shed_summarize

#!/usr/bin/env python

from collections import defaultdict
import re
import sys

import scipy.stats as stats
import yaml

if __name__ == '__main__':
    data = defaultdict(list)
    pattern = re.compile(r'^\w+:')
    for f in sys.argv[1:]:
        with open(f) as f:
            y = ''
            for line in f:
                if pattern.match(line):
                    y += line
            d = yaml.load(y)
            for k, v in d.items():
                data[k].append(v)

    o = {}
    for k, v in data.items():
        d = stats.describe(v)
        o[k] = {
            'mean': float(d.mean),
            'stddev': float(d.variance**(0.5)),
        }
    print(yaml.dump(o))