Add "describe" to output kmer sizes, molecules, scaled, and num hashes for each signature

[olgabot]

Hello! I find myself forgetting exactly the parameters I used for each signature and then get frustrated/confused when I sourmash compare and then there's no compatible kmer sizes or molecules. So this feature would let the user know what kinds of signatures are inside each signature.

Here's what I hacked up. I know this uses json instead of ijson but my brain got so turned around trying to use the iterator so I used json for now. I also couldn't figure out how to access the 'scaled' option.

import itertools
import json


keys_for_length = ('mins',)

keys_for_values = ('ksize', 'molecule')

keys_to_print = keys_for_length + keys_for_values


def _describe_single(signature):
    data = []
    for sig in signature['signatures']:
        d = dict(name=signature['name'])
        for key, value in sig.items():
            if key not in keys_to_print:
                continue

            if key in keys_for_length:
                v = len(value)
                k = 'n_' + key
            elif key in keys_for_values:
                v = value
                k = key
            d[k] = v
        data.append(d)
    return data


def describe(filename):
    with open(filename) as f:
        signature = json.load(f)
    data = itertools.chain(*map(_describe_single, signature))

    description = pd.DataFrame(list(data))
    return description

This outputs the following pandas dataframe:

[olgabot]

I know pandas isn't a dependency so we can use the csv tool instead.

[ctb]

beautiful. yes, we should add this as a basic command line option I think!

[taylorreiter]

I was just thinking that something like this would be very useful as I ran head on a signature and had to scroll back up a million lines in my terminal to see the information I wanted. Thank you @olgabot!

[ctb]

Added as a command line feature, 'sourmash sig info', in #587. Outputs a CSV which can be used as above.