Better representation of changes in JSON structures in audit logs

[nhoening]

Right now, the string representation of changes can become pretty long (showing both whole before and after dicts), for example.

Updated Field: sensors_to_show, From: [{'title': 'Prices', 'sensors': [58, 70]}, 16, {'title': 'Battttery activity', 'sensors ... e': 'Battttery activity', 'sensors': [5, 4]}, {'title': 'No Title', 'sensors': [15]}, {'title': 'No Title 5', 'sensors': []}]

Note the "..." part, as we have limited space, strings are cut to fit 255 characters.

We have some code in another project using dictdiffer, which we can use for all two existing structures (sensors_to_show, flex_context) and of course later also flex_model:

from dictdiffer import diff

def pretty_diff(result) -> list[str]:
    string = []
    for change_type, key, value in result:
        if isinstance(key, list):
            _key = key[0]
            for k in key[1:]:
                _key += f"[{k}]" if isinstance(k, int) else f".{k}"
            key = _key
        if change_type == "change":
            string.append(f"change {key}: {value[0]} -> {value[1]}")
        elif change_type == "add":
            for v in value:
                key += f"[{v[0]}]" if isinstance(v[0], int) else f".{v[0]}"
                string.append(f"add {key}: {v[1]}")
        elif change_type == "remove":
            for v in value:
                if key:
                    key += f"[{v[0]}]" if isinstance(v[0], int) else f".{v[0]}"
                    string.append(f"remove {key}: {v[1]}")
                else:
                    string.append(f"remove {v[0]}: {v[1]}")
        else:
            string.append(f"{change_type} {key}: {value}")
    return string


def enumerated_dict_diff(d1: dict, d2: dict) -> str:
    result = diff(d1, d2)
    string = ""
    for index, item in enumerate(pretty_diff(result), start=1):
        string += f"{index}. {item}\n"
    return string

This hopefully will look informative, at least if there are not too many changes at once.