Files
Sheerka-Old/tests/test_DefaultParser.py
T

362 lines
11 KiB
Python

import pytest
import ast
from core.builtin_concepts import ParserResultConcept, BuiltinConcepts, ReturnValueConcept
from core.concept import Concept
from core.sheerka import Sheerka, ExecutionContext
from parsers.ConceptLexerParser import OrderedChoice, StrMatch, ConceptMatch, RegexParser
from parsers.PythonParser import PythonParser, PythonNode
from core.tokenizer import Keywords, Tokenizer
from parsers.DefaultParser import DefaultParser, NameNode, SyntaxErrorNode
from parsers.DefaultParser import UnexpectedTokenErrorNode, DefConceptNode
# def nop():
# return NopNode()
#
#
# def n(number):
# return NumberNode([], number)
#
#
# def s(string, quote="'"):
# return StringNode([], string, quote)
#
#
# def v(name):
# return VariableNode([], name)
#
#
# def t():
# return TrueNode([])
#
#
# def f():
# return FalseNode([])
#
#
# def null():
# return NullNode([])
#
#
# def b(operator, left, right):
# return BinaryNode([], operator, left, right)
#
# def compare_ast(left, right):
# left_as_string = ast.dump(left)
# left_as_string = left_as_string.replace(", ctx=Load()", "")
# left_as_string = left_as_string.replace(", kind=None", "")
#
# right_as_string = right if isinstance(right, str) else ast.dump(right)
# right_as_string = right_as_string.replace(", ctx=Load()", "")
# right_as_string = right_as_string.replace(", kind=None", "")
#
# return left_as_string == right_as_string
#
def get_concept(name, where=None, pre=None, post=None, body=None, definition=None):
concept = DefConceptNode([], name=NameNode(list(Tokenizer(name))))
if body:
concept.body = get_concept_part(body)
if where:
concept.where = get_concept_part(where)
if pre:
concept.pre = get_concept_part(pre)
if post:
concept.post = get_concept_part(post)
if definition:
concept.definition = ReturnValueConcept(
"Parsers:RegexParser",
True,
definition)
return concept
def get_context():
sheerka = Sheerka(skip_builtins_in_db=True)
sheerka.initialize("mem://")
return ExecutionContext("test", "xxx", sheerka)
def get_concept_part(part):
if isinstance(part, str):
node = PythonNode(part, ast.parse(part, mode="eval"))
return ReturnValueConcept(
who="Parsers:DefaultParser",
status=True,
value=ParserResultConcept(
source=part,
parser=PythonParser(),
value=node))
if isinstance(part, PythonNode):
return ReturnValueConcept(
who="Parsers:DefaultParser",
status=True,
value=ParserResultConcept(
source=part.source,
parser=PythonParser(),
value=part))
if isinstance(part, ReturnValueConcept):
return part
# @pytest.mark.parametrize("text, expected", [
# ("1", n(1)),
# ("+1", n(1)),
# ("-1", n(-1)),
# ("'foo'", s("foo")),
# ("identifier", v("identifier")),
# ("true", t()),
# ("false", f()),
# ("null", null()),
# ("1 * 2", b(TokenKind.STAR, n(1), n(2))),
# ("1 * 2/3", b(TokenKind.STAR, n(1), b(TokenKind.SLASH, n(2), n(3)))),
# ("1 + 2", b(TokenKind.PLUS, n(1), n(2))),
# ("1 + 2 - 3", b(TokenKind.PLUS, n(1), b(TokenKind.MINUS, n(2), n(3)))),
# ("1 + 2-3", b(TokenKind.PLUS, n(1), b(TokenKind.PLUS, n(2), n(-3)))),
# ("1 + 2 +-3", b(TokenKind.PLUS, n(1), b(TokenKind.PLUS, n(2), n(-3)))),
# ("1 + 2 * 3", b(TokenKind.PLUS, n(1), b(TokenKind.STAR, n(2), n(3)))),
# ("1 * 2 + 3", b(TokenKind.PLUS, b(TokenKind.STAR, n(1), n(2)), n(3))),
# ("(1 + 2) * 3", b(TokenKind.STAR, b(TokenKind.PLUS, n(1), n(2)), n(3))),
# ("1 * (2 + 3)", b(TokenKind.STAR, n(1), b(TokenKind.PLUS, n(2), n(3)))),
# ])
# def test_i_can_parse_simple_expression(text, expected):
# parser = DefaultParser(text, None)
# ast = parser.parse()
# assert ast.is_same(expected)
#
#
# @pytest.mark.parametrize("text, token_found, expected_tokens", [
# ("1+", TokenKind.EOF,
# [TokenKind.NUMBER, TokenKind.STRING, TokenKind.IDENTIFIER, 'true', 'false', 'null', TokenKind.LPAR]),
# ("(1+1", TokenKind.EOF, [TokenKind.RPAR])
# ])
# def test_i_can_detect_unexpected_end_of_code(text, token_found, expected_tokens):
# parser = DefaultParser(text, None)
# parser.parse()
#
# assert parser.has_error
# assert parser.error_sink[0].tokens[0].type == token_found
# assert parser.error_sink[0].expected_tokens == expected_tokens
@pytest.mark.parametrize("text, expected", [
("def concept hello", get_concept(name="hello")),
("def concept hello ", get_concept(name="hello")),
("def concept a + b", get_concept(name="a + b")),
