Implement functions from np.random

unumpy/__init__.py

@@ -151,6 +151,7 @@
 True
 """
 from ._multimethods import *
+from .random import *
 
 from ._version import get_versions
 

unumpy/numpy_backend.py

@@ -1,6 +1,6 @@
 import numpy as np
 from uarray import Dispatchable, wrap_single_convertor
-from unumpy import ufunc, ufunc_list, ndarray, dtype
+from unumpy import ufunc, ufunc_list, ndarray, dtype, random
 import unumpy
 import functools
 
@@ -24,11 +24,13 @@ def __ua_function__(method, args, kwargs):
     if method in _implementations:
         return _implementations[method](*args, **kwargs)
 
-    if not hasattr(np, method.__name__):
+    if hasattr(np, method.__name__):
+        return getattr(np, method.__name__)(*args, **kwargs)
+    elif hasattr(np.random, method.__name__):
+        return getattr(np.random, method.__name__)(*args, **kwargs)
+    else:
         return NotImplemented
 
-    return getattr(np, method.__name__)(*args, **kwargs)
-
 
 @wrap_single_convertor
 def __ua_convert__(value, dispatch_type, coerce):

unumpy/random/__init__.py

@@ -0,0 +1 @@
+from ._multimethods import *

unumpy/random/_multimethods.py

@@ -0,0 +1,306 @@
+import functools
+import operator
+from uarray import create_multimethod, mark_as, all_of_type, Dispatchable
+import builtins
+
+create_numpy = functools.partial(create_multimethod, domain="numpy")
+
+from .._multimethods import ndarray, _self_argreplacer
+
+
+@create_numpy(_self_argreplacer)
+@all_of_type(ndarray)
+def rand(*tup):
+    return tup
+
+
+@create_numpy(_self_argreplacer)
+@all_of_type(ndarray)
+def randn(*tup):
+    return tup
+
+
+@create_numpy(_self_argreplacer)
+@all_of_type(ndarray)
+def randint(low, high=None, size=None, dtype="l"):
+    return low
+
+
+@create_numpy(_self_argreplacer)
+@all_of_type(ndarray)
+def random_integers(low, high=None, size=None):
+    return low
+
+
+@create_numpy(_self_argreplacer)
+@all_of_type(ndarray)
+def random_sample(size=None):
+    return size
+
+
+@create_numpy(_self_argreplacer)
+@all_of_type(ndarray)
+def random(size=None):
+    return size
+
+
+@create_numpy(_self_argreplacer)
+@all_of_type(ndarray)
+def ranf(size=None):
+    return size
+
+
+@create_numpy(_self_argreplacer)
+@all_of_type(ndarray)
+def sample(size=None):
+    return size
+
+
+@create_numpy(_self_argreplacer)
+@all_of_type(ndarray)
+def choice(a, size=None, replace=True, p=None):
+    return (a, size, replace, p)
+
+
+@create_numpy(_self_argreplacer)
+@all_of_type(ndarray)
+def bytes(length):
+    return length
+
+
+@create_numpy(_self_argreplacer)
+@all_of_type(ndarray)
+def shuffle(x):
+    return x
+
+
+@create_numpy(_self_argreplacer)
+@all_of_type(ndarray)
+def permutation(x):
+    return x
+
+
+@create_numpy(_self_argreplacer)
+@all_of_type(ndarray)
+def beta(a, b, size=None):
+    return (a, b, size)
+
+
+@create_numpy(_self_argreplacer)
+@all_of_type(ndarray)
+def binomial(n, p, size=None):
+    return (n, p, size)
+
+
+@create_numpy(_self_argreplacer)
+@all_of_type(ndarray)
+def dirichlet(alpha, size=None):
+    return (alpha, size)
+
+
+@create_numpy(_self_argreplacer)
+@all_of_type(ndarray)
+def exponential(scale, size=None):
+    return (scale, size)
+
+
+@create_numpy(_self_argreplacer)
+@all_of_type(ndarray)
+def f(dfnum, dfden, size=None):
+    return (dfnum, dfden, size)
+
+
+@create_numpy(_self_argreplacer)
+@all_of_type(ndarray)
+def gamma(shape, scale=1.0, size=None):
+    return (shape, scale, size)
+
+
+@create_numpy(_self_argreplacer)
+@all_of_type(ndarray)
+def geometric(p, size=None):
+    return (p, size)
+
+
+@create_numpy(_self_argreplacer)
+@all_of_type(ndarray)
+def gumbel(loc=0.0, scale=1.0, size=None):
+    return (loc, scale, size)
