add feedback mechanism for grammar and spelling errors

feature_extractor.py

@@ -43,7 +43,7 @@ class FeatureExtractor(object):
                 self._mean_spelling_errors=sum(e_set._spelling_errors)/float(len(e_set._spelling_errors))
                 self._spell_errors_per_character=sum(e_set._spelling_errors)/float(sum([len(t) for t in e_set._text]))
                 self._grammar_errors_per_character=sum(self._get_grammar_errors
-                    (e_set._pos,e_set._text,e_set._tokens))/float(len(e_set._text))
+                    (e_set._pos,e_set._text,e_set._tokens))/float(sum([len(t) for t in e_set._text]))
                 ret = "ok"
             else:
                 raise util_functions.InputError(e_set, "needs to be an essay set of the train type.")
@@ -156,4 +156,20 @@ class FeatureExtractor(object):
 
         prompt_arr = numpy.array((prompt_overlap, prompt_overlap_prop, expand_overlap, expand_overlap_prop)).transpose()
 
-        return prompt_arr.copy()
+        return prompt_arr.copy()
\ No newline at end of file
+
+    def gen_feedback(self, e_set):
+        set_grammar=self._get_grammar_errors(e_set._pos,e_set._text,e_set._tokens)
+        set_grammar_per_character=[set_grammar[m]/len(e_set._text[m]) for m in xrange(0,len(e_set._text))]
+        set_spell_errors_per_character=[e_set._spelling_errors[m]/len(e_set._text[m]) for m in xrange(0,len(e_set._text))]
+        e_set._spelling_errors
+        all_feedback=[]
+        for m in xrange(0,len(e_set._text)):
+            individual_feedback=[]
+            if set_grammar_per_character[m]<self._grammar_errors_per_character:
+                individual_feedback.append("Potential grammatical errors.")
+            if set_spell_errors_per_character[m]>self._spell_errors_per_character:
+                individual_feedback.append("Potential spelling errors.")
+            all_feedback.append(individual_feedback)
+
+        return all_feedback

grade.py

@@ -20,7 +20,7 @@ import feature_extractor
 import sklearn.ensemble
 
 def grade(grader_path,submission,sandbox):
-    results = {'errors': [],'tests': [],'correct': False,'score': 0}
+    results = {'errors': [],'tests': [],'correct': False,'score': 0, 'feedback' : []}
 
     #Try to find and load the model file
 
@@ -40,7 +40,7 @@ def grade(grader_path,submission,sandbox):
     #Try to extract features from submission and assign score via the model
     try:
         grader_feats=grader_data['extractor'].gen_feats(grader_set)
-
+        results['feedback']=grader_data['extractor'].gen_feedback(grader_set)
         results['score']=int(grader_data['model'].predict(grader_feats)[0])
     except:
         results['errors'].append("Could not extract features and score essay.")

tests/tmpfile

- in order to replicate this experiment i would need to know , first how to calculate the mass of the samples . i would also need t o know what the four samples are , as if is not stated in the procedure . lastly to repeat this expirement i would need to know how much vinegar was placed into each container . 
+ in order to replicate this experiment , additional information is needed . first , the amount of vinegar would be needed . also , the type of wood used would be needed because different woods have different densities . finally , the temperatures of the room , where the solutions sat , needs to be known . 
\ No newline at end of file