linspace -> geomspace

core_algorithm/utils/response_plot.py

@@ -8,7 +8,6 @@
 
 
 def get_table_values(table, col_name, input_cnt):
-
     bar_values: dict[str:float] = {}
 
     col_index = None
@@ -32,7 +31,6 @@ def get_table_values(table, col_name, input_cnt):
 
 
 def plot_bars(filepath, plot_name, best_graph, table, units):
-
     input_cnt = len(best_graph.inputs)
     output_cnt = len(best_graph.outputs)
 
@@ -54,7 +52,7 @@ def plot_bars(filepath, plot_name, best_graph, table, units):
             outputs_dict[k]['prev_node'] = best_graph.find_prev(outputs_dict[k]['node']).name
 
     for k, o in outputs_dict.items():
-        if prev_node:=o['prev_node']:
+        if prev_node := o['prev_node']:
             outputs_dict[k]['name'] = prev_node
             outputs_dict[k]['scores_pre_HR'] = get_table_values(table, prev_node, input_cnt)
             outputs_dict[k]['scores_post_HR'] = get_table_values(table, k, input_cnt)
@@ -64,8 +62,9 @@ def plot_bars(filepath, plot_name, best_graph, table, units):
 
     # fig = plt.subplots()
     px = 1 / plt.rcParams['figure.dpi']
-    plt.figure(figsize=(px*(640 + 180*input_cnt), px*(120 + output_cnt*(280 + 120*input_cnt))))
-    plt.subplots_adjust(top=0.8 + (0.03*output_cnt), hspace=0.5, wspace=0.5)
+    plt.figure(
+        figsize=(px * (640 + 180 * input_cnt), px * (120 + output_cnt * (280 + 120 * input_cnt))))
+    plt.subplots_adjust(top=0.8 + (0.03 * output_cnt), hspace=0.5, wspace=0.5)
     if plot_name.endswith('.UCF'):
         plot_name = plot_name[:-4]
     plt.suptitle(f'{plot_name}: Response Plots', fontsize=15, y=.95, fontweight='bold')
@@ -80,38 +79,39 @@ def plot_bars(filepath, plot_name, best_graph, table, units):
     for output, data in outputs_dict.items():
 
         # Output Response Plots (all outputs)
-        ax1 = plt.subplot(output_cnt, 3, 3*i - 2)
+        ax1 = plt.subplot(output_cnt, 3, 3 * i - 2)
         ax1.set_title(f'{data["name"]}', fontsize=12)
-        plt.bar(list(data['scores_pre_HR'].keys()), list(data['scores_pre_HR'].values()), color='gray')
+        plt.bar(list(data['scores_pre_HR'].keys()), list(data['scores_pre_HR'].values()),
+                color='gray')
         plt.yscale('log')
-        plt.ylim(ax1.get_ylim()[0]*0.1, ax1.get_ylim()[1]*10)
+        plt.ylim(ax1.get_ylim()[0] * 0.1, ax1.get_ylim()[1] * 10)
         plt.ylabel(f'Output ({units})', fontsize=12, fontweight='bold')
 
         # Outputs with Hill Response functions (i.e. Communication Molecules) only...
         if data['prev_node']:
 
             # Hill Response Curve
-            curve = lambda ymax, ymin, kd, n, x : ymin + (ymax - ymin) / (1.0 + (kd / x)**n)
-            ax2 = plt.subplot(output_cnt, 3, 3*i - 1)
-            x = np.linspace(ax1.get_ylim()[0], ax1.get_ylim()[1], 1000)
+            curve = lambda ymax, ymin, kd, n, x: ymin + (ymax - ymin) / (1.0 + (kd / x) ** n)
+            ax2 = plt.subplot(output_cnt, 3, 3 * i - 1)
+            # for logspace, need ints corresponding to powers; geom uses real vals as endpoints
+            x = np.geomspace(ax1.get_ylim()[0] * 0.1, ax1.get_ylim()[1], 10000)
             ax2.plot(x, curve(data['params']['ymax'], data['params']['ymin'], data['params']['kd'],
                               data['params']['n'], x))
             ax2.set_title(f'{output} Hill', fontsize=12)
-            plt.yscale('log', )
+            plt.yscale('log')
             plt.xscale('log')
             if i == last:
                 plt.xlabel(f'Input ({units})', fontsize=12, fontweight='bold')
 
             # Promoter Activity Bars
-            ax3 = plt.subplot(output_cnt, 3, 3*i)
+            ax3 = plt.subplot(output_cnt, 3, 3 * i)
             ax3.sharey(ax2)
             ax3.set_title(f'{output} Promoter', fontsize=12)
             plt.bar(list(data['scores_post_HR'].keys()), list(data['scores_post_HR'].values()))
             plt.yscale('log')
 
         i += 1
 
-
     # Suppress (useless) console output
     matplotlib_logger = logging.getLogger("matplotlib")
     matplotlib_logger.setLevel(logging.INFO)