+
+
+@create_numpy(_self_argreplacer)
+@all_of_type(ndarray)
+def hypergeometric(ngood, nbad, nsample, size=None):
+    return (ngood, nbad, nsample, size)
+
+
+@create_numpy(_self_argreplacer)
+@all_of_type(ndarray)
+def laplace(loc=0.0, scale=1.0, size=None):
+    return (loc, scale, size)
+
+
+@create_numpy(_self_argreplacer)
+@all_of_type(ndarray)
+def logistic(loc=0.0, scale=1.0, size=None):
+    return (loc, scale, size)
+
+
+@create_numpy(_self_argreplacer)
+@all_of_type(ndarray)
+def lognormal(mean=0.0, sigma=1.0, size=None):
+    return (mean, sigma, size)
+
+
+@create_numpy(_self_argreplacer)
+@all_of_type(ndarray)
+def logseries(p, size=None):
+    return (p, size)
+
+
+@create_numpy(_self_argreplacer)
+@all_of_type(ndarray)
+def multinomial(n, pvals, size=None):
+    return (n, pvals, size)
+
+
+# check
+@create_numpy(_self_argreplacer)
+@all_of_type(ndarray)
+def multivariate_normal(mean, cov, size=None, check_valid=None, tol=None):
+    return (mean, cov, size, check_valid, tol)
+
+
+@create_numpy(_self_argreplacer)
+@all_of_type(ndarray)
+def negative_binomial(n, p, size=None):
+    return (n, p, size)
+
+
+@create_numpy(_self_argreplacer)
+@all_of_type(ndarray)
+def noncentral_chisquare(df, nonc, size=None):
+    return (df, nonc, size)
+
+
+@create_numpy(_self_argreplacer)
+@all_of_type(ndarray)
+def noncentral_f(dfnum, dfden, nonc, size=None):
+    return (dfnum, dfden, nonc, size)
+
+
+@create_numpy(_self_argreplacer)
+@all_of_type(ndarray)
+def normal(loc=0.0, scale=1.0, size=None):
+    return (loc, scale, size)
+
+
+@create_numpy(_self_argreplacer)
+@all_of_type(ndarray)
+def pareto(a, size=None):
+    return (a, size)
+
+
+@create_numpy(_self_argreplacer)
+@all_of_type(ndarray)
+def poisson(lam=1.0, size=None):
+    return (lam, size)
+
+
+@create_numpy(_self_argreplacer)
+@all_of_type(ndarray)
+def power(a, size=None):
+    return (a, size)
+
+
+@create_numpy(_self_argreplacer)
+@all_of_type(ndarray)
+def rayleigh(scale=1.0, size=None):
+    return (scale, size)
+
+
+@create_numpy(_self_argreplacer)
+@all_of_type(ndarray)
+def standard_cauchy(size=None):
+    return size
+
+
+@create_numpy(_self_argreplacer)
+@all_of_type(ndarray)
+def standard_exponential(size=None):
+    return size
+
+
+@create_numpy(_self_argreplacer)
+@all_of_type(ndarray)
+def standard_gamma(shape, size=None):
+    return (shape, size)
+
+
+@create_numpy(_self_argreplacer)
+@all_of_type(ndarray)
+def standard_normal(size=None):
+    return size
+
+
+@create_numpy(_self_argreplacer)
+@all_of_type(ndarray)
+def standard_t(df, size=None):
+    return (df, size)
+
+
+@create_numpy(_self_argreplacer)
+@all_of_type(ndarray)
+def triangular(left, mode, right, size=None):
+    return (left, mode, right, size)
+
+
+@create_numpy(_self_argreplacer)
+@all_of_type(ndarray)
+def uniform(low=0.0, high=1.0, size=None):
+    return (low, high, size)
+
+
+@create_numpy(_self_argreplacer)
+@all_of_type(ndarray)
+def vonmises(mu, kappa, size=None):
+    return (mu, kappa, size)
+
+
+@create_numpy(_self_argreplacer)
+@all_of_type(ndarray)
+def wald(mean, scale, size=None):
+    return (mean, scale, size)
+
+
+@create_numpy(_self_argreplacer)
+@all_of_type(ndarray)
+def weibull(a, size=None):
+    return (a, size)
+
+
+@create_numpy(_self_argreplacer)
+@all_of_type(ndarray)
+def zipf(a, size=None):
+    return (a, size)
+
+
+# discuss: RandomState
+
+
+@create_numpy(_self_argreplacer)
+@all_of_type(ndarray)
+def seed(seed=None):
+    return seed
+
+
+@create_numpy(_self_argreplacer)
+@all_of_type(ndarray)
+def get_state():
+    return
+
+
+@create_numpy(_self_argreplacer)
+@all_of_type(ndarray)
+def set_state(state):
+    return state