import lasagne.nonlinearities as to
import numpy as np
import pymc3 as pm
from lasagne import layers as llayers
from gelato.layers import DenseLayer, InputLayer
from gelato.layers import get_output, find_parent, find_root
from gelato.specs.dist import NormalSpec, LognormalSpec
from .datasets import generate_linear_regression
class TestWorkflow(object):
@classmethod
def setup_class(cls):
cls.intercept = 1
cls.slope = 3
cls.sd = .1
cls.x, cls.y = generate_linear_regression(cls.intercept, cls.slope, sd=cls.sd)
cls.x = np.matrix(cls.x).T
cls.y = np.matrix(cls.y).T
def test_workflow(self):
inp = InputLayer(self.x.shape)
out = DenseLayer(inp, 1, W=NormalSpec(sd=LognormalSpec()), nonlinearity=to.identity)
out = DenseLayer(out, 1, W=NormalSpec(sd=LognormalSpec()), nonlinearity=to.identity)
assert out.root is inp
with out:
pm.Normal('y', mu=get_output(out),
sd=self.sd,
observed=self.y)
def test_find_parent_and_root(self):
inp = InputLayer(self.x.shape)
middle = llayers.DenseLayer(inp, 10)
middle1 = DenseLayer(middle, 10)
middle2 = llayers.DenseLayer(middle1, 4)
assert find_parent(middle2) is middle1
assert find_root(middle2) is inp
