Oba: Converting a Python iterable object to an attribute class

Introduction

Converting iterable objects (such as dictionaries, tuples, lists, sets, and subclasses) to access values using class attributes instead of square brackets []. Supports recursive operations and supports both getting and setting values.

Installation

pip install oba

Usage

from oba import Oba

# Convert a dictionary to an attribute class
d = dict(a=[1, 2, 3], b=[4, dict(x=1)], c=dict(l='hello'))
o = Obj(d)

# Get values
print(o.a[2])  # => 3
print(o.c.l)  # => hello

# Set values
o.b[1].x = 4
print(Obj.raw(o.b[1]))  # => {'x': 4}

points = [dict(x=1, y=2), dict(x=-1, y=0), dict(x=0, y=1)]
o = Obj(points)

o[0].x += 1
print(Obj.raw(o[0]))  # => {'x': 2, 'y': 2}

Comparison

namedtuple (built-in module)

from collections import namedtuple

o = namedtuple('Struct', d.keys())(*d.values())

print(o)  # => Struct(a=[1, 2, 3], b=[4, {'x': 1}], c={'l': 'hello'})
print(o.a)  # => [1, 2, 3]

o.x = 2  # => AttributeError

bunch (package)

from bunch import Bunch

o = Bunch(d)

print(o.a)  # => [1, 2, 3]
print(o.b['x'])  # => 1
print(o.b.x)  # => KeyError

o.x = 2  # OK, editable

o = Bunch(points)  # => AttributeError

json (built-in module)

import json

class Obj(object):
    def __init__(self, d):
        self.__dict__.update(d)

o = json.loads(json.dumps(d), object_hook=Obj)

print(o.a[2])  # => 3
print(o.c.l)  # => hello

o.x = 2  # OK, editable

o = Obj(points)

o[0].x += 1
print(o[0].x)  # => 2

mock (package)

from mock import Mock

o = Mock(**d)

print(o.a)  # => [1, 2, 3]
print(o.b['x'])  # => 1
print(o.b.x)  # => KeyError

o.x = 2  # OK

o = Mock(*points)  # => TypeError

Summary

	namedtuple	bunch (2011)	json	mock (2020)	oba (2022)
built-in	✓	✗ 11K	✓	✗ 28K	✗ 2K
recursive	✗	✗	✓	✗	✓
revert to raw	✗	✓ (with extra operations)	✗	✗	✓
editable	✗	✓	✓	✓	✓
iterable	✓ (no keys)	✓	✗	✗	✓
support dict	✓	✓	✓	✓	✓
support list	✓	✗	✓	✗	✓
support tuple	✗	✗	✓	✗	✓
tolerable	✗	✗	✗	✗	✓

Features

Additionally, Oba also has a unique tolerance for unknown attributes. In cases where some attributes do not exist, for example:

d = dict(a=1)
print(d['b'])  # KeyError

Other libraries will immediately raise an error. However, in some scenarios (such as reading configuration files), the absence of sub-attributes is a common problem, and we hope to be able to tolerate and monitor the existence of such errors.

from oba import Obj

d = dict(a=1)
o = Obj(d)

print('x' in o)  # => False
if not o.x.y.z:  # OK
    print('not exist')  # => not exist
print(o.x.y.z)  # => ValueError: NoneObj (x.y.z)  # locating the non-existent attribute chain

Its internal implementation is that when an attribute does not exist, the object automatically switches to the NoneObj class and records the attribute chain.

License

MIT

Links

• GitHub
• PyPI
• Documentation
• Chinese version