Page MenuHomeHEPForge
Files F7877222

No OneTemporary
Actions

View Options

diff --git a/pyext/professor2/sampling.py b/pyext/professor2/sampling.py
--- a/pyext/professor2/sampling.py
+++ b/pyext/professor2/sampling.py
@@ -1,89 +1,92 @@
# -*- python -*-
# http://stackoverflow.com/questions/3025162/statistics-combinations-in-python
def numCombs(n, k):
"""
n choose k algorithm
"""
from operator import mul
from fractions import Fraction
return int( reduce(mul, (Fraction(n-i, i+1) for i in range(k)), 1) )
def xrandomUniqueCombinations(items, nchoose, howmany=None):
""" Generator-like function for n choose k items """
seencombs = []
# Max number safeguard against infinite loops
maxnum = numCombs(len(items), nchoose)
import random
if howmany is None or howmany > maxnum:
print "Only %i possible combinations"%maxnum
howmany = maxnum
while len(seencombs) < howmany:
temp = random.sample(items, nchoose)
temp.sort()
if not sorted(temp) in seencombs:
seencombs.append(temp)
yield temp
## Define a sampler type
class Sampler(object):
def __init__(self, low, high, bias=None):
self.low = float(low)
self.high = float(high)
self.f, self.invf = None, None
if bias:
## Import clever machinery
try:
import sympy as sp
from sympy.abc import x, y
import numpy as np
except ImportError:
print "Bias functions require SymPy and NumPy to be installed... exiting"
exit(1) #< TODO: don't exit from inside a lib function...
## Make transformation and its inverse
try:
#print bias
f_expr = sp.sympify(bias)
except sp.SympifyError, e:
print "Bias function could not be parsed by SymPy:"
print e
exit(1) #< TODO: don't exit from inside a lib function...
try:
finv_exprs = sp.solve(sp.Eq(y, f_expr), x)
finv_expr = finv_exprs[0]
#print f_expr, finv_exprs
self.f = sp.lambdify(x, f_expr, "numpy")
self.finv = sp.lambdify(y, finv_expr, "numpy")
self.lowf, self.highf = self.f(self.low), self.f(self.high)
except Exception, e:
print "Bias function could not be used/inverted by SymPy:"
print e
exit(1) #< TODO: don't exit from inside a lib function...
def shoot(self):
import random
if not self.f:
## Just uniform sampling between low..high
val = random.uniform(self.low, self.high)
else:
## Uniform sample in transformed space, and transform the result back
valf = random.uniform(self.lowf, self.highf)
val = self.finv(valf)
return val
def __call__(self):
return self.shoot()
+ def __repr__(self):
+ return "<%s with x in %f ... %f>" % (self.__class__.__name__, self.low, self.high)
+
## Test biased sampler machinery if run as main
if __name__ == "__main__":
s = Sampler(1, 10, "exp(x)")
import yoda
h = yoda.Histo1D(20, 0, 10)
for _ in xrange(10000):
h.fill( s() )
yoda.plot(h, "foo.pdf")

File Metadata

Mime Type
text/x-diff
Expires
Tue, Nov 19, 2:55 PM (1 d, 12 h)
Storage Engine
blob
Storage Format
Raw Data
Storage Handle
3804873
Default Alt Text
(3 KB)

Event Timeline