Files
Sheerka-Old/tests/parsers/test_PythonWithConceptsParser.py
kodjo 7dcaa9c111 Fixed #29: Parsers: Implement parsing memoization
Fixed #77 : Parser: ShortTermMemoryParser should be called separately
Fixed #78 : Remove VariableNode usage
Fixed #79 : ConceptManager: Implement compile caching
Fixed #80 : SheerkaExecute : parsers_key is not correctly computed
Fixed #81 : ValidateConceptEvaluator : Validate concept's where and pre clauses right after the parsing
Fixed #82 : SheerkaIsAManager: isa() failed when the set as a body
Fixed #83 : ValidateConceptEvaluator : Support BNF and SYA Concepts
Fixed #84 : ExpressionParser: Implement the parser as a standard parser
Fixed #85 : Services: Give order to services
Fixed #86 : cannot manage smart_get_attr(the short, color)
2021-06-07 21:14:03 +02:00

267 lines
12 KiB
Python

import ast
import pytest
from core.builtin_concepts import ParserResultConcept, BuiltinConcepts, ReturnValueConcept
from core.concept import Concept
from core.rule import Rule
from core.sheerka.services.SheerkaExecute import ParserInput
from core.tokenizer import Token, TokenKind, Tokenizer
from core.var_ref import VariableRef
from parsers.BaseNodeParser import ConceptNode, UnrecognizedTokensNode, RuleNode, VariableNode
from parsers.BnfNodeParser import BnfNodeParser
from parsers.PythonParser import PythonNode
from parsers.PythonWithConceptsParser import PythonWithConceptsParser
from parsers.SequenceNodeParser import SequenceNodeParser
from parsers.UnrecognizedNodeParser import UnrecognizedNodeParser
from tests.TestUsingMemoryBasedSheerka import TestUsingMemoryBasedSheerka
from tests.parsers.parsers_utils import get_source_code_node
unrecognized_nodes_parser = UnrecognizedNodeParser()
def ret_val(*args):
result = []
index = 0
for item in args:
if isinstance(item, Concept):
tokens = [Token(TokenKind.IDENTIFIER, item.name, 0, 0, 0)]
result.append(ConceptNode(item, index, index, tokens, item.name))
index += 1
elif isinstance(item, Rule):
tokens = [Token(TokenKind.RULE, (None, item.id), 0, 0, 0)]
result.append(RuleNode(item, index, index, tokens, f"r:|{item.id}:"))
index += 1
elif isinstance(item, VariableRef):
tokens = list(Tokenizer(item.prop, yield_eof=False))
result.append(VariableNode(item.obj, item.prop, index, index + len(tokens) - 1, tokens, f"{item.prop}"))
index += len(tokens)
else:
tokens = list(Tokenizer(item, yield_eof=False))
result.append(UnrecognizedTokensNode(index, index + len(tokens) - 1, tokens))
index += len(tokens)
return ReturnValueConcept("who", False, ParserResultConcept(parser=unrecognized_nodes_parser, value=result))
def to_str_ast(expression):
return PythonNode.get_dump(ast.parse(expression, mode="eval"))
class TestPythonWithConceptsParser(TestUsingMemoryBasedSheerka):
@pytest.mark.parametrize("text, interested", [
("not parser result", False),
(ParserInput("not parser result"), False),
(ParserResultConcept(parser="not multiple_concepts_parser"), False),
(ParserResultConcept(parser=unrecognized_nodes_parser, value=[UnrecognizedTokensNode(0, 0, [])]), True),
])
def test_not_interested(self, text, interested):
context = self.get_context()
res = PythonWithConceptsParser().parse(context, text)
if interested:
assert res is not None
else:
assert res is None
def test_i_can_parse_concepts_python_and_variable_ref(self):
context = self.get_context()
foo = Concept("foo")
input_return_value = ret_val(foo, " + 1 + ", VariableRef(foo, "var_name"))
parser = PythonWithConceptsParser()
result = parser.parse(context, input_return_value.body)
wrapper = result.value
return_value = result.value.value
assert result.status
assert result.who == parser.name
assert context.sheerka.isinstance(wrapper, BuiltinConcepts.PARSER_RESULT)
