Added tests for gridkey, horizontal plots, delta text, summary bars, swarm bars, and contrast bars

Author: JAnns98

dabest/misc_tools.py

@@ -872,7 +872,7 @@ def get_plot_groups(is_paired, idx, proportional, all_plot_groups):
     return temp_idx, temp_all_plot_groups
 
 
-def add_counts_to_ticks(plot_data, xvar, yvar, rawdata_axes, plot_kwargs, horizontal):
+def add_counts_to_ticks(plot_data, xvar, yvar, rawdata_axes, plot_kwargs, flow, horizontal):
     """
 
     Add the counts to the raw data axes labels.
@@ -889,6 +889,8 @@ def add_counts_to_ticks(plot_data, xvar, yvar, rawdata_axes, plot_kwargs, horizo
         The raw data axes.
     plot_kwargs : dict
         Kwargs passed to the plot function.
+    flow : bool
+        Whether sankey flow is enabled or not.
     horizontal : bool
         A boolean flag to determine if the plot is for horizontal plotting.
     """
@@ -919,7 +921,12 @@ def lookup_value(text):
 
     for ticklab in get_label():
         t = ticklab.get_text()
-        te = t.split('\n')[-1]  # Get the last line of the label
+
+        if horizontal and not flow:
+            te = t.split('v.s. ')[-1]  # Get the last line of the label
+        else:
+            te = t.split('\n')[-1]  # Get the last line of the label
+
         value = lookup_value(te)
         if horizontal:
             ticks_with_counts.append(f"{t} (N={value})")
@@ -992,7 +999,7 @@ def extract_contrast_plotting_ticks(is_paired, show_pairs, two_col_sankey, plot_
     return ticks_to_skip, ticks_to_plot, ticks_to_skip_contrast, ticks_to_start_twocol_sankey
 
 def set_xaxis_ticks_and_lims(show_delta2, show_mini_meta, rawdata_axes, contrast_axes, show_pairs, float_contrast,
-                             ticks_to_skip, contrast_xtick_labels, plot_kwargs, horizontal):
+                             ticks_to_skip, contrast_xtick_labels, plot_kwargs, proportional, horizontal):
     """
     Set the x-axis/yaxis ticks and limits for the plotter function.
 
@@ -1016,6 +1023,8 @@ def set_xaxis_ticks_and_lims(show_delta2, show_mini_meta, rawdata_axes, contrast
         A list of contrast xtick labels.
     plot_kwargs : dict
         Kwargs passed to the plot function.
+    proportional: bool
+        A boolean flag to determine if the plot is a proportional plot.
     horizontal : bool
         A boolean flag to determine if the plot is for horizontal plotting.
     """
@@ -1034,6 +1043,9 @@ def set_xaxis_ticks_and_lims(show_delta2, show_mini_meta, rawdata_axes, contrast
             max_x = contrast_axes.get_ylim()[1]
             rawdata_axes.set_ylim(-0.375, max_x)
 
+            if proportional:
+                rawdata_axes.set_ylim(-0.375, max_x+0.1)
+
         if show_delta2 or show_mini_meta:
             # Increase the ylim of raw data by 2
             temp = rawdata_axes.get_ylim()

dabest/plot_tools.py

@@ -850,8 +850,8 @@ def sankeydiag(
         sankey_ticks = (
             [f"{left}" for left in broadcasted_left]
             if flow
-            else [
-                f"{left}\n v.s.\n{right}"
+            else [f"{left} v.s. {right}" if horizontal
+                else f"{left}\n v.s.\n{right}"
                 for left, right in zip(broadcasted_left, right_idx)
             ]
         )
@@ -917,9 +917,9 @@ def summary_bars_plotter(summary_bars: list, results: object, ax_to_plot: object
     else:
         summary_xmin, summary_xmax = ax_to_plot.get_xlim()
         summary_bars_colors = (
-            [summary_bars_kwargs.get('color')]*(max(ticks_to_plot)+1)
+            [summary_bars_kwargs.get('color')]*int(max(ticks_to_plot)+1)
             if summary_bars_kwargs.get('color') is not None
-            else ['black']*(max(ticks_to_plot)+1)
+            else ['black']*int(max(ticks_to_plot)+1)
             if color_col is not None or (proportional and is_paired) or is_paired 
             else list(plot_palette_raw.values())
         )
@@ -932,7 +932,7 @@ def summary_bars_plotter(summary_bars: list, results: object, ax_to_plot: object
                 summary_ci_low = results.pct_low[summary_index]
                 summary_ci_high = results.pct_high[summary_index]
 
