Add missing docstrings

Author: alexgdebeer

deep_tensor/references/reference.py

@@ -2,7 +2,9 @@
 from typing import Tuple
 import warnings
 
+import torch
 from torch import Tensor
+from torch.quasirandom import SobolEngine
 
 from ..domains import Domain
 
@@ -110,10 +112,10 @@ def eval_potential(self, rs: Tensor) -> Tuple[Tensor, Tensor]:
         """
         pass
 
-    @abc.abstractmethod
     def random(self, d: int, n: int) -> Tensor:
-        """Generates a set of samples from the reference density using
-        the inverse CDF method.
+        r"""Generates a set of random samples.
+
+        Generates a set of random samples from the reference density.
         
         Parameters
         ----------
@@ -125,15 +127,18 @@ def random(self, d: int, n: int) -> Tensor:
         Returns
         -------
         rs:
-            An n * d matrix containing the generated samples.
+            An $n \times d$ matrix containing the generated samples.
 
         """
-        pass
-    
-    @abc.abstractmethod
+        zs = torch.rand(n, d)
+        rs = self.invert_cdf(zs)
+        return rs
+        
     def sobol(self, d: int, n: int) -> Tensor:
-        """Generates a set of QMC samples from the reference density 
-        using a Sobol sequence.
+        r"""Generates a set of QMC samples.
+
+        Generates a set of QMC samples from the reference density using 
+        a Sobol sequence.
         
         Parameters
         ----------
@@ -145,10 +150,13 @@ def sobol(self, d: int, n: int) -> Tensor:
         Returns
         -------
         rs:
-            An n * d matrix containing the generated samples.
+            An $n \times d$ matrix containing the generated samples.
         
         """
-        pass
+        S = SobolEngine(dimension=d)
+        zs = S.draw(n)
+        rs = self.invert_cdf(zs)
+        return rs
 
     def _out_domain(self, rs: Tensor) -> Tensor:
         outside = (rs < self.domain.left) | (self.domain.right < rs)

deep_tensor/references/symmetric_reference.py

@@ -4,7 +4,6 @@
 
 import torch
 from torch import Tensor
-from torch.quasirandom import SobolEngine
 
 from .reference import Reference
 from ..domains import BoundedDomain, Domain
@@ -162,15 +161,4 @@ def eval_potential(self, rs: Tensor) -> Tuple[Tensor, Tensor]:
         d_rs = rs.shape[1]
         log_ps, log_dpdrs = self.eval_unit_potential(rs)
         log_ps = log_ps + d_rs * math.log(self.norm)
-        return log_ps, log_dpdrs
-    
-    def random(self, d: int, n: int) -> Tensor:
-        zs = torch.rand(n, d)
-        rs = self.invert_cdf(zs)
-        return rs
-        
-    def sobol(self, d: int, n: int) -> Tensor:
-        S = SobolEngine(dimension=d)
-        zs = S.draw(n)
-        rs = self.invert_cdf(zs)
-        return rs
+        return log_ps, log_dpdrs

deep_tensor/references/uniform_reference.py

@@ -2,7 +2,6 @@
 
 import torch
 from torch import Tensor 
-from torch.quasirandom import SobolEngine
 
 from .reference import Reference
 from ..domains import BoundedDomain
@@ -37,13 +36,4 @@ def eval_potential(self, rs: Tensor) -> Tuple[Tensor, Tensor]:
         n_rs = rs.shape[0]
         log_ps = torch.zeros(n_rs)
         log_dpdrs = torch.zeros(n_rs)  # TODO: look at this more closely.
-        return log_ps, log_dpdrs
-    
-    def random(self, d: int, n: int) -> Tensor:
-        rs = torch.rand(n, d)
-        return rs
-    
-    def sobol(self, d: int, n: int) -> Tensor:
-        S = SobolEngine(dimension=d)
-        rs = S.draw(n)
-        return rs
+        return log_ps, log_dpdrs

docs/objects.json

@@ -30,6 +30,8 @@ GaussianReference.random(d: int, n: int)
 
 Generates a set of random samples.
 
+Generates a set of random samples from the reference density.
+
 #### Parameters {.doc-section .doc-section-parameters}
 
 <code>[**d**]{.parameter-name} [:]{.parameter-annotation-sep} [int]{.parameter-annotation}</code>
@@ -54,6 +56,9 @@ GaussianReference.sobol(d: int, n: int)
 
 Generates a set of QMC samples.
 
+Generates a set of QMC samples from the reference density using 
+a Sobol sequence.
+
 #### Parameters {.doc-section .doc-section-parameters}
 
 <code>[**d**]{.parameter-name} [:]{.parameter-annotation-sep} [int]{.parameter-annotation}</code>

docs/reference/GaussianReference.qmd

@@ -21,6 +21,8 @@ UniformReference.random(d: int, n: int)
 
 Generates a set of random samples.
 
+Generates a set of random samples from the reference density.
+
 #### Parameters {.doc-section .doc-section-parameters}
 
 <code>[**d**]{.parameter-name} [:]{.parameter-annotation-sep} [int]{.parameter-annotation}</code>
@@ -45,6 +47,9 @@ UniformReference.sobol(d: int, n: int)
 
 Generates a set of QMC samples.
 
+Generates a set of QMC samples from the reference density using 
+a Sobol sequence.
+
 #### Parameters {.doc-section .doc-section-parameters}
 
 <code>[**d**]{.parameter-name} [:]{.parameter-annotation-sep} [int]{.parameter-annotation}</code>