#!/usr/bin/env python
# By Joshua E Cook, joshua at joshuacook dot com

from __future__ import with_statement
import re

LogFile = "o10kx100.ap"

# size of the read buffer for the file returned by builtin open()
BufferLength = 0x40000 

# mode passed to builtin open()
ReadMode = "rb"


Fuzz = "GET /ongoing/When"

def fuzzy(fuzz, seq):
    """Yields a counted sequences of substrings starting at the beginning of fuzz
    up to the first space character after the end of fuzz.
    """
    len_fuzz = len(fuzz)
    for line in (line for line in seq if fuzz in line):
        start = line.find(fuzz)
        end = line.find(' ',start + len_fuzz)
        yield line[start : end], 1
        
def count(seqn):
    """Sums the number of occurances of unique items in a counted sequence."""
    count = {}
    setdefault = count.setdefault
    for line, n in seqn:
        count[line] = n + setdefault(line,0)
    return count.iteritems()


Pattern = re.compile(r"^GET /ongoing/When/\d\d\dx/(\d\d\d\d/\d\d/\d\d/[^ .]+)$")

def actual(search, seqn):
    """Performs a precise regular expression search on a counted sequence.

    Yields a counted seqence containing the first subgroup of the match.

    Omits unmatched items.
    """            
    for match, n in ((search(item),n) for item,n in seqn):
        if match:
            yield match.group(1),n

with open(LogFile,ReadMode,BufferLength) as log:
    counts = actual(Pattern.search,count(fuzzy(Fuzz,log)))

second = lambda seq: seq[1]
for k, v in sorted(counts,reverse=True,key=second)[:10]:
    print "%s: %d"%(k,v)