analysis: Improve errors

Improve the errors in the ema_workbench/analysis, by:
- Making them more explicit
- Including what the variable value or type not expected currently is
- Including what the variable value or type not expected should be
- Stating the explicit type of warning or error (DeprecationWarning, ValueError, etc.)
- Converting them to F-strings (for better in-code readability)
- Formatting warnings and errors with a Capital letter, while leaving log-statements lowercase.

This will make debugging easier, allowing more time for model development and experimentation.

ema_workbench/analysis/b_and_w_plotting.py

@@ -148,7 +148,7 @@ def set_ax_collections_to_bw(ax, style, colormap):
         try:
             converter_func = _collection_converter[collection_type]
         except KeyError:
-            raise EMAError(f"converter for {collection_type} not implemented")
+            raise EMAError(f"Converter for collection type {collection_type} not implemented")
         else:
             converter_func(collection, ax, style, colormap)
 

ema_workbench/analysis/cart.py

@@ -60,7 +60,7 @@ def setup_cart(results, classify, incl_unc=None, mass_min=0.05):
         y = classify(outcomes)
         mode = sdutil.RuleInductionType.BINARY
     else:
-        raise TypeError("unknown type for classify")
+        raise TypeError(f"Unknown type for classify: {type(classify)}")
 
     return CART(x, y, mass_min, mode=mode)
 
@@ -329,7 +329,7 @@ def show_tree(self, mplfig=True, format="png"):
         # but just in case, we raise an error if assumption of len==1 does
         # not hold
         if len(graphs) > 1:
-            raise EMAError("trying to visualize more than one tree")
+            raise EMAError(f"Expected a single tree for visualization, but found {len(graphs)} trees.")
 
         graph = graphs[0]
 
@@ -343,6 +343,6 @@ def show_tree(self, mplfig=True, format="png"):
         elif format == "svg":
             img = graph.create_svg()
         else:
-            raise TypeError("""format must be in {'png', 'svg'}""")
+            raise TypeError(f"format must be 'png' or 'svg' (instead of {format}).")
 
         return img

ema_workbench/analysis/feature_scoring.py

@@ -112,7 +112,7 @@ def _prepare_outcomes(outcomes, classify):
         y = classify(outcomes)
         categorical = True
     else:
-        raise TypeError("unknown type for classify")
+        raise TypeError(f"Unknown type for classify: {type(classify)}")
 
     return y, categorical
 

ema_workbench/analysis/plotting.py

@@ -747,7 +747,7 @@ def multiple_densities(
 
         # making of grid
         if not points_in_time:
-            raise EMAError("no points in time specified")
+            raise EMAError("No points in time specified, should be a list of length 1-6")
         if len(points_in_time) == 1:
             ax_env = plt.subplot2grid((2, 3), (0, 0), colspan=3)
             ax1 = plt.subplot2grid((2, 3), (1, 1), sharey=ax_env)
@@ -788,7 +788,7 @@ def multiple_densities(
             ax6 = plt.subplot2grid((2, 6), (1, 5), sharex=ax1, sharey=ax_env)
             kde_axes = [ax1, ax2, ax3, ax4, ax5, ax6]
         else:
-            raise EMAError("too many points in time provided")
+            raise EMAError(f"Too many points in time specified: {len(points_in_time)}, max is 6")
 
         axes_dict["main plot"] = ax_env
         for n, entry in enumerate(kde_axes):

ema_workbench/analysis/plotting_util.py

@@ -284,7 +284,7 @@ def group_density(ax_d, density, outcomes, outcome_to_plot, group_labels, log=Fa
     elif density == Density.BOXENPLOT:
         plot_boxenplot(ax_d, values, log, group_labels=group_labels)
     else:
-        raise EMAError(f"unknown density type: {density}")
+        raise EMAError(f"Unknown density plot type: {density}")
 
