#!/usr/bin/python

# Output sample data for "interesting" entries.
# Entry is deemed "interesting" if is shows up in top-<NUM>
# in any of the provided profiles.

class benchmark:
    def __init__(self):
        self.data = {}
        self.interesting_symbols = {}

import sys

def main(argv):
    import csv
    import optparse
    import collections

    parser = optparse.OptionParser()
    num_interesting = 5
    parser.add_option("-n", "--num", dest="num_interesting", default=5,
                      help="number of interesting enteresting entries")
    parser.add_option("-q", "--quiet",
                      action="store_false", dest="verbose", default=True,
                      help="don't print status messages to stdout")

    (options, args) = parser.parse_args()

    header = []
    output_fieldnames = []
    input_csvs = []
    for arg in args:
        icsv = csv.reader(file(arg, 'r'))
        if not header:
            header = icsv.next()
            output_fieldnames = [header[0], header[1]]
        else:
            header2 = icsv.next()
            assert header2 == header

        csv_id = str.split(arg, ".csv")[0]
        for s in header[2:]:
            output_fieldnames += [csv_id + ":" + s]

        input_csvs.append(icsv)

    benchmarks = collections.OrderedDict()

    for icsv in input_csvs:
        cur_bmk=""
        for line in icsv:
            if cur_bmk != line[0]:
                # Entry for a new benchmark
                cur_interesting = num_interesting
                cur_bmk = line[0]
                if not benchmarks.has_key(cur_bmk):
                    benchmarks[cur_bmk] = benchmark()
            else:
                # Continueing with current benchmark
                cur_interesting -= 1

            symbol = line[1]

            if not benchmarks[cur_bmk].data.has_key(symbol):
                benchmarks[cur_bmk].data[symbol] = {}
            benchmarks[cur_bmk].data[symbol][icsv] = line[2:]

            if cur_interesting > 0:
                if not benchmarks[cur_bmk].interesting_symbols.has_key(symbol):
                    benchmarks[cur_bmk].interesting_symbols[symbol] = 0
                benchmarks[cur_bmk].interesting_symbols[symbol] += cur_interesting

    output_csv = csv.writer(sys.stdout)
    output_csv.writerow(output_fieldnames)

    for cur_bmk in benchmarks:
        for symbol, total_count in sorted(benchmarks[cur_bmk].interesting_symbols.iteritems(), key=lambda (k,v): (v,k), reverse=True):
            row = [cur_bmk, symbol]
            for icsv in input_csvs:
                key = (icsv, cur_bmk, symbol)
                if benchmarks[cur_bmk].data[symbol].has_key(icsv):
                    line2 = benchmarks[cur_bmk].data[symbol][icsv]
                else:
                    line2 = []
                    for s in header[2:]:
                        line2.append("0")

                row += line2

            output_csv.writerow(row)

if __name__ == "__main__":
   main(sys.argv[1:])