avg number of comps per country since 2010 in WCA

# this program and the extract of the wca export must be in the same folder
# to run, just type "python program.py"

# avg number of comps per country since 2010

# sum the number of competitions from 2010 and 2018 and divide by 9 seems booring,
# so we are taking the average of the number of comps/year since 2010.

import csv

country_list = []
comp_year = []
avg_list = []

base_year = 2010
count = 0
discard = 1 # ignore countries with 1 competition

def avg(l):
    out = 0.0
    if len(l)==0: return 0.0
    for x in l:
        out += x
    return 1.0*out/len(l)

with open('WCA_export_Competitions.tsv','rb') as tsvin:
    tsvin = csv.reader(tsvin, delimiter='\t')

    for line in tsvin:

        count += 1
        if count == 1: continue # discard the header. I'm too lazy to search for a better way to do this.

        country = line[3]

        year = int(line[5])

        if year < base_year: continue

        if country not in country_list:
            country_list.append(country)
            comp_year.append([])

        i = country_list.index(country)

        while len(comp_year[i]) < year-base_year+1:
            comp_year[i].append(0)

        comp_year[i][year-base_year] += 1 # 2010 is 0, 2011 is 1, etc. we add 1 for each comp

    for x in comp_year:
        avg_list.append(avg(x))

    prev = 0 # tied stats
    count = 1
    for (x, y, z) in sorted(zip(avg_list, country_list, comp_year))[::-1]:
        if sum(z) <= discard: continue
        print "%3s"%(str(count).zfill(2)+")" if x!=prev else ""), ("%.2f"%x).zfill(5), y, z
        count += 1
        prev = x