objectutils

Installation

pip install objectutils

About

Tiny functions that extend python json-like objects functionality as highly customizable:

• traversing
• transformation
• flattening
• sum
• diff

operations on json-like python objects(lists, dicts)

Allows writing comprehensions without comprehensions 🙃

Only python 3.10+ supported

Provided as python library and made to be used from python directly.

Inspired by:

• jmespath
• jq

Examples

Having the following response from some API:

obj = {"computers": [
        {
            "computername": "1",
            "software": ["s1", "s2"],
        },
        {
            "computername": "2",
            "software": ["s2", "s3"],
        },
        {
            "computername": "3",
            "software": ["s1", "s3"],
        },
    ]
}

traverse

You should write something like that to get the Counter of the software installed in total:

from itertools import chain

c = Counter(chain.from_iterable([computer["software"] for computer in obj["computers"]]))

Such expressions getting even worse in more complicated cases. With traverse method provided by this tiny lib you should do the following to get the same Counter:

from objectutils import traverse

c = traverse(obj, [Counter, chain.from_iterable, "computers", [], "software"])

traverse supports callable objects in its path, as well as the keys of object. [] considered as all the possible values in iterable, as 'asterisk'(*).

If applicable, calls the funcs and callable objects with unpacked iterable from the right. On exception that was predicted in this case, tries to call with single argument

As for me, it is much clearer approach than writing comprehensions or loops in such cases.

Transformation

It is also possible to transform the data selected by traverse using dicts on the path, but this may be a bit tricky:

traverse(
    {"old1": {"old2": "old3"}, "old4": "old5"}, 
    [
        {"new2": []}, 
        [{"new3": ["old1", "old2"], "new4": ["old4"]}, 
        {"anothernew4": ["old4"]}]
    ]
)

#Result {'new2': [{'new3': 'old3', 'new4': 'old5'}, {'anothernew4': 'old5'}]}

flatten

from objectutils import flatten

Use flatten(obj) to get a flat dictionary in form {path: plain value} For the data above, the result is the following:

{
    ("computers", 0, "computername"): "1",
    ("computers", 0, "software", 0): "s1",
    ("computers", 0, "software", 1): "s2",
    ("computers", 1, "computername"): "2",
    ("computers", 1, "software", 0): "s2",
    ("computers", 1, "software", 1): "s3",
    ("computers", 2, "computername"): "3",
    ("computers", 2, "software", 0): "s1",
    ("computers", 2, "software", 1): "s3",
}

The keys are the paths that may be used in traverse to get the value next to them. However intended usage is to reduce the path somehow, using ".".join() or something like that. The second argument is a function that will be applied to result keys.

zip_dicts

Used to join values of two similar dicts on same paths. May be used to find a diff betweens two dicts, join values as a sum

from objectutils import zip_dicts

d1 = {
    1: {
        2: 3,
        3: 3,
    }
}
d2 = {
    1: {
        2: 4,
        3: 3,
    }
}
zip_dicts(d1, d2, lambda a, b: a+b, lambda a, b: a==b) # {1: {3: 6}} - "find a sum on all same paths where values are equal"

Third argument is a function for two items on the same paths, fourth is the filter for leaf dict values.

Using as diff (default mapping and filter)

d1 = {
    1: {
        3: 3,
        4: 2,
    }
}
d2 = {
    1: {
        3: 3,
        4: 3,
    }
}
zip_dicts(d1, d2) # {1: {4: (2, 3)}} - elements on path [1, 4] not equal(2 and 3 correspondingly)