Merge pull request #153 from lambda-feedback/tr129-experiment-with-syntactical-equivalence

Further attempt at generating correct feedback string

Author: KarlLundengaard

app/criteria_graph_utilities.py

@@ -29,10 +29,13 @@ def __init__(self, label, summary, details, evaluate, replacement=None):
             return
 
     class Criterion(Node):
-        def __init__(self, label, summary, details, tags=None, feedback_string_generator=lambda x: None):
+        def __init__(self, label, summary, details, tags=None, feedback_string_generator=None):
             super().__init__(label, summary, details)
             self.consequences = self.outgoing
-            self.feedback_string_generator = feedback_string_generator
+            if feedback_string_generator is not None:
+                self.feedback_string_generator = feedback_string_generator
+            else:
+                self.feedback_string_generator = lambda x: None
             if tags is None:
                 self.tags = set()
             else:
@@ -210,7 +213,7 @@ def add_criterion_node(self, label, summary, details, sufficiencies=None, evalua
             raise Exception(f"Criterion node {label} is already defined.")
         if sufficiencies is not None:
             raise Exception(f"Criterion nodes cannot have sufficiencies.")
-        node = CriteriaGraph.Criterion(label, summary, details, feedback_string_generator)
+        node = CriteriaGraph.Criterion(label, summary=summary, details=details, feedback_string_generator=feedback_string_generator)
         self.criteria.update({label: node})
         self.sufficiencies.update({label: sufficiencies})
         return node
@@ -268,7 +271,7 @@ def attach(self, source_label, target_label, summary=None, details=None, suffici
                     if summary is None or details is None:
                         raise Exception(f"Unknown node {target_label}. If you wish to create a new node summary and details must be specified.")
                     else:
-                        target = target_generator(target_label, summary, details, sufficiencies, evaluate)
+                        target = target_generator(target_label, summary=summary, details=details, sufficiencies=sufficiencies, evaluate=evaluate, feedback_string_generator=feedback_string_generator)
                 else:
                     raise Exception(f"Both {source_label} and {target_label} are {type_name} nodes. Only {other_type_name} nodes can be attached to {type_name} nodes.")
 

app/feedback/symbolic_comparison.py

@@ -1,25 +1,25 @@
-from ..criteria_utilities import Criterion, no_feedback
+from ..criteria_utilities import Criterion
 
 # TODO: Find better way of identifying reference criteria
 # equivalences dictionary should contain a list of variations that are likely to be produced by the following procedure:
 # - rewrite critera as expr=0,
 # - parse left hand side of rewritten critera as a sympy expression
 # - turn sympy expression into a string and remove all whitespace
-equivalences = dict()  
+equivalences = dict()
 criteria = dict()
 
 criteria["RESPONSE_EQUAL_ANSWER"] = Criterion("response=answer")
-equivalences.update({"RESPONSE_EQUAL_ANSWER": ["response=answer","answer=response","answer-response=0","-answer+response=0","answer/response=1","response/answer-1=0"]})
+equivalences.update({"RESPONSE_EQUAL_ANSWER": ["response=answer", "answer=response", "answer-response=0", "-answer+response=0", "answer/response=1", "response/answer-1=0"]})
 criteria["RESPONSE_EQUAL_ANSWER"][True] = lambda inputs: "The response matches the expected answer."
 criteria["RESPONSE_EQUAL_ANSWER"][False] = lambda inputs: "The response does not match the expected answer."
 
 criteria["RESPONSE_DOUBLE_ANSWER"] = Criterion("response=2*answer")
-equivalences.update({"RESPONSE_DOUBLE_ANSWER": ["response=2*answer","response/answer=2","2*answer=response","answer=response/2","answer-response/2","-answer+response/2","-2*answer+response","2*answer-response","-2+answer/response","-2+response/answer","answer-1*response/2","-answer+1*response/2","-2+1*answer/response","-2+1*response/answer"]})
+equivalences.update({"RESPONSE_DOUBLE_ANSWER": ["response=2*answer","response/answer=2", "2*answer=response", "answer=response/2", "answer-response/2", "-answer+response/2", "-2*answer+response", "2*answer-response", "-2+answer/response", "-2+response/answer", "answer-1*response/2", "-answer+1*response/2", "-2+1*answer/response", "-2+1*response/answer"]})
 criteria["RESPONSE_DOUBLE_ANSWER"][True] = lambda inputs: "The response is the expected answer multiplied by 2."
 criteria["RESPONSE_DOUBLE_ANSWER"][False] = lambda inputs: "The response is not the expected answer multiplied by 2."
 
 criteria["RESPONSE_NEGATIVE_ANSWER"] = Criterion("response=-answer")
-equivalences.update({"RESPONSE_NEGATIVE_ANSWER": ["response=-answer","answer=-response","answer+response=0","answer+response","answer/response=-1","response/answer+1"]})
+equivalences.update({"RESPONSE_NEGATIVE_ANSWER": ["response=-answer", "answer=-response", "answer+response=0", "answer+response","answer/response=-1", "response/answer+1"]})
 criteria["RESPONSE_NEGATIVE_ANSWER"][True] = lambda inputs: "The response is the expected answer multiplied by -1."
 criteria["RESPONSE_NEGATIVE_ANSWER"][False] = lambda inputs: "The response is not the expected answer multiplied by -1."
 
@@ -36,7 +36,7 @@
     "EXPRESSION_NOT_EQUALITY": "The response was an expression but was expected to be an equality.",
     "EQUALITY_NOT_EXPRESSION": "The response was an equality but was expected to be an expression.",
     "WITHIN_TOLERANCE": "",  # "The difference between the response the answer is within specified error tolerance.",
-    "NOT_NUMERICAL": "", #"The expression cannot be evaluated numerically.",
+    "NOT_NUMERICAL": "",  # "The expression cannot be evaluated numerically.",
 }
 
 # Format for feedback string entry: criteria["eval_tag"]("criteria_tag", inputs) = "formatted string" | None
@@ -65,7 +65,7 @@
     "EXPRESSION_NOT_EQUALITY": "The response was an expression but was expected to be an equality.",
     "EQUALITY_NOT_EXPRESSION": "The response was an equality but was expected to be an expression.",
     "WITHIN_TOLERANCE": None,  # "The difference between the response the answer is within specified error tolerance.",
-    "NOT_NUMERICAL": None, #"The expression cannot be evaluated numerically.",
+    "NOT_NUMERICAL": None,  #"The expression cannot be evaluated numerically.",
 }[tag]
 feedback_generators["GENERIC"] = lambda tag: lambda inputs: {
     "TRUE": None,
@@ -98,7 +98,7 @@
     "FALSE": "The response can be simplified further.",
 }[tag]
 feedback_generators["SAME_FORM"] = lambda tag: lambda inputs: {
-    "CARTESIAN": "Response and answer are both written on Cartesian form", # None,
-    "EXPONENTIAL": "Response and answer are both written on exponential form", # None,
+    "CARTESIAN": "Response and answer are both written on Cartesian form",  # None,
+    "EXPONENTIAL": "Response and answer are both written on exponential form",  # None,
     "UNKNOWN": "The response is not written on the expected form.",
-}[tag]
+}[tag]