#!/usr/bin/env python
import time
import sys
import os
import concurrent.futures

################################################################################
def analyze(file):
	contents = open(fi).read()
	# somthing will happen here
	return "" #return list of all ids in this file
################################################################################
def workthelist(filelist):
	#iterate over list and analyze
	print("analyzing", len(filelist), "files")
	for f in filelist:
		analyze(os.path.join(folder, f))
	return "" #later this year return the list of all ids in this filelist
################################################################################


# start stopwatch
timeStart = time.perf_counter()

# read folder from commandline
if len(sys.argv)<2:
	print("specify path please.")
	exit(1)
folder = sys.argv[1]

# count number of files
filelist = os.listdir(folder)
print("number of files, predicted: ", len(filelist))

halfcount = round(len(filelist)/2)
firsthalf = filelist[:halfcount]
secondhalf = filelist[halfcount:]

# sum size
sumsize = 0
filecount = 0
for f in filelist:
	filecount += 1
	fi = os.path.join(folder, f)
	fileinfo = os.stat(fi)
	sumsize += fileinfo.st_size
#	analyze(fi)
print("number of files, counted: ", filecount)
print("size summarized: ", round(sumsize/1024/1024), "M")
print("mean size per file: ", round(sumsize/1024/filecount), "K")

# parse each file via regex
# use parallel shit and stuff
executor = concurrent.futures.ProcessPoolExecutor(max_workers=2)
a = executor.submit(workthelist(firsthalf))
b = executor.submit(workthelist(secondhalf))

# save list of matches to new file
# ahhh, not yet
# stop stopwatch
timeEnd = time.perf_counter()
print("Execution time: ", timeEnd-timeStart, "seconds")