assert wrapper.source == "foo + 1 + var_name"
assert isinstance(return_value, PythonNode)
assert return_value.source == "__C__foo__C__ + 1 + __V__foo__var_name__V__"
assert return_value.original_source == "foo + 1 + var_name"
assert return_value.get_dump(return_value.ast_) == to_str_ast("__C__foo__C__ + 1 + __V__foo__var_name__V__")
assert return_value.objects == {"__C__foo__C__": foo,
"__V__foo__var_name__V__": VariableRef(foo, "var_name")}
def test_i_can_parse_concepts_and_python_when_concept_is_known(self):
context = self.get_context()
foo = Concept("foo")
foo = context.sheerka.create_new_concept(context, foo).body.body
input_return_value = ret_val(foo, " + 1")
parser = PythonWithConceptsParser()
result = parser.parse(context, input_return_value.body)
wrapper = result.value
return_value = result.value.value
assert result.status
assert result.who == parser.name
assert context.sheerka.isinstance(wrapper, BuiltinConcepts.PARSER_RESULT)
assert wrapper.source == "foo + 1"
assert isinstance(return_value, PythonNode)
assert return_value.source == "__C__foo__1001__C__ + 1"
assert return_value.original_source == "foo + 1"
assert return_value.get_dump(return_value.ast_) == to_str_ast("__C__foo__1001__C__ + 1")
assert return_value.objects["__C__foo__1001__C__"] == foo
def test_i_can_parse_when_concept_name_has_invalid_characters(self):
context = self.get_context()
foo = Concept("foo et > (,")
foo = context.sheerka.create_new_concept(context, foo).body.body
input_return_value = ret_val(foo, " + 1")
parser = PythonWithConceptsParser()
result = parser.parse(context, input_return_value.body)
return_value = result.value.value
assert result.status
assert return_value.objects["__C__foo0et000000__1001__C__"] == foo
def test_i_can_parse_when_multiple_concepts(self):
sheerka, context, foo, bar = self.init_concepts("foo", "bar")
input_return_value = ret_val("func(", foo, ", ", bar, ")")
parser = PythonWithConceptsParser()
result = parser.parse(context, input_return_value.body)
parser_result = result.value
return_value = result.value.value
assert result.status
assert result.who == parser.name
assert context.sheerka.isinstance(parser_result, BuiltinConcepts.PARSER_RESULT)
assert parser_result.source == "func(foo, bar)"
assert isinstance(return_value, PythonNode)
assert return_value.source == "func(__C__foo__1001__C__, __C__bar__1002__C__)"
assert return_value.original_source == "func(foo, bar)"
assert return_value.get_dump(return_value.ast_) == to_str_ast("func(__C__foo__1001__C__, __C__bar__1002__C__)")
assert return_value.objects["__C__foo__1001__C__"] == foo
assert return_value.objects["__C__bar__1002__C__"] == bar
def test_i_can_parse_when_python_and_rule(self):
context = self.get_context()
rule = Rule().set_id("rule_id")
input_return_value = ret_val(rule, " + 1")
parser = PythonWithConceptsParser()
result = parser.parse(context, input_return_value.body)
wrapper = result.value
return_value = result.value.value
assert result.status
assert result.who == parser.name
assert context.sheerka.isinstance(wrapper, BuiltinConcepts.PARSER_RESULT)
assert wrapper.source == "r:|rule_id: + 1"
assert isinstance(return_value, PythonNode)
assert return_value.source == "__R____rule_id__R__ + 1"
assert return_value.original_source == "r:|rule_id: + 1"
assert return_value.get_dump(return_value.ast_) == to_str_ast("__R____rule_id__R__ + 1")
assert return_value.objects["__R____rule_id__R__"] == rule
def test_python_ids_mappings_are_correct_when_rules_with_the_same_id(self):
context = self.get_context()
rule1 = Rule().set_id("rule_id")
rule2 = Rule().set_id("rule_id")
input_return_value = ret_val(rule1, "+", rule2)
