ANN: pyprimes 0.1 released

[Steven D'Aprano]

I am happy to announce the first public release of pyprimes, a pure-
Python module for generating prime numbers, primality tests, and prime 
factorisation.

http://pypi.python.org/pypi/pyprimes/0.1.1a

With pyprimes, you can compare a number of algorithms for generating and 
testing primes. It includes fast algorithms for lazily generating primes 
on demand, such as the Croft Spiral, as well as terrible examples of 
algorithms to avoid, such as Turner's Sieve (sometimes known as "the 
Sleight on Eratosthenes").

Examples:

py> import pyprimes 
py> pyprimes.isprime(3300454933)
True
py> pyprimes.factors(3300454934)
[2, 7, 14969, 15749L]
py> for p in pyprimes.primes():
...     if p < 500000: continue
...     if p > 500100: break
...     print p
... 
500009
500029
500041
500057
500069
500083


A note on performance:

Generating large prime numbers (hundreds of digits) is extremely 
computationally intensive. A pure-Python solution like pyprimes is 
unlikely to be suitable for such large numbers, and I make no claim that 
it will break any world-records.

But many popular and common prime number generators can't even deal well 
with *five* digit primes, slowing down to a crawl. pyprimes is in some 
cases hundreds of times faster than such prime generators. pyprimes 
includes examples of these algorithms so you can compare them for 
yourself:

py> from time import time
py> def test(generator):
...     t = time()
...     for p in generator():
...             if p > 10000: break
...     return time() - t
... 
py> test(pyprimes.primes)
0.013000965118408203
py> test(pyprimes.naive_primes1)
4.1489889621734619



-- 
Steven