Uniform endpoints

Documentation/CHANGELOG.md

@@ -47,6 +47,7 @@ Release Date: TBD
 * Renames `DiscreteDistributionXRA` to `DiscreteDistributionLabeled` and updates methods [#1170](https://github.com/econ-ark/HARK/pull/1170)
 * Renames `HARK.numba` to `HARK.numba_tools` [#1183](https://github.com/econ-ark/HARK/pull/1183)
 * Adds the RNG seed as a property of `DiscreteDistributionLabeled` [#1184](https://github.com/econ-ark/HARK/pull/1184)
+* Updates the `approx` method of `HARK.distributions.Uniform` to include the endpoints of the distribution with infinitesimally small (zero) probability mass. [#1180](https://github.com/econ-ark/HARK/pull/1180)
 
 ### 0.12.0
 

HARK/distribution.py

@@ -742,14 +742,16 @@ def draw(self, N):
             )
         return draws[0] if len(draws) == 1 else draws
 
-    def approx(self, N):
+    def approx(self, N, endpoint=False):
         """
         Makes a discrete approximation to this uniform distribution.
 
         Parameters
         ----------
         N : int
-            The number of points in the discrete approximation
+            The number of points in the discrete approximation.
+        endpoint : bool
+            Whether to include the endpoints in the approximation.
 
         Returns
         -------
@@ -758,11 +760,17 @@ def approx(self, N):
             points for discrete probability mass function.
         """
         pmv = np.ones(N) / float(N)
+
         center = (self.top + self.bot) / 2.0
         width = (self.top - self.bot) / 2.0
         atoms = center + width * np.linspace(-(N - 1.0) / 2.0, (N - 1.0) / 2.0, N) / (
             N / 2.0
         )
+
+        if endpoint:  # insert endpoints with infinitesimally small mass
+            atoms = np.concatenate(([self.bot], atoms, [self.top]))
+            pmv = np.concatenate(([0.0], pmv, [0.0]))
+
         return DiscreteDistribution(
             pmv, atoms, seed=self.RNG.randint(0, 2**31 - 1, dtype="int32")
         )

HARK/tests/test_distribution.py

@@ -312,16 +312,16 @@ def test_MVNormal(self):
         ## Are these tests generator/backend specific?
         dist = MVNormal()
 
-        #self.assertTrue(
+        # self.assertTrue(
         #    np.allclose(dist.draw(1)[0], np.array([2.76405235, 1.40015721]))
-        #)
+        # )
 
         dist.draw(100)
         dist.reset()
 
-        #self.assertTrue(
+        # self.assertTrue(
         #    np.allclose(dist.draw(1)[0], np.array([2.76405235, 1.40015721]))
-        #)
+        # )
 
     def test_Weibull(self):
         Weibull().draw(1)[0]
@@ -333,6 +333,12 @@ def test_Uniform(self):
 
         self.assertEqual(calc_expectation(uni.approx(10)), 0.5)
 
+        uni_discrete = uni.approx(10, endpoint=True)
+
+        self.assertEqual(uni_discrete.atoms[0][0], 0.0)
+        self.assertEqual(uni_discrete.atoms[0][-1], 1.0)
+        self.assertEqual(calc_expectation(uni.approx(10)), 0.5)
+
     def test_Bernoulli(self):
         Bernoulli().draw(1)[0]