-            summary_color = summary_bars_colors[ticks_to_plot[summary_index]]
+            summary_color = summary_bars_colors[int(ticks_to_plot[summary_index])]
 
             ax_to_plot.add_patch(mpatches.Rectangle((summary_xmin,summary_ci_low),summary_xmax+1, 
             summary_ci_high-summary_ci_low, zorder=-2, color=summary_color, **summary_bars_kwargs))
@@ -981,14 +981,14 @@ def contrast_bars_plotter(results: object, ax_to_plot: object,  swarm_plot_ax: o
 
     contrast_means = []
     for j, tick in enumerate(ticks_to_plot):
-        contrast_means.append(results.difference[j])
+        contrast_means.append(results.difference[int(j)])
 
     unpacked_idx = [element for innerList in idx for element in innerList] 
 
     contrast_bars_colors = (
-        [contrast_bars_kwargs.get('color')] * (max(ticks_to_plot) + 1) 
+        [contrast_bars_kwargs.get('color')] * int(max(ticks_to_plot) + 1) 
         if contrast_bars_kwargs.get('color') is not None 
-        else ['black'] * (max(ticks_to_plot) + 1) 
+        else ['black'] * int(max(ticks_to_plot) + 1) 
         if color_col is not None or is_paired 
         else plot_palette_raw
     )
@@ -1136,9 +1136,9 @@ def delta_text_plotter(results: object, ax_to_plot: object, swarm_plot_ax: objec
     delta_text_kwargs.pop('x_location')
 
     delta_text_colors = (
-        [delta_text_kwargs.get('color')]*(max(ticks_to_plot)+1)
+        [delta_text_kwargs.get('color')]*int(max(ticks_to_plot)+1)
         if delta_text_kwargs.get('color') is not None
-        else ['black']*(max(ticks_to_plot)+1)
+        else ['black']*int(max(ticks_to_plot)+1)
         if color_col is not None or (proportional and is_paired) or is_paired
         else plot_palette_raw
     )
@@ -1156,7 +1156,7 @@ def delta_text_plotter(results: object, ax_to_plot: object, swarm_plot_ax: objec
     # Collect the Y-values for the delta text
     Delta_Values = []
     for j, tick in enumerate(ticks_to_plot):
-        Delta_Values.append(results.difference[j])
+        Delta_Values.append(results.difference[int(j)])
     if show_delta2: Delta_Values.append(delta_delta.difference)
     if show_mini_meta: Delta_Values.append(mini_meta_delta.difference)
 
@@ -1197,9 +1197,9 @@ def delta_text_plotter(results: object, ax_to_plot: object, swarm_plot_ax: objec
     for x,y,t,tick in zip(delta_text_x_coordinates, delta_text_y_coordinates,Delta_Values,ticks_to_plot):
         Delta_Text = np.format_float_positional(t, precision=2, sign=True, trim="k", min_digits=2)
         idx_selector = (
-            tick
+            int(tick)
             if type(delta_text_colors) == list 
-            else unpacked_idx[tick]
+            else unpacked_idx[int(tick)]
             )
         ax_to_plot.text(x, y, Delta_Text, color=delta_text_colors[idx_selector], zorder=5, **delta_text_kwargs)
 
@@ -1349,7 +1349,16 @@ def slopegraph_plotter(dabest_obj, plot_data, xvar, yvar, color_col, plot_palett
         current_pair = pivoted_plot_data.loc[
             :, pd.MultiIndex.from_product([pivot_values, current_tuple])
         ].dropna()
+
+        # Check for correct pairing
+        if len(current_pair) == 0:
+            raise ValueError('There are no pairs to plot... check original dataframe for correct ID pairing')
+
+        current_pair = pivoted_plot_data.loc[
+            :, pd.MultiIndex.from_product([pivot_values, current_tuple])
+        ]
         grp_count = len(current_tuple)
+
         # Iterate through the data for the current tuple.
         for ID, observation in current_pair.iterrows():
             x_points = [t + 0.15*jitter*rng.standard_t(df=6, size=None) for t in range(x_start, x_start + grp_count)] # devMJBL
@@ -1545,16 +1554,16 @@ def effect_size_curve_plotter(ticks_to_plot, results, ci_type, contrast_axes, vi
     # Plot the curves
     contrast_xtick_labels = []
     for j, tick in enumerate(ticks_to_plot):
-        current_group = results.test[j]
-        current_control = results.control[j]
-        current_bootstrap = results.bootstraps[j]
-        current_effsize = results.difference[j]
+        current_group = results.test[int(j)]
+        current_control = results.control[int(j)]
+        current_bootstrap = results.bootstraps[int(j)]
+        current_effsize = results.difference[int(j)]
         if ci_type == "bca":
-            current_ci_low = results.bca_low[j]
-            current_ci_high = results.bca_high[j]
+            current_ci_low = results.bca_low[int(j)]
+            current_ci_high = results.bca_high[int(j)]
         else:
-            current_ci_low = results.pct_low[j]
-            current_ci_high = results.pct_high[j]
+            current_ci_low = results.pct_low[int(j)]
+            current_ci_high = results.pct_high[int(j)]
 
         # Create the violinplot.
         # New in v0.2.6: drop negative infinities before plotting.
@@ -2195,6 +2204,8 @@ def __init__(
         h = (ax.get_position().ymax - ax.get_position().ymin) * figh
         ax_xspan = ax.get_xlim()[1] - ax.get_xlim()[0]
         ax_yspan = ax.get_ylim()[1] - ax.get_ylim()[0]
+        if horizontal:
+            ax_xspan, ax_yspan = ax_yspan, ax_xspan
 
