MyDict

Version 2.X is a breaking-API step regarding the functions that transforms the MyDict object into/from something else. Those methods have been moved to another module: mydict.jsonify. Everything else remains the same, and Python 2.X is totally discouraged at this point.

A Python dict subclass which tries to act like JavaScript objects, so you can use the dot notation (d.foo) to access members of the object. If the member doesn't exist yet then it's created when you assign a value to it. Brackets notation (d['foo']) is also accepted.

Installation

$ pip install mydict

Examples

Let's give it a try.

d = MyDict()
d.foo = 'bar'

print(d.foo)
# ==> 'bar'

If you try to get the value of a non-existing member then a None value is returned

d = MyDict()
if d.foo is None:
    print('"foo" does not exist yet!')

If that value is "complex" (a dict or another MyDict instance), then it's also recursively transformed into a MyDict object, so you can access it in the same way

d = MyDict()
d.foo = {'bar': 'baz', 'lst': [{'a': 123}]}

print(d.foo.bar)
# ==> 'baz'

print(d.foo.lst[0].a)
# ==> 123

Values in lists are accessed, as you expect, with the brackets notation (d[0]):

d = MyDict()
d.foo = [1, 2, 3]

print(d.foo[2])
# ==> 3

We can instantiate it from a dict of any level of complexity:

d = MyDict({'foo': 'bar', 'baz': [1, 2, {'foo': 'bar', 'baz': 'Hello, world!'}]})

print(d.foo)
# ==> 'bar'

print(d.baz[0])
# ==> 1

print(d.baz[2].foo)
# ==> 'bar'

with keywords in the constructor:

d = MyDict(a=1, b=2.2, c=[1, 2, 3], d=[{'x': 1, 'y': [100, 200, 300]}])
# ...
d.a == 1
d.b == 2.2
d.c[0] == 1
d.d[0].x == 1
d.d[0].y[1] == 200

or both:

d = MyDict({'foo': 'bar'}, baz=123)
# ...
d.foo == 'bar'
d.baz == 123

Please, take into account that keyword initialization has precedence over the dict (first parameter of the constructor):

d = MyDict({'foo': 'bar'}, foo='BAR')
# ...
d.foo == 'BAR'

It's also possible to access members using a path with get or brackets notation (d['...']):

d = MyDict(foo={'bar': 'baz'})
# ...
d['foo.bar'] == 'baz'
d.get('foo.bar') == 'baz'

But when those keys with dots exists in the tree they are accessed using the corresponding key:

d = MyDict({'foo.bar': 'baz'})
# ...
# 'foo.bar' is not interpreted as a path because the key exists
d['foo.bar'] = 'baz'

But there's a particular case, if a dotted key exists and match an existing path, then this ain't work properly, or work in a different way depending on the method of access used, to be correct:

d = MyDict({'foo': {'bar': 'baz'}, 'foo.bar': 'BAZ'})
# ...
d['foo.bar'] = 'BAZ'  # the "dotted field" ('foo.bar') has precedence over the path
d.foo.bar = 'baz'  # it's not possible to detect a "dotted key" using "dot notation"

Personally, I don't see this as a great issue because I generally avoid using dots in keys, like in the previous case.

Transformation

You have at your disposal a couple of functions to retrieve the MyDict object transformed into something else. For the version 2 the original methods (to_json, from_json, get_dict) have been moved to another module: mydict.jsonify.

Types of case

The available types of case are:

• mydict.SNAKE_CASE : snake_case
• mydict.CAMEL_CASE : camelCase
• mydict.PASCAL_CASE : PascalCase

More on this later on.

mydict.jsonify.to_json

Returns the MyDict object as a JSON string (str):

d = MyDict(foo="bar", arr=[1, 2, {"three": 3}])
mydict.jsonify.to_json(d)
# '{"foo": "bar", "arr": [1, 2, {"three": 3}]}'

In addition, it's also possible to handle the case type of the keys inside the object. For example, we can use snake_case in MyDict object and then "export" it with those keys in camelCase. Let's see it in action:

d = MyDict(my_foo='bar', my_arr=[1, 2, {"other_key": 3}])
mydict.jsonify.to_json(d, case_type=mydict.CAMEL_CASE)
# '{"myFoo": "bar", "myArr": [1, 2, {"otherKey": 3}]}'

mydict.jsonify.get_dict

In some occasions you'll need a plain old Python dict representation of the MyDict object, though is a dict subclass:

d = MyDict(foo="bar", arr=[{"one": 1}, {"two": 2}])
mydict.jsonify.get_dict(d)
# {'foo': 'bar', 'arr': [{'one': 1}, {'two': 2}]}

In addition, it's also possible to handle the case type of the keys inside the object, in the same way to_json works. For example, we can use snake_case in MyDict object and then "export" it with those keys in camelCase. Let's see it in action:

d = MyDict(my_foo='bar', my_arr=[1, 2, {"other_key": 3}])
mydict.jsonify.get_dict(d, case_type=mydict.CAMEL_CASE)
# {'myArr': [1, 2, {'otherKey': 3}], 'myFoo': 'bar'}

Initialization from JSON

It's also possible to load a JSON from str, bytes, and file-like objects (with a .read() method) using the function mydict.jsonify.from_json:

d = mydict.jsonify.from_json('{"foo": "bar"}')
# d.foo == 'bar'

d = mydict.jsonify.from_json(b'{"foo": "bar"}')
# d.foo == 'bar'

d = mydict.jsonify.from_json(open('/path/to/file.json', 'r'))
# d = mydict.jsonify.from_json(open('/path/to/file.json', 'rb')) also works

from io import StringIO, BytesIO

s = StringIO()
s.write('{"foo": "bar"}')

d_from_s = mydict.jsonify.from_json(s)
# d_from_s.foo == 'bar'

b = BytesIO()
b.write(b'{"foo": "bar"}')
# b.write('{"foo": "bar"}'.encode('utf8')) is equivalent

d_from_b = mydict.jsonify.from_json(b)
# d_from_b.foo == 'bar'

Please, notice whether the source is string or bytes the result is always string.

In addition, there's also a param case_type in the from_json function. It works in the same way we previously mentioned for to_json and get_dict. For example:

d = mydict.jsonify.from_json('{"myFoo": "bar", "myArr": [1, 2, {"otherKey": 3}]}', case_type=mydict.SNAKE_CASE)
# d.my_foo == 'bar'
# d.my_arr == [1, 2, {'other_key': 3}]
# d.my_arr[2].other_key == 3

Very useful when we collect data from an API which uses camelCase for its keys but we want a more pythonic way for those keys.

The tests passed successfully with Python 3.6. Python 2.X is totally discouraged at this stage of the library. We recommend using Python +3.X

$ pip install pytest
$ pytest mydict -v