Implemented some enhancement request and fixed some bugs

Fixed #2 : Variables are not recognized when inside a rule token
Fixed #15 : Rule: rete attributes are lost when a new ontology is created
Fixed #14 : ReteNetwork: Format rules must not be added to Rete network
Fixed #16 : DefConcept: Variables are not recognized when they are keyword arguments
Fixed #4 : Comparison are not correctly set when comparison property is a concept
Fixed #14 : Parser: merge FunctionParser.NamesNode and ExpressionParser.NamesNode
Fixed #18 : Parser: Add SourceCodeNode test to UnrecognizedNodeParser
Fixed #20 : At startup Number concept is saved in db a numerous number of time
Fixed #21 : CacheManager: I can remove all elements from a ListIfNeededCache and fill it again
Fixed #22 : CacheManager: I can remove all elements from a SetCache and fill it again
Fixed #23 : HistoryManager: history() no longer works
Fixed #24 : HistoryManager: history() no longer works after creating an exec rule
Fixed #25 : SheerkaMemory: Use MemoryObject instead of sheerka.local
Fixed #26 : Debugger: add the list all available services..
Fixed #27 : CONCEPTS_GRAMMARS_ENTRY does not seems to be in use any more
Fixed #28 : Give order to services

src/evaluators/AddToMemoryEvaluator.py

@@ -23,11 +23,5 @@ class AddToMemoryEvaluator(OneReturnValueEvaluator):
         return len(context.sheerka.services[SheerkaMemory.NAME].registration) > 0
 
     def eval(self, context, return_value):
-        if context.sheerka.during_initialisation:
-            from core.sheerka.services.SheerkaMemory import SheerkaMemory
-            service = context.sheerka.services[SheerkaMemory.NAME]
-            service.registration.clear()
-            return None
-
         context.sheerka.commit_registered_objects(context)
         return None  # no need to have a second pass

src/evaluators/DefConceptEvaluator.py

@@ -168,9 +168,16 @@ class DefConceptEvaluator(OneReturnValueEvaluator):
             source = ret_value.value.source.as_text() if isinstance(ret_value.value.source,
                                                                     ParserInput) else ret_value.value.source
             tokens = ret_value.value.tokens or list(Tokenizer(source, yield_eof=False))
-            tokens = [t.str_value for t in tokens]
+            possible_vars = set()
+            for t in tokens:
+                if t.type == TokenKind.RULE:
+                    for v in [v for v in t.value if v is not None]:
+                        possible_vars.add(v)
+                else:
+                    possible_vars.add(t.str_value)
+
             for identifier in [i for i in concept_name if str(i).isalnum()]:
-                if identifier in tokens:
+                if identifier in possible_vars:
                     variables.add(identifier)
             debugger.debug_var("names", variables, hint="from concept")
             return variables

src/evaluators/PythonEvaluator.py

@@ -11,9 +11,10 @@ from core.concept import ConceptParts, Concept
 from core.global_symbols import NotInit, NotFound
 from core.rule import Rule
 from core.sheerka.ExecutionContext import ExecutionContext
+from core.sheerka.services.SheerkaMemory import SheerkaMemory
 from core.tokenizer import Token, TokenKind
 from evaluators.BaseEvaluator import OneReturnValueEvaluator
-from parsers.PythonParser import PythonNode, get_python_node
+from parsers.PythonParser import PythonNode
 
 TO_DISABLED = ["breakpoint", "callable", "compile", "delattr", "eval", "exec", "exit", "input", "locals", "open",
                "print", "quit", "setattr"]
@@ -97,7 +98,7 @@ class PythonEvaluator(OneReturnValueEvaluator):
 
         debugger = context.get_debugger(PythonEvaluator.NAME, "eval")
         debugger.debug_entering(node=node)
-        exception_debugger = context.get_debugger("Exceptions", PythonEvaluator.NAME + ".eval")
+        exception_debugger = context.get_debugger("Exceptions", PythonEvaluator.NAME + "-eval")
         get_trace_back = exception_debugger.is_enabled()
 
         context.log(f"Evaluating python node {node}.", self.name)
@@ -140,8 +141,8 @@ class PythonEvaluator(OneReturnValueEvaluator):
                     context.log("Evaluating using 'exec'.", self.name)
                     evaluated = self.exec_with_return(node.ast_, globals_, my_locals)
 
-                # TODO find a better implementation using SheerkaMemory
-                sheerka.locals.update(my_locals)
+                for k, v in my_locals.items():
+                    sheerka.services[SheerkaMemory.NAME].add_to_memory(context, k, v)
 
                 if not expect_success or evaluated:
                     break  # in this first version, we stop once a success is found
@@ -231,17 +232,18 @@ class PythonEvaluator(OneReturnValueEvaluator):
                 my_globals[name] = Expando("sheerka", bag)
                 continue
 
-            # search in local variables. To remove when local variables will be merged with memory
-            if name in context.sheerka.locals:
-                my_globals[name] = context.sheerka.locals[name]
-                continue
-
             # search in short term memory
             if (obj := context.get_from_short_term_memory(name)) is not NotFound:
                 context.log(f"Resolving '{name}'. Using value found in STM.", self.name)
                 my_globals[name] = obj
                 continue
 
+            # search in memory
+            if (obj := context.sheerka.get_last_from_memory(context, name)) is not NotFound:
+                context.log(f"Resolving '{name}'. Using value found in Long Term Memory.", self.name)
+                my_globals[name] = obj.obj
+                continue
+
             # search in sheerka methods
             if (method := self.get_sheerka_method(context, name, expression_only)) is not None:
                 my_globals[name] = method