Calculate Chi^2 values for getAdj- formulas

copycat/copycat.py

@@ -67,6 +67,10 @@ class Copycat(object):
 
     def run(self, initial, modified, target, iterations, testAdjFormulas=False):
         self.workspace.resetWithStrings(initial, modified, target)
+
+        # I (LSaldyt) am very sorry for writing code like this.
+        # It will soon be deleted. I promise.
+
         if testAdjFormulas:
             formulas = self.temperature.adj_formulas()
         else:
@@ -90,9 +94,9 @@ class Copycat(object):
             for answer, d in answers.items():
                 d['avgtemp'] = d.pop('sumtemp') / d['count']
                 d['avgtime'] = d.pop('sumtime') / d['count']
-            formulaList.append(answers)
+            formulaList.append((formula, answers))
         if not testAdjFormulas:
-            return formulaList[0]
+            return formulaList[0][1]
         else:
             return formulaList
 

multi-run.py

@@ -1,6 +1,7 @@
 #!/usr/bin/env python3
 import argparse, logging
 from copycat import Copycat, Reporter, plot_answers, save_answers
+from collections import defaultdict
 
 class SimpleReporter(Reporter):
     """Reports results from a single run."""
@@ -22,9 +23,30 @@ def main():
             line = line.replace('\n', '')
             a, b, c = line.split(',')
             answerList = copycat.run(a, b, c, options.iterations, True)
-            for answers in answerList:
+            results = dict()
-                for answer, d in sorted(iter(answers.items()), key=lambda kv: kv[1]['avgtemp']):
+            for formula, answers in answerList:
-                    print('%s: %d (avg time %.1f, avg temp %.1f)' % (answer, d['count'], d['avgtime'], d['avgtemp']))
+                answers = sorted(answers.items(), key=lambda kv : kv[1]['count'])
+                keys    = [k for k, v in answers]
+                counts  = [v['count'] for k, v in answers]
+                results[formula] = (keys, counts)
+
+            originalCounts = defaultdict(lambda : 0)
+            originalCounts.update(dict(zip(*results['original'])))
+            
+            for formula, (keys, counts) in results.items():
+                if formula != 'original':
+                    chi2 = 0
+                    for answer, count in zip(keys, counts):
+                        originalCount = originalCounts[answer] 
+                        if originalCount != 0:
+                            chi2 += (count + originalCount) ** 2 / originalCount
+                    print((formula, chi2))
+
+
+
+                #for answer, d in sorted(iter(answers.items()), key=lambda kv: kv[1]['avgtemp']):
+                    #print('%s: %d (avg time %.1f, avg temp %.1f)' % (answer, d['count'], d['avgtime'], d['avgtemp']))
+
 
             #filename = 'output/{}-{}-{}.csv'.format(a, b, c)
             #save_answers(answers, filename)