#!/usr/bin/python
# -*- coding: utf8 -*-

import numpy as np
import pylab as pl
import urllib
import StringIO
import os
import re
from lxml import etree

website = 'http://results.scc-events.com/2014/?event=MAL&pid=list&num_results=100'

def fetch_times_from_page(pagename):
    page = urllib.urlopen(pagename)
    page_text = page.read()
    page_xml = etree.parse(StringIO.StringIO(page_text), etree.HTMLParser())
    
    times_txt = []
    for element in page_xml.iter('tr'):
        if len(element) == 10:
            t_text = element[8].text
            match = re.match('\A(\d{2}):(\d{2}):(\d{2})\Z', t_text)
            if match == None:
                continue
            hms = match.groups()
            if len(hms) == 3:
                times_txt.append(t_text)
    return times_txt

def fetch_list(pagename):
    print 'fetching data from website'
    times_txt = []
    i = 0
    while True:
        print 'fetching set', i
        new_times = fetch_times_from_page(pagename + '&page=' + str(i))
        if len(new_times) == 0:
            print 'no more new entries on page', i
            break
        else:
            times_txt = times_txt + new_times
        i += 1
    return times_txt

times = {}
for gender in ['M', 'W']:

    fname = 'marathon40_' + gender + '.txt'
    if not os.path.isfile(fname):
        # have no local file -> need to download data
        pagename = website + '&search[se' + 'x]=' + gender
        times_list = fetch_list(pagename)
        f = open(fname, 'w')
        f.write('\n'.join(times_list))
        f.close()
    
    f = open(fname, 'r')
    times[gender] = []
    for line in f.readlines():
        match = re.match('\A(\d{2}):(\d{2}):(\d{2})\Z', line.strip())
        if match == None:
            continue
        hms = match.groups()
        if len(hms) == 3:
            t = float(hms[0]) + float(hms[1]) / 60. + float(hms[2]) / 3600.
            times[gender].append(t)
    f.close()

pl.figure(figsize=(50/9., 35/9.))
pl.rc('path', simplify_threshold=0.5)
pl.plot(sorted(times['M']), np.linspace(0, 1, len(times['M'])),
    'b-', label='m', linewidth=3)
pl.plot(sorted(times['W']), np.linspace(0, 1, len(times['W'])),
    'r-', label='w', linewidth=3)
pl.xlim(2, 7)
pl.xlabel('t [h]')
pl.ylabel('F')
pl.title('41. Berlin Marathon running times')
pl.grid(True)
pl.legend(loc='center right')
pl.tight_layout()
pl.savefig('running_times_41Berlin_marathon_cumulative.svg')