用 python 做了一个简单 dict 封装成 document，大家帮助 review 一下。：）(修改成 gist）

[openroc]

主要目的针对dict少写[]和“”，如 a["test"]["aaa"], 可以简化成a.test.aaa

	#!/user/bin/python
	import json
	class DocumentBase(object):
	def traversal_for_init_dict(self, items):
	for k in items:
	if type(items[k]) == dict:
	setattr(self, k, DocumentDict(**items[k]))
	elif type(items[k]) == list:
	setattr(self, k, self.traversal_for_init_list(items[k]))
	else:
	setattr(self, k, items[k])
	return self.__dict__
	def traversal_for_init_list(self, items):
	rc = []
	for n in xrange(len(items)):
	if type(items[n]) == list:
	rc.append(self.traversal_for_init_list(items[n]))
	elif type(items[n]) == dict:
	rc.append(DocumentDict(**items[n]))
	else:
	rc.append(items[n])
	return rc
	def traversal_to_dict(self, rc, items):
	for k in items:
	if type(items[k]) == DocumentDict:
	rc[k] = items[k].to_dict()
	elif type(items[k]) == list:
	rc[k] = self.traversal_to_list(items[k])
	else:
	rc[k] = items[k]
	def traversal_to_list(self, items):
	rc = []
	for k in xrange(len(items)):
	if type(items[k]) == list:
	rc.append(self.traversal_to_list(items[k]))
	elif type(items[k]) == DocumentDict:
	rc.append(items[k].to_dict())
	else:
	rc.append(items[k])
	return rc
	class DocumentDict(DocumentBase):
	def __init__(self, **kwargs):
	self.traversal_for_init_dict(kwargs)
	def __setitem__(self, key, value):
	setattr(self, key, value)
	def __getitem__(self, key):
	return getattr(self, key)
	def to_dict(self):
	rc = {}
	self.traversal_to_dict(rc, self.__dict__)
	return rc
	class DocumentList(DocumentBase):
	def __init__(self, items):
	if type(items) != list:
	raise Exception("not list: %s" % str(items))
	self.__list__ = self.traversal_for_init_list(items)
	def __getitem__(self, index):
	return self.__list__[index]
	def to_dict(self):
	return self.traversal_to_list(self.__list__)
	if __name__ == '__main__':
	print 'test...'
	test_case = {
	"a": 1,
	"b": 12.0,
	"c": 'cccc',
	"d": { "d1": 3333, "d2": "xxx"},
	"e": [
	{"e1": 6666},
	{"e2": 8888, "e3": { "e31":-1, "e32":[ "ss1", "ss2"] } },
	{"e4": [100, 200, 300] }
	],
	"f": [ [0, 1, 3], ['a', 'b'], [ {"f1":1}, {"f2": 2} ]]
	}
	test_case_list = [ test_case, test_case, test_case]
	print json.dumps(test_case)
	print json.dumps(test_case_list)
	print 100, '='*80
	dd = DocumentDict()
	dd2 = DocumentDict()
	dd2.aaa_x=1
	dd2.bbb_y=2
	dd.a = 1
	dd.b = '2'
	dd.c = []
	dd.d = dd2
	dd.c.append(dd2)
	dd.c.append(dd2)
	dd.c.append(dd2)
	#print dd.a, dd.b, dd.c, dd.d
	print dd.to_dict()
	print 111, '='*80
	dd3 = DocumentDict(**test_case)
	print dd3.to_dict()
	print 222, '='*80
	print dd3.a, dd3.b, dd3.d, dd3.e, dd3.d.d1, dd3.d.d2
	print dd3.e[0].e1
	print dd3.e[1].e2, dd3.e[1].e3.e31, dd3.e[1].e3.e32, dd3.e[1].e3.e32[0]
	print dd3.e[2].e4, dd3.e[2].e4[0]
	print dd3.f[0], dd3.f[2][0].f1, dd3.f[2][1].f2
	print 333, '='*80
	ddl = DocumentList(test_case_list)
	print ddl.to_dict()
	print ddl[0].e[0].e1, ddl[1].e[1].e2
	DocumentDict()
	print 'done'

view raw DocumentDict.py hosted with ❤ by GitHub

[openroc]

paste a gist,

https://gist.github.com/openroc/8c42550c7a31bab5c338

[binux]

class Obict(dict):
....__getattr__ = dict.__getitem__
....__setattr__ = dict.__setitem__

[openroc]

@blnux, 要支持list和dict，嵌套。：）

[lifemaxer]

@binux 又见牛人！上次的“uuid.uuid1()”

[binux]

@openroc 嵌套就是 return 的时候 Obict 包一下就完了，list 同理
而且，建议扩展基础类型的时候从 dict， 或者 UserDict 继承，保证类型兼容性。不然你让 keys, items, values, for loop 怎么办？

[openroc]

@binux, 嗯，类型兼容的确没有考虑，谢过。

[clino]

这个有类似的作用:
https://code.google.com/p/ulipad/source/browse/trunk/modules/dict4ini.py
不过是用在ini文件上

d = DictIni('t2.ini', format="%s:%s", hideData=True)
d.a.a = 'mama'
d.a.b = 'lubs me!'

d.b.a = 'i'
d.b.b = 'lub bosunmi!'

d.c.a = 'dada'
d.c.b = 'lubs me too!'
d.save()

[chuangbo]

赞同 @binux，继承 dict 是必须的。
web.py 的 Storage 是一个比较完整的实现。
https://github.com/webpy/webpy/blob/master/web/utils.py#L52