It seems you are losing precision in this line:
pi = pi * Decimal(12)/Decimal(640320**(1.5))
Try using:
pi = pi * Decimal(12)/Decimal(640320**Decimal(1.5))
This happens because even though Python can handle arbitrary scale integers, it doesn’t do so well with floats.
Bonus
A single line implementation using another algorithm (the BBP formula):
from decimal import Decimal, getcontext
getcontext().prec=100
print sum(1/Decimal(16)**k *
(Decimal(4)/(8*k+1) -
Decimal(2)/(8*k+4) -
Decimal(1)/(8*k+5) -
Decimal(1)/(8*k+6)) for k in range(100))