import threading
def concurrent_map(func, data):
"""
Similar to the bultin function map(). But spawn a thread for each argument
and apply `func` concurrently.
Note: unlike map(), we cannot take an iterable argument. `data` should be an
indexable sequence.
"""
N = len(data)
result = [None] * N
# wrapper to dispose the result in the right slot
def task_wrapper(i):
result[i] = func(data[i])
threads = [threading.Thread(target=task_wrapper, args=(i,)) for i in xrange(N)]
for t in threads:
t.start()
for t in threads:
t.join()
return result
