Ability to selectively camelize/underscorize fields

[levic]

Some time ago I looked at using this project for converting JSON responses from underscore case to camelcase but the problem I ran into is that it doesn't distinguish between keys that are values and keys that are field names

For example:

import djangorestframework_camel_case.util as util
data = {
    'first_name': 'Mary',
    'last_name': 'Smith',
    'servers': {
        'server_west.mycompany.com': '...',
        'server_south.mycompany.com': '...',
    },
}
print(json.dumps(util.camelize(data), indent=4))

Output is:

{
    "firstName": "Mary",
    "lastName": "Smith",
    "servers": {
        "serverWest.mycompany.com": "...",
        "serverSouth.mycompany.com": "..."
    }
}

There's no way of indicating that the keys in the servers object shouldn't be modified because it's an associative array where the keys are user-supplied data.

I created an alternate implementation that adds an ignore parameter, so eg you could do the following:

util.camelize(data, ignore=['servers'])

Anything in the tree that matches the ignore specifier will be ignored (it will still recurse into that node but won't change the case on the key, and also handles arrays; ignore=['servers', 'servers.*.user_accounts'] would allow you to correctly transform something like:

{
    "first_name": "Mary",
    "last_name": "Smith",
    "servers": {
        "server_west.mycompany.com": {
            'user_accounts': {
                'mary_smith': {
                    'home_dir': '/home/mary_smith',
                },
            },
        },
    },
}

into

{
    "firstName": "Mary",
    "lastName": "Smith",
    "servers": {
        "server_west.mycompany.com": {
            'userAccounts': {
                'mary_smith': {
                    'homeDir': '/home/mary_smith',
                },
            },
        },
    },
}

If I cleaned up the code is it something that you'd be interesting in accepting a PR for or outside the scope of this project?

[th0th]

@levic can you please share your camelize with ignore implementation? I have the same issue, with a JSONField. However, I am not sure that setting a global ignore pattern would work, the same key might be used in different places,

[asmaps]

I could use a feature like that too, but the other way round, so to selectively not decamelize some fields. Probably this should be defined in the serializer Meta class.

[vbabiy]

@levic I would like to see how you have done this.

[gartmeier]

I could use a feature like that too, but the other way round, so to selectively not decamelize some fields. Probably this should be defined in the serializer Meta class.

I copied the CamelCaseJSONParser class quick and dirty to use a underscoreize function which skips items in api_settings.JSON_UNDERSCOREIZE['skip_keys'].

https://gist.github.com/torfeld6/7d794b26531f0cc03043ff666980fe50

I'd like it to be in the serializer too, but I don't think it can be with a parser. It would have to be a serializer mixin.

[stuart-bradley]

@torfeld6 This was really helpful, I modified your code slightly to handle nesting, and also created a renderer:

• Parser: https://gist.github.com/lutrasdebtra/799faf2c8f7cb6cc80c6e41f3ef9f08e
• Renderer: https://gist.github.com/lutrasdebtra/8be2bca78c61d1426568500503c8983c

[haruanm]

Is there a chance that a feature like that would be added to the project? I can help with something if there is space to it.

[dazigna]

I am facing the same issue, this feature would be such a great addition to the package.

[vbabiy]

I am all for this, we just need to come up with a full PR. With docs and tests around this.

[sergei-maertens]

Am I wrong in thinking this is possible (to some extent) through the json_underscoreize attributes of the parser/renderer classes?