如何获取没有转义符的正确JSON，之前使用fileInfo存储的？

Question

我无法获取我存储在Maya的fileInfo中的确切 JSON 字符串（json.dumps）。

>>> import pymel.core as pc
>>> json.dumps({'foo': 123})
'{"foo": 123}'
>>> pc.fileInfo['foo'] = json.dumps({'foo': 123})
>>> pc.fileInfo['foo']
u'{\\"foo\\": 123}'
>>> json.loads(pc.fileInfo['foo']) # this produces an error because of the escape sequence
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "c:\Program Files\Autodesk\Maya2011\bin\python26.zip\json\__init__.py", line 307, in loads

  File "c:\Program Files\Autodesk\Maya2011\bin\python26.zip\json\decoder.py", line 319, in decode
  File "c:\Program Files\Autodesk\Maya2011\bin\python26.zip\json\decoder.py", line 336, in raw_decode
  File "c:\Program Files\Autodesk\Maya2011\bin\python26.zip\json\scanner.py", line 55, in iterscan
  File "c:\Program Files\Autodesk\Maya2011\bin\python26.zip\json\decoder.py", line 171, in JSONObject
ValueError: Expecting property name: line 1 column 1 (char 1)

问题在于，当Maya存储一个值时，它会确保这个值被正确地转义。但当我试图取出这个字符串时，它却变得不符合json.loads的格式。我该如何解决这个问题呢？

我基本上想要的是一种方法，可以取消转义 MEL在保存值时引入的转义序列？到目前为止，我还没有找到任何可以做到这一点的东西。

MEL和C语言有很多相似之处，所以它的转义序列基本上和C语言的一样，如上所示。

Answer 1

我遇到了同样的问题。当你从fileInfo命令中获取字符串后，只需像这样替换掉转义的引号：

# Lets say you stored some json data in a fileInfo called 'tester'
import maya.cmds as cmds
import json

s = cmds.fileInfo( 'tester', q = True )[ 0 ]
s = s.replace( '\\"', '"' ) # Handle escaped quotes
j = json.loads( s )
print( json.dumps( j, indent = 4 ) )

Answer 2

确保你在文件信息中存取的数据保持完整，最好的办法就是在存入时进行base64编码，取出时再解码。Maya不会自动帮你处理这些，而且它提供的转义方式，如你所见，是有点奇怪的。

下面是我使用的一个方法；它使用了yaml格式，但你可以轻松换成json格式。我用这个方法来保存yaml数据，正好符合你的需求。

'''
mayaPersist - namespace for functions related to storing data in fileInfo objects inside the current Maya file

I've tested this (a little) with as many as 100,000 integers - it works but it's slooow at that size
< 1000 values seems imperceptible
'''

import yaml
import base64
from  maya.cmds import fileInfo
import itertools

def save(key, value):
    '''
    save the specified value as a base64 encoded yaml dunp at key 'key'
    '''
    encoded =encode(value)
    fileInfo(key, encoded)

def load(key):
    '''
    return the value stored at 'key', or None if the value can't be found

    @note it is possible to store a 'None' in the value, so this doesn't prove that the key does not exist !
    '''
    answer = fileInfo(key, q=True)
    if not answer:
        return None
    return decode(answer[0])

def exists(key):
    '''
    returns true if the specified key exists
    '''
    answer = fileInfo(key, q=True)
    return len(answer) != 0

def ls():
    '''
    a generator that returns all of the key-value pairs in this file's fileInfo

    @note:  these are not decoded, because they contain a mix of native stirngs and b64 values
    '''
    all_values = fileInfo(q=True)
    keys = itertools.islice(all_values, 0, None, 2)
    values = itertools.islice(all_values, 1, None, 2)
    return itertools.izip(keys, values)


def delete(key):
    '''
    remove the key and any data stored with it from this file
    '''
    fileInfo(rm=key)

def decode(value):
    '''
    convert a base64'ed yaml object back into a maya object

    if the object is not encoded (eg, one of the default string values) return it untouched
    '''
    try:
        val = base64.b64decode(value)
        return yaml.load(val)
    except TypeError:  
        return value



def encode (value):
    '''
    return the supplied value encoded into base64-packed YAML dump
    '''
    return  base64.b64encode(yaml.dump(value))

Answer 3

@theodox 的回答可能是对的。不过，如果你知道这个 json 字符串会被转义的话，你可以很简单地把它还原成原来的未转义版本，方法是：

mystring.decode('unicode_escape')

所以在上面的情况下，正确的做法是把存储的信息反序列化为：

json.loads(pc.fileInfo['foo'].decode('unicode_escape'))

不过这样看起来有点丑，所以你可以把它放在一个函数里。

def safe_json_loads(str):
    return json.loads(str.decode('unicode_escape'))


data = safe_json_loads(pc.fileInfo['foo'])

祝好，