types¶

Media type-to-Serializer map.

Type:	dict
Value:	{}

extensions¶

Filename extension-to-Serializer map.

Type:	dict
Value:	{}

serializers¶

List of available serializers.

Type:	list
Value:	[]

readers¶

Iterate serializers able to read, or unserialize, data.

for ser in reg.readers:
  print ser

Return type:	generator

writers¶

Iterate serializers able to write, or serialize, data.

for ser in reg.writers:
  print ser

Return type:	generator

register(serializer)¶

Register a new Serializer.

serializer should be the class of the serializer to register, not an instance.

unregister(serializer=None)¶: Unregister a previously registered Serializer or all registered serializers, if serializer is None.

find(media_type_or_extension)¶

Find a Serializer associated with a media type or an extension.

Return type:	`Serializer`
Returns:	`None` if not found.

associate(serializer, media_type=None, extension=None, override_existing=True)¶: Associate a Serializer with formats and/or extensions.

class smisk.serialization.Serializer(object)¶

Abstract baseclass for serializers.

All members described here are class members (serializers are never instantiated).

name¶

A human readable short and descriptive name of the serializer.

Type:	str
Value:	“Untitled serializer”

extensions¶

Filename extensions this serializer can handle.

Must contain at least one item. The first item will be used as the primary extension.

Type:	collection
Value:	tuple()

media_types¶

Media types this serializer can handle.

Must contain at least one item. The first item will be used as the primary media type.

Type:	collection
Value:	tuple()

charset¶

Preferred character encoding.

If the value is None the serializer is considered “binary”.

Type:	str
Value:	None

unicode_errors¶

How to handle unicode conversions.

Possible values: "strict", "ignore", "replace", "xmlcharrefreplace", "backslashreplace"

Type:	str
Value:	“strict”

handles_empty_response¶

If enabled, serialize() will be called even when leafs does not generate a response body. (i.e. params=None passed to serialize())

Some serialization formats does not allow empty responses (RPC-variants for instance) in which case this feature come in handy.

Type:	bool
Value:	False

can_serialize¶

Declares where there or not this serializer can write/encode/serialize data.

New in version 1.1.3.

Type:	bool
Value:	False

can_unserialize¶

Declares where there or not this serializer can read/decode/unserialize data.

New in version 1.1.3.

Type:	bool
Value:	False

serialize(params, charset) → tuple¶

Serialize the data structure params with the user/outside-prefered character encoding charset.

If the serializer is binary; the first item in the tuple returned should be None. If the serializer is textual; the character encoding actually used should be the first item of the tuple returned. It is constructed this way to allow serializers only to accept a few character encodings or to be able to enforce a specific one at the same time as some serializers strictly follow the charset requested by the user/outside.

Parameters:	params (dict) – Parameters charset (str) – Destination charset. Might be discarded, so care about the returned charset.
Returns:	Tuple of `(str charset, str data)` where charset is the name of the actual charset used and might be `None` if binary or unknown. data must be a str.
Return type:	tuple

unserialize(file, length=-1, charset=None) → tuple¶

Unserialize bytes representing some kind of data structure.

Parameters:	file (object) – A file-like object implementing at least the read() method length (int) – Number of bytes to read from file or `-1` if unknown. charset (str) – Character encoding of input if the serializer is textual. Might be `None` if unknown in which case the `charset` attribute should be considered.
Returns:	Tuple of `(list args, dict params)`. args and params might be `None`.
Return type:	tuple

serialize_error(status, params, charset) → tuple¶

Encode an error.

Returning None indicates that someone else should handle the error encoding. (Normally this means that smisk.mvc.Application.error() serializes the error using best guess).

“params” will always contain:

code (int) — Error code (i.e. 123).

name (unicode) — Name of the error. i.e. “404 Not Found”

description (unicode) — Description of the error or the emtpy string.

server (unicode) — Short one line description of the server name, port and software.

“params” might contain:

traceback (list) — A list of strings.

Parameters:	status (smisk.mvc.http.Status) – HTTP status params (dict) – Parameters charset (str) – Destination charset. Might be discarded, so care about the returned charset.
Returns:	Tuple of (str charset, str data) where charset is the name of the actual charset used and might be None if binary or unknown.
Return type:	tuple

add_content_type_header(response, charset)¶: Sets "Content-Type" header if missing in response. This method is called by smisk.mvc.Application when completing a HTTP transaction and should not be overridden in subclasses (as it)

directions() → list¶

List of possible directions, or read/write capabilities.

Returns:	`["read", "write"]`, `["read"]`, `["write"]` or `[]`
Return type:	list

did_register(registry)¶

Called when this serializer has been successfully registered in a Registry (the registry instance is passed as the registry argument).

Default implementation does nothing. This is meant to be overridden in subclasses to allow a kind of initialization routine, setting up the serializer if needed.

Parameters:	registry (Registry) – The registry in which this serializer was registered.
Return type:	None

did_unregister(registry)¶

Called when this serializer has been removed from a Registry.

Default implementation does nothing. This is meant to be overridden in subclasses to allow a kind of tear-down routine, finalizing the serializer if needed.