

Marshalling

Marshalling converts Python data (int, str, float, list, dict, data classes, named tuples) into a format that can be send across the wire. The user can request a particular format to be returned using the Accept request HTTP header:

• application/json: accept JSON output
• application/xml: accept XML output
• text/csv: accept CSV output
• */*: accept anything

Default supported types

By default, pytcher accepts application/json and returns a JSON format output.

Out of the box, it can marshall the followings:

Type	Description	Example	Output
dict	dictionary	{'key': 'value'}	{"key":"value}
list	list	['abc', 'cde']	["abc","cde"]
namedtuple	named tuples	Item('apple', 1.25)	{"name":"apple","price":1.25}
@dataclass	data class	Item('apple', 1.25)	{"name":"apple","price":1.25}
enum	enum	Color.RED	"RED"
str	string	'apple'	"apple"
int	integer	5	5
float	float	5.12	5.12
date	date	datetime.date(2019, 2, 3)	"2019-02-03"
datetime	datetime (no timezone)	datetime.datetime(2019, 2, 3, 12, 32, 1)	"2019-02-03T12:32:01.000"
datetime	datetime (with timezone)	datetime.datetime(2019, 2, 3, 12, 32, 1, tzinfo=pytz.UTC)	"2019-02-03T12:32:01.000+00:00"

A namedtuple can be defined as follows:

from collections import namedtuple

Item = namedtuple('Item', ['name', 'price'])

apple = Item('apple', 1.25)

A dataclass can be defined as follows:

from dataclasses import dataclass

@dataclass
class Item(object):
    name: str
    price: float

apple = Item('apple', 1.25)    

An enum can be defined as follows:

from enum import Enum
class Color(Enum):
    RED = 1
    GREEN = 2

red = Color.RED

Additional types

One can also add additional types to support as follows:

TODO

Adding new accept formats