Bug fix for plotting order in plot_comparison (fixes mismatch between labels and plot)

bambi/backend/terms.py

@@ -441,7 +441,7 @@ def build(self, bmb_model):
             contribution = pt.concatenate(contribution_list)[indexes_to_unsort]
         # If there are no groups, it's a single dot product
         else:
-            contribution = phi @ coeffs
+            contribution = pt.dot(phi, coeffs)  # "@" operator is not working as expected
 
         output = pm.Deterministic(label, contribution, dims=contribution_dims)
         return output

bambi/interpret/create_data.py

@@ -70,6 +70,11 @@ def _grid_level(
     elif kind == "slopes":
         pairwise_grid = enforce_dtypes(condition_info.model.data, pairwise_grid, variable_info.name)
 
+    # After computing default values, fractional values may have been computed.
+    # Enforcing the dtype of "int" may create duplicate rows as it will round
+    # the fractional values.
+    pairwise_grid = pairwise_grid.drop_duplicates()
+
     return pairwise_grid
 
 

bambi/interpret/plot_types.py

@@ -3,7 +3,7 @@
 import numpy as np
 import pandas as pd
 
-from bambi.interpret.utils import Covariates, get_unique_levels, get_group_offset, identity
+from bambi.interpret.utils import Covariates, get_group_offset, identity
 
 
 def plot_numeric(
@@ -49,7 +49,7 @@ def plot_numeric(
         ax.fill_between(values_main, y_hat_bounds[0], y_hat_bounds[1], alpha=0.4)
     elif "group" in covariates and not "panel" in covariates:
         ax = axes[0]
-        colors = get_unique_levels(plot_data[color])
+        colors = np.unique(plot_data[color])
         for i, clr in enumerate(colors):
             idx = (plot_data[color] == clr).to_numpy()
             values_main = transform_main(plot_data.loc[idx, main])
@@ -62,16 +62,16 @@ def plot_numeric(
                 color=f"C{i}",
             )
     elif not "group" in covariates and "panel" in covariates:
-        panels = get_unique_levels(plot_data[panel])
+        panels = np.unique(plot_data[panel])
         for ax, pnl in zip(axes.ravel(), panels):
             idx = (plot_data[panel] == pnl).to_numpy()
             values_main = transform_main(plot_data.loc[idx, main])
             ax.plot(values_main, y_hat_mean[idx], solid_capstyle="butt")
             ax.fill_between(values_main, y_hat_bounds[0][idx], y_hat_bounds[1][idx], alpha=0.4)
             ax.set(title=f"{panel} = {pnl}")
     elif "group" in covariates and "panel" in covariates:
-        colors = get_unique_levels(plot_data[color])
-        panels = get_unique_levels(plot_data[panel])
+        colors = np.unique(plot_data[color])
+        panels = np.unique(plot_data[panel])
         if color == panel:
             for i, (ax, pnl) in enumerate(zip(axes.ravel(), panels)):
                 idx = (plot_data[panel] == pnl).to_numpy()
@@ -138,20 +138,20 @@ def plot_categoric(covariates: Covariates, plot_data: pd.DataFrame, legend: bool
 
     main, color, panel = covariates.main, covariates.group, covariates.panel
     covariates = {k: v for k, v in vars(covariates).items() if v is not None}
-    main_levels = get_unique_levels(plot_data[main])
+    main_levels = np.unique(plot_data[main])
     main_levels_n = len(main_levels)
     idxs_main = np.arange(main_levels_n)
     y_hat_mean = plot_data["estimate"]
     y_hat_bounds = np.transpose(plot_data[plot_data.columns[-2:]].values)
 
     if "group" in covariates:
-        colors = get_unique_levels(plot_data[color])
+        colors = np.unique(plot_data[color])
         colors_n = len(colors)
         offset_bounds = get_group_offset(colors_n)
         colors_offset = np.linspace(-offset_bounds, offset_bounds, colors_n)
 
     if "panel" in covariates:
-        panels = get_unique_levels(plot_data[panel])
+        panels = np.unique(plot_data[panel])
 
     if len(covariates) == 1:
         ax = axes[0]

bambi/interpret/plotting.py

@@ -315,13 +315,16 @@ def plot_comparisons(
     if conditional is None and average_by is None:
         raise ValueError("Must specify at least one of 'conditional' or 'average_by'.")
 
-    if conditional is not None:
-        if not isinstance(conditional, str):
-            if len(conditional) > 3 and average_by is None:
-                raise ValueError(
-                    "Must specify a covariate to 'average_by' when number of covariates"
-                    "passed to 'conditional' is greater than 3."
-                )
+    if isinstance(conditional, dict):
+        conditional = {key: sorted(value) for key, value in conditional.items()}
+    elif conditional is not None:
+        conditional = listify(conditional)
+        if len(conditional) > 3 and average_by is None:
+            raise ValueError(
+                "Must specify a covariate to 'average_by' when number of covariates"
+                "passed to 'conditional' is greater than 3."
+            )
+
     if average_by is True:
         raise ValueError(
             "Plotting when 'average_by = True' is not possible as 'True' marginalizes "
@@ -469,13 +472,16 @@ def plot_slopes(
 
     if conditional is None and average_by is None:
         raise ValueError("Must specify at least one of 'conditional' or 'average_by'.")
-    if conditional is not None:
-        if not isinstance(conditional, str):
-            if len(conditional) > 3 and average_by is None:
-                raise ValueError(
-                    "Must specify a covariate to 'average_by' when number of covariates"
-                    "passed to 'conditional' is greater than 3."
-                )
+
+    if isinstance(conditional, dict):
+        conditional = {key: sorted(value) for key, value in conditional.items()}
+    elif conditional is not None:
+        conditional = listify(conditional)
+        if len(conditional) > 3 and average_by is None:
+            raise ValueError(
+                "Must specify a covariate to 'average_by' when number of covariates"
+                "passed to 'conditional' is greater than 3."
+            )
 
     if average_by is True:
         raise ValueError(

bambi/interpret/utils.py

@@ -124,7 +124,7 @@ def set_default_variable_values(self) -> np.ndarray:
                                         predictor_data, self.eps, dtype
                                     )
                             elif component.kind == "categoric":
-                                values = get_unique_levels(predictor_data)
+                                values = np.unique(predictor_data)
 
         return values
 
@@ -377,17 +377,6 @@ def make_group_values(x: np.ndarray, groups_n: int = 5) -> np.ndarray:
     raise ValueError("Group covariate must be numeric or categoric.")
 
 
-def get_unique_levels(x: np.ndarray) -> np.ndarray:
-    """
-    Get unique levels of a categoric variable.
-    """
-    if hasattr(x, "dtype") and hasattr(x.dtype, "categories"):
-        levels = np.array((x.dtype.categories))
-    else:
-        levels = np.unique(x)
-    return levels
-
-
 def get_group_offset(n, lower: float = 0.05, upper: float = 0.4) -> np.ndarray:
     # Complementary log log function, scaled.
     # See following code to have an idea of how this function looks like