parser = PythonWithConceptsParser()
result = parser.parse(context, input_return_value.body)
return_value = result.value.value
assert result.status
assert return_value.objects["__R____rule_id__R__"] == rule1
assert return_value.objects["__R____rule_id_1__R__"] == rule2
def test_python_ids_mappings_are_correct_when_concepts_with_the_same_name(self):
context = self.get_context()
foo1 = Concept("foo")
foo2 = Concept("foo")
foo3 = context.sheerka.create_new_concept(context, Concept("foo", body="foo3")).body.body
foo4 = context.sheerka.create_new_concept(context, Concept("foo", body="foo4")).body.body
input_return_value = ret_val(foo1, "+", foo2, "+", foo3, "+", foo4)
parser = PythonWithConceptsParser()
result = parser.parse(context, input_return_value.body)
return_value = result.value.value
assert result.status
assert return_value.objects["__C__foo__C__"] == foo1
assert return_value.objects["__C__foo_1__C__"] == foo2
assert return_value.objects["__C__foo__1001__C__"] == foo3
assert return_value.objects["__C__foo__1002__C__"] == foo4
def test_i_cannot_parse_if_syntax_error(self):
context = self.get_context()
foo = Concept("foo")
foo = context.sheerka.create_new_concept(context, foo).body.body
input_return_value = ret_val(foo, " + ")
parser = PythonWithConceptsParser()
result = parser.parse(context, input_return_value.body)
assert not result.status
assert context.sheerka.isinstance(result.value, BuiltinConcepts.NOT_FOR_ME)
def test_i_can_parse_nodes_when_source_code_node(self):
context = self.get_context()
foo = Concept("foo")
bar = Concept("bar")
nodes = [
UnrecognizedTokensNode(0, 1, list(Tokenizer("not ", yield_eof=False))),
get_source_code_node(2, "foo == 1", {"foo": foo}),
UnrecognizedTokensNode(7, 9, list(Tokenizer(" and ", yield_eof=False))),
get_source_code_node(10, "bar < 1", {"bar": bar}),
]
expected_ast = ast.parse('not __C__foo__C__ == 1 and __C__bar__C__ < 1', "<source>", 'eval')
parser = PythonWithConceptsParser()
result = parser.parse_nodes(context, nodes)
result_python_node = result.value.value
assert isinstance(result_python_node, PythonNode)
assert result_python_node.source == 'not __C__foo__C__ == 1 and __C__bar__C__ < 1'
assert result_python_node.ast_str == PythonNode.get_dump(expected_ast)
assert result_python_node.original_source == "not foo == 1 and bar < 1"
assert result_python_node.objects == {"__C__foo__C__": foo, "__C__bar__C__": bar}
def test_can_parse_after_unrecognized_bnf(self):
sheerka, context, one, two, twenties = self.init_concepts(
Concept("one", body="1"),
Concept("two", body="2"),
Concept("twenties", definition="'twenty' (one|two)=n", body='20 + n').def_var("n"),
create_new=True
)
bnf_parser_ret_val = BnfNodeParser().parse(context, ParserInput("a + twenty one"))
unrec_node_ret_val = UnrecognizedNodeParser().parse(context, bnf_parser_ret_val.body)
parser = PythonWithConceptsParser()
result = parser.parse(context, unrec_node_ret_val.body)
assert result.status
python_node = result.body.body
assert isinstance(python_node, PythonNode)
assert python_node.source == 'a + __C__twenties__1003__C__'
assert "__C__twenties__1003__C__" in python_node.objects
def test_i_cannot_parse_unrecognized_sequence(self):
sheerka, context, one, two, twenties = self.init_concepts(
Concept("one", body="1"),
Concept("two", body="2"),
Concept("twenties", definition="'twenty' (one|two)=n", body='20 + n').def_var("n"),
create_new=True
)
sequence_parser_ret_val = SequenceNodeParser().parse(context, ParserInput("a + twenty one"))
unrec_node_ret_val = UnrecognizedNodeParser().parse(context, sequence_parser_ret_val.body)
parser = PythonWithConceptsParser()
result = parser.parse(context, unrec_node_ret_val.body)
assert not result.status