progress monitor now shows current value on plot

gastop/progmon.py

@@ -11,27 +11,62 @@ class ProgMon():
     def __init__(self,progress_display,num_generations):
 
         self.progress_display = progress_display
+        self.num_gens = num_generations
+        self.orderofgen = int(np.log10(self.num_gens))*10
         self.pop_progress = []
+        self.pop_start = []
 
         if self.progress_display == 2:  # check if figure method of progress monitoring is requested
             # initialize plot:
             fig = plt.figure()
             self.ax1 = fig.add_subplot(1,1,1) #does this need self.?
-            plt.xlim(0, num_generations)
-            plt.ylabel('fitscore')
-            plt.xlabel('iteration')
-            plt.xlim(0,num_generations)
+            plt.xlim(0, self.num_gens)
+            plt.ylabel('Minimum Fitness Score')
+            plt.xlabel('Iteration')
+
+            #plt.yscale('log')
+            #
+            # self.ax2 = fig.add_subplot(1,2,2) #does this need self.?
+            # plt.xlim(0, num_generations)
+            # plt.ylabel('Min Fitness Score')
+            # plt.xlabel('Iteration')
+            # plt.xlim(0,num_generations)
         #
 
     def progress_monitor(self, current_gen, population):
 
         # three options: plot, progress bar ish thing, no output just append
         # calc population diversity and plot stuff or show current results
         fitscore = [i.fitness_score for i in population] #extract factor of safety from each truss object in population
+        fitscore_min = np.amin(fitscore)
+
         self.pop_progress.append(population) #append to history
-        if self.progress_display == 1:
-            print(current_gen,np.amin(fitscore))
-        elif self.progress_display == 2:
-            #print(current_gen,min(fitscore))
-            self.ax1.scatter(current_gen,np.amin(fitscore),c=[0,0,0]) #plot minimum fitscore for current gen in black
+        #if self.progress_display == 1:
+        #    test = np.amin(fitscore)
+        if self.progress_display == 2:
+            if current_gen==0:
+                self.pop_start = fitscore_min # store initial min fitscore (should be worst)
+            #     fitscore_range_scaled = 1.0
+            #     #self.pop_prop(current_gen) = (np.amax(fitscore) - np.amin(fitscore))/2.0
+            #     self.pop_prop.append([(np.amax(fitscore) - np.amin(fitscore))/2.0])
+            # else:
+            #     fitscore_range = (np.amax(fitscore) - np.amin(fitscore))/2.0
+            #     fitscore_range_scaled = fitscore_range/self.pop_prop[current_gen-1]
+            #     self.pop_prop.append([fitscore_range])
+            #     #self.pop_prop(current_gen) = fitscore_range
+            #
+            # fitscore_med = np.median(fitscore)
+            #self.pop_prop.append(fitscore_range)
+
+            #self.ax1.errorbar(current_gen, fitscore_med, yerr=fitscore_range_scaled, fmt='o',c=[0,0,0])
+            #self.ax1.scatter(current_gen,fitscore_med,c=[0,0,0])
+
+            self.ax1.scatter(current_gen,fitscore_min,c=[1,0,0])
+            # set text with current min fitscore
+            plot_text=self.ax1.text(self.num_gens-self.orderofgen, self.pop_start, round(fitscore_min,3),bbox=dict(facecolor='white', alpha=1))
+            # set box to same size
+            plot_text._bbox_patch._mutation_aspect = 0.1
+            plot_text.get_bbox_patch().set_boxstyle("square", pad=1)
+
+            #self.ax1.scatter(current_gen,np.amin(fitscore),c=[0,0,0]) #plot minimum fitscore for current gen in black
             plt.pause(0.0001) #pause for 0.0001s to allow plot to update, can potentially remove this

junk/susan_scratch.py

@@ -4,6 +4,7 @@
 from matplotlib import style
 import time
 from tqdm import tqdm
+import numpy as np
 
 ##plt.ion() #look into multithreading this
 #style.use('fivethirtyeight')
@@ -17,35 +18,56 @@ def run(self):
     print(y)
 
 def counter(n):
-    y = 0
+    y0 = np.array([1,2,3])
     style.use('fivethirtyeight')
     fig = plt.figure()
-    ax1 = fig.add_subplot(1,1,1)
+    ax1 = fig.add_subplot(1,2,1)
     plt.ylabel('convergence')
     plt.xlabel('iteration')
+    plt.xlim(0,n)
 
+    ax2 = fig.add_subplot(1,2,2)
+    plt.ylabel('test')
+    plt.xlabel('iteration')
+    #plt.yscale('log')
+    #plt.text(2,4,"0")
+    plt.xlim(0,n)
+    #plt.ylim(0,12)
+    test  = []
 
-    for i in tqdm(range(n)):
 
-        y = 2.0*i
-        progress(i,y,ax1)
-        #print(y)
+    for i in tqdm(range(n)):
+        y = 10.0-i*y0
+        if i == 0:
+            test = np.min(y)
+        # else:
+        #     test.append(test[i-1])
+        progress(i,y,ax1,ax2,n,test)
+        print(np.min(y))
     plt.show()
 
     return y
 
 
-def progress(i,y,ax1):
+def progress(i,y,ax1,ax2,n,test):
 #    if i==1:
         #style.use('fivethirtyeight')
         #fig = plt.figure()
         #ax1 = fig.add_subplot(1,1,1)
         #plt.ylabel('convergence')
         #plt.xlabel('iteration')
 #    else:
-    ax1.scatter(i,y,c=[0,0,0])
-    plt.pause(0.0001) #time it waits for plot to update
+    err_range = (np.amax(y) - np.amin(y))/2.0
+    ax1.errorbar(i, np.mean(y), yerr=err_range, fmt='o')
+    #ax2.text(n,4,np.amin(y),bbox=dict(facecolor='white', alpha=1))
+    ax2.text(n-1, test-1, np.amin(y),
+        bbox=dict(facecolor='white', alpha=1))
+    ax2.scatter(i,np.amin(y),c=[0,0,0])
+
+
+    #ax2.text()
+    plt.pause(0.5) #time it waits for plot to update
 
 
-y = counter(10)
+y = counter(4)
 print(y)