         # increases jitter distance based on number of swarms that is going to be drawn
         jitter = jitter * (1 + 0.05 * (math.log(ax_xspan)))

dabest/plotter.py

@@ -175,6 +175,7 @@ def effectsize_df_plotter(effectsize_df, **plot_kwargs):
                                                     )
         if not proportional:
             # Plot the raw data as a slopegraph.
+
             slopegraph_plotter(
                 dabest_obj=dabest_obj, 
                 plot_data=plot_data, 
@@ -188,7 +189,7 @@ def effectsize_df_plotter(effectsize_df, **plot_kwargs):
                 temp_idx=temp_idx,
                 horizontal=horizontal
                 )
-
+            
             # DELTA PTS ON CONTRAST PLOT WIP
             show_delta_dots = plot_kwargs["delta_dot"]
             if show_delta_dots and is_paired is not None:
@@ -313,6 +314,7 @@ def effectsize_df_plotter(effectsize_df, **plot_kwargs):
             yvar=yvar, 
             rawdata_axes=rawdata_axes, 
             plot_kwargs=plot_kwargs,
+            flow = sankey_kwargs["flow"],
             horizontal=horizontal,
             )
 
@@ -388,6 +390,7 @@ def effectsize_df_plotter(effectsize_df, **plot_kwargs):
                         ticks_to_skip=ticks_to_skip, 
                         contrast_xtick_labels=contrast_xtick_labels, 
                         plot_kwargs=plot_kwargs,
+                        proportional=proportional,
                         horizontal=horizontal,
                         )
     # Legend

nbs/API/misc_tools.ipynb

@@ -925,7 +925,7 @@
     "    return temp_idx, temp_all_plot_groups\n",
     "\n",
     "\n",
-    "def add_counts_to_ticks(plot_data, xvar, yvar, rawdata_axes, plot_kwargs, horizontal):\n",
+    "def add_counts_to_ticks(plot_data, xvar, yvar, rawdata_axes, plot_kwargs, flow, horizontal):\n",
     "    \"\"\"\n",
     "\n",
     "    Add the counts to the raw data axes labels.\n",
@@ -942,6 +942,8 @@
     "        The raw data axes.\n",
     "    plot_kwargs : dict\n",
     "        Kwargs passed to the plot function.\n",
+    "    flow : bool\n",
+    "        Whether sankey flow is enabled or not.\n",
     "    horizontal : bool\n",
     "        A boolean flag to determine if the plot is for horizontal plotting.\n",
     "    \"\"\"\n",
@@ -972,7 +974,12 @@
     "    \n",
     "    for ticklab in get_label():\n",
     "        t = ticklab.get_text()\n",
-    "        te = t.split('\\n')[-1]  # Get the last line of the label\n",
+    "\n",
+    "        if horizontal and not flow:\n",
+    "            te = t.split('v.s. ')[-1]  # Get the last line of the label\n",
+    "        else:\n",
+    "            te = t.split('\\n')[-1]  # Get the last line of the label\n",
+    "\n",
     "        value = lookup_value(te)\n",
     "        if horizontal:\n",
     "            ticks_with_counts.append(f\"{t} (N={value})\")\n",
@@ -1045,7 +1052,7 @@
     "    return ticks_to_skip, ticks_to_plot, ticks_to_skip_contrast, ticks_to_start_twocol_sankey\n",
     "\n",
     "def set_xaxis_ticks_and_lims(show_delta2, show_mini_meta, rawdata_axes, contrast_axes, show_pairs, float_contrast,\n",
-    "                             ticks_to_skip, contrast_xtick_labels, plot_kwargs, horizontal):\n",
+    "                             ticks_to_skip, contrast_xtick_labels, plot_kwargs, proportional, horizontal):\n",
     "    \"\"\"\n",
     "    Set the x-axis/yaxis ticks and limits for the plotter function.\n",
     "\n",
@@ -1069,6 +1076,8 @@
     "        A list of contrast xtick labels.\n",
     "    plot_kwargs : dict\n",
     "        Kwargs passed to the plot function.\n",
+    "    proportional: bool\n",
+    "        A boolean flag to determine if the plot is a proportional plot.\n",
     "    horizontal : bool\n",
     "        A boolean flag to determine if the plot is for horizontal plotting.\n",
     "    \"\"\"\n",
@@ -1087,6 +1096,9 @@
     "            max_x = contrast_axes.get_ylim()[1]\n",
     "            rawdata_axes.set_ylim(-0.375, max_x)\n",
     "\n",
+    "            if proportional:\n",
+    "                rawdata_axes.set_ylim(-0.375, max_x+0.1)\n",
+    "\n",
     "        if show_delta2 or show_mini_meta:\n",
     "            # Increase the ylim of raw data by 2\n",
     "            temp = rawdata_axes.get_ylim()\n",