Support returning indexes of list value in keypaths() utility function

[momocow]

For example,

data = {"items": ["a", {"c": 1}]}

assert benedict(data).keypaths() == [
    "items",
    "items[0]",
    "items[1]",
    "items[1].c"
]

In my use case, I have a regular expression as a filter for keypaths, the filtered keypaths are passed to subset() to make a new filtered dict.

For example,

import re

filter = "items[\d+].name"

data = {"items": [{"name": "a", "value": "b"}, {"name": "c", "value": "d"}]}
b_data = benedict(data)


filtered_keypaths = [kp for kp in b_data.keypaths() if re.fullmatch(filter, kp)]
filtered_data = b_data.subset(filtered_keypaths)

assert filtered_data == {"items": [{"name": "a"}, {"name": "b"}]} 

Just like what map() works in the functional programming.

[fabiocaccamo]

This is really an interesting feature.

I think that subset could be improved to accept also a regex to do exactly what you need with just 1 line of code.

[fabiocaccamo]

@momocow with 0.21.0 version you can retrieve keypaths including indexes:

data = {"items": ["a", {"c": 1}]}
kps = benedict(data).keypaths(indexes=True)
assert kps == [
    "items",
    "items[0]",
    "items[1]",
    "items[1].c"
]