#!/usr/bin/env python
import os
import re
from subprocess import Popen, PIPE
import time


NUM_PARALLEL_PYLINT_JOBS = 8
PYLINT_PASS_THRESHOLD = 10
NUM_FILES = [10, 25, 50, 100, 150, 200, 400]


def find_all_python_files(num_files_to_find):
    python_files = []
    for root, dirs, files in os.walk('.'):
        for file in files:
            if file.endswith('.py'):
                if len(python_files) < num_files_to_find:
                    python_files.append(os.path.join(root, file))
                else:
                    break
    return python_files


def run_pylint_on_files(py_files):
    # input: list of filepaths
    py_files = " ".join(py_files)

    pylint = Popen(
        ("pylint -j %d -f text %s" % (NUM_PARALLEL_PYLINT_JOBS, py_files)).split(),
        stdout=PIPE)
    pylint.wait()
    output = pylint.stdout.read()

    result_re = re.compile(r"Your code has been rated at ([\d\.]+)/10")
    result = float(result_re.findall(output)[0])

    if result < PYLINT_PASS_THRESHOLD:
        print 'git: fatal: commit failed, Pylint tests failing.'
    else:
        print 'all good; pylint passed'


def profile_pylint(filesets, iterations=3):
    times_spent = {}
    for fs in filesets:
        sum_times_spent = 0.0
        for i in range(iterations):
            start_time = time.time()
            run_pylint_on_files(fs)
            end_time = time.time()
            sum_times_spent += (end_time - start_time)
        avg_time_spent = sum_times_spent / iterations
        times_spent[len(fs)] = avg_time_spent
    return times_spent


filesets = []
for n_files in NUM_FILES:
    files = find_all_python_files(num_files_to_find=n_files)
    filesets.append(files)

times_spent = profile_pylint(filesets=filesets, iterations=3)
for t in times_spent:
    print '# Files: {} \t Avg. Seconds Spent: {}'.format(t, times_spent[t])