     ax_d.set_xlabel("")
     ax_d.set_ylabel("")
@@ -316,7 +316,7 @@ def simple_density(density, value, ax_d, ax, log):
     elif density == Density.BOXENPLOT:
         plot_boxenplot(ax_d, [value[:, -1]], log)
     else:
-        raise EMAError("unknown density plot type")
+        raise EMAError(f"Unknown density plot type: {density}")
 
     ax_d.get_yaxis().set_view_interval(
         ax.get_yaxis().get_view_interval()[0], ax.get_yaxis().get_view_interval()[1]
@@ -671,8 +671,12 @@ def prepare_pairs_data(
     filter_scalar : bool, optional
 
     """
-    if isinstance(outcomes_to_show, str):
-        raise EMAError("for pair wise plotting, more than one outcome needs to be provided")
+    if outcomes_to_show is not None:
+        if not isinstance(outcomes_to_show, list):
+            raise TypeError(f"For pair-wise plotting multiple outcomes need to be provided.\n"
+                            f"outcomes_to_show must be a list of strings or None, instead of a {type(outcomes_to_show)}")
+        elif len(outcomes_to_show) < 2:
+            raise ValueError(f"Only {len(outcomes_to_show)} outcome provided, at least two are needed for pair-wise plotting.")
 
     experiments, outcomes, outcomes_to_show, time, grouping_labels = prepare_data(
         experiments, None, outcomes, outcomes_to_show, group_by, grouping_specifiers, filter_scalar
@@ -755,7 +759,7 @@ def prepare_data(
         if not grouping_specifiers:
             # no grouping specifier, so infer from the data
             if group_by == "index":
-                raise EMAError("no grouping specifiers provided while " "trying to group on index")
+                raise EMAError("No grouping specifiers provided while trying to group on index")
             else:
                 column_to_group_by = experiments[group_by]
                 if column_to_group_by.dtype in (object, "category"):

ema_workbench/analysis/prim.py

@@ -108,7 +108,7 @@ def pca_preprocess(experiments, y, subsets=None, exclude=set()):
     if not x.select_dtypes(exclude=np.number).empty:
         raise RuntimeError("X includes non numeric columns")
     if not set(np.unique(y)) == {0, 1}:
-        raise RuntimeError("y should only contain 0s and 1s")
+        raise RuntimeError(f"y should only contain 0s and 1s, currently y contains {set(np.unique(y))}.")
 
     # if no subsets are provided all uncertainties with non dtype object
     # are in the same subset, the name of this is r, for rotation
@@ -416,7 +416,7 @@ def inspect(self, i=None, style="table", **kwargs):
             i = [i]
 
         if not all(isinstance(x, int) for x in i):
-            raise TypeError("i must be an integer or list of integers")
+            raise TypeError(f"i must be an integer or list of integers, not {type(i)}")
 
         return [self._inspect(entry, style=style, **kwargs) for entry in i]
 
@@ -450,7 +450,7 @@ def _inspect(self, i=None, style="table", **kwargs):
         elif style == "data":
             return self._inspect_data(i, uncs, qp_values)
         else:
-            raise ValueError("style must be one of graph, table or data")
+            raise ValueError(f"style must be one of 'graph', 'table' or 'data', not {style}.")
 
     def _inspect_data(self, i, uncs, qp_values):
         """Helper method for inspecting boxes,
@@ -1044,9 +1044,9 @@ def __init__(
         self._update_yi_remaining = self._update_functions[update_function]
 
         if len(self.y.shape) > 1:
-            raise PrimException("y is not a 1-d array")
+            raise PrimException(f"y is not a 1-d array, but has shape {self.y.shape}")
         if self.y.shape[0] != len(self.x):
-            raise PrimException("len(y) != len(x)")
+            raise PrimException(f"len(y) != len(x): {self.y.shape[0]} != {len(self.x)}")
 
         # store the remainder of the parameters
         self.paste_alpha = paste_alpha