Documentation improvements

.github/workflows/pypi_release.yml

@@ -14,6 +14,7 @@ jobs:
       uses: actions/setup-python@v4
       with:
         python-version: "3.x"
+        cache: `pip`
     - name: Install pypa/build
       run: >-
         python3 -m

.github/workflows/sphinx.yml

@@ -26,4 +26,3 @@ jobs:
       uses: sphinx-notes/pages@v3
       with:
         documentation_path: docs/source
-        cache: true

fairret/loss/projection.py

@@ -27,13 +27,13 @@ class ProjectionLoss(FairnessLoss):
 
     The projections are computed using cvxpy. Hence, any subclass is expected to implement the statistical distance
     between distributions in both cvxpy and PyTorch by implementing the
-    :py:meth:`~projection.ProjectionLoss.cvxpy_distance` method and the
-    :py:meth:`~projection.ProjectionLoss.torch_distance` method respectively.
+    :py:meth:`~fairret.loss.projection.ProjectionLoss.cvxpy_distance` method and the
+    :py:meth:`~fairret.loss.projection.ProjectionLoss.torch_distance` method respectively.
 
-    Optionally, the :py:meth:`~projection.ProjectionLoss.torch_distance_with_logits` method can be overwritten to
-    provide a more numerically stable handling of predictions that are provided as logits. If left unimplemented,
-    :py:meth:`~projection.ProjectionLoss.torch_distance` will be called instead, after applying the sigmoid function to
-    the predictions.
+    Optionally, the :py:meth:`~fairret.loss.projection.ProjectionLoss.torch_distance_with_logits` method can be
+    overwritten to provide a more numerically stable handling of predictions that are provided as logits. If left
+    unimplemented, :py:meth:`~fairret.loss.projection.ProjectionLoss.torch_distance` will be called instead,
+    after applying the sigmoid function to the predictions.
 
     Note:
         We use 'statistical distance' in a broad sense here, and do not require that the distance is a metric. See
@@ -131,8 +131,8 @@ def torch_distance(self, pred: torch.Tensor, proj: torch.Tensor) -> torch.Tensor
 
     def torch_distance_with_logits(self, pred, proj):
         """
-        A more numerically stable alternative method to :py:meth:`~projection.ProjectionLoss.torch_distance`, where `pred`
-        is assumed to be logits.
+        A more numerically stable alternative method to
+        :py:meth:`~fairret.loss.projection.ProjectionLoss.torch_distance`, where `pred` is assumed to be logits.
 
         Args:
             pred (torch.Tensor): The predicted distribution as logits, in shape (N,1). As we assume binary

fairret/loss/violation.py

@@ -43,7 +43,7 @@ def forward(self, pred: torch.Tensor, sens: torch.Tensor, *stat_args, pred_as_lo
                 target_statistic: Optional[torch.Tensor] = None, **stat_kwargs: Any) -> torch.Tensor:
         """
         Calculate the violation vector in relation to the `target_statistic` and penalize this violation using the
-        :py:meth:`~violation.ViolationLoss.penalize_violation` method implemented by the subclass.
+        :py:meth:`~fairret.loss.violation.ViolationLoss.penalize_violation` method implemented by the subclass.
 
         Args:
             pred (torch.Tensor): Predictions of shape :math:`(N, 1)`, as we assume to be performing binary

fairret/statistic/linear_fractional.py

@@ -595,14 +595,15 @@ def denom_slope(self, label: torch.Tensor) -> torch.Tensor:
 class StackedLinearFractionalStatistic(LinearFractionalStatistic):
     """
     A vector-valued statistic that combines the outputs of K
-    :py:class:`fairret.statistic.linear_fractional.LinearFractionalStatistic` 's into a single statistic with output
+    :py:class:`~fairret.statistic.linear_fractional.LinearFractionalStatistic` 's into a single statistic with output
     (K, S) by stacking all outputs in the second-to-last dimension (`dim=-2`).
     """
 
     def __init__(self, *statistics: LinearFractionalStatistic):
         """
         Args:
-            *statistics: The :py:class:`fairret.statistic.linear_fractional.LinearFractionalStatistic` 's to be stacked.
+            *statistics: The :py:class:`~fairret.statistic.linear_fractional.LinearFractionalStatistic` 's to be
+            stacked.
         """
 
         super().__init__()