("def concept a+b", get_concept(name="a + b")),
("def concept 'a+b'+c", get_concept(name="'a+b' + c")),
("def concept 'as if'", get_concept(name="'as if'")),
("def concept 'as' if", get_concept(name="'as if'")),
("def concept hello as 'hello'", get_concept(name="hello", body="'hello'")),
("def concept hello as 1", get_concept(name="hello", body="1")),
("def concept hello as 1 + 1", get_concept(name="hello", body="1 + 1")),
])
def test_i_can_parse_def_concept(text, expected):
parser = DefaultParser()
res = parser.parse(get_context(), text)
node = res.value.value
assert res.status
assert res.who == parser.name
assert res.value.source == text
assert isinstance(res.value, ParserResultConcept)
assert node == expected
def test_i_can_parse_complex_def_concept_statement():
text = """def concept a plus b
where a,b
pre isinstance(a, int) and isinstance(b, float)
post isinstance(res, int)
as res = a + b
"""
parser = DefaultParser()
res = parser.parse(get_context(), text)
return_value = res.value
expected_concept = get_concept(
name="a plus b",
where="a,b",
pre="isinstance(a, int) and isinstance(b, float)",
post="isinstance(res, int)",
body=PythonNode("res = a + b", ast.parse("res = a + b", mode="exec"))
)
assert res.status
assert isinstance(return_value, ParserResultConcept)
assert return_value.value == expected_concept
def test_i_can_have_mutilines_declarations():
text = """
def concept add one to a as
def func(x):
return x+1
func(a)
"""
expected_concept = get_concept(
name="add one to a ",
body=PythonNode(
"def func(x):\n return x+1\nfunc(a)",
ast.parse("def func(x):\n return x+1\nfunc(a)", mode="exec"))
)
parser = DefaultParser()
res = parser.parse(get_context(), text)
return_value = res.value
assert res.status
assert isinstance(return_value, ParserResultConcept)
assert return_value.value == expected_concept
def test_i_can_use_colon_to_use_indentation():
text = """
def concept add one to a as:
def func(x):
return x+1
func(a)
"""
expected_concept = get_concept(
name="add one to a ",
body=PythonNode(
"def func(x):\n return x+1\nfunc(a)",
ast.parse("def func(x):\n return x+1\nfunc(a)", mode="exec"))
)
parser = DefaultParser()
res = parser.parse(get_context(), text)
return_value = res.value
assert res.status
assert isinstance(return_value, ParserResultConcept)
assert return_value.value == expected_concept
def test_indentation_is_mandatory_after_a_colon():
text = """
def concept add one to a as:
def func(x):
return x+1
func(a)
"""
parser = DefaultParser()
res = parser.parse(get_context(), text)
return_value = res.value
assert not res.status
assert isinstance(return_value, ParserResultConcept)
assert isinstance(return_value.value[0], SyntaxErrorNode)
assert return_value.value[0].message == "Indentation not found."
def test_indentation_is_not_allowed_if_the_colon_is_missing():
text = """
def concept add one to a as
def func(x):
return x+1
func(a)
"""
context = get_context()
sheerka = context.sheerka
parser = DefaultParser()
res = parser.parse(context, text)
return_value = res.value
assert not res.status
assert isinstance(return_value, ParserResultConcept)
assert sheerka.isinstance(return_value.value[0], BuiltinConcepts.TOO_MANY_ERRORS)
def test_name_is_mandatory():
text = "def concept as 'hello'"
parser = DefaultParser()
res = parser.parse(get_context(), text)
return_value = res.value
assert not res.status
assert isinstance(return_value, ParserResultConcept)
assert isinstance(return_value.value[0], SyntaxErrorNode)
assert return_value.value[0].message == "Name is mandatory"
def test_concept_keyword_is_mandatory_but_the_concept_is_recognized():
text = "def hello as a where b pre c post d"
expected_concept = get_concept(name="hello", body="a", where="b", pre="c", post="d")
parser = DefaultParser()
res = parser.parse(get_context(), text)
return_value = res.value
assert not res.status
assert isinstance(return_value, ParserResultConcept)
assert isinstance(return_value.value[0], UnexpectedTokenErrorNode)
assert return_value.value[0].message == "Syntax error."
assert return_value.value[0].expected_tokens == [Keywords.CONCEPT]
assert return_value.try_parsed == expected_concept
@pytest.mark.parametrize("text", [
"def concept hello where 1+",
"def concept hello pre 1+",
"def concept hello post 1+",
"def concept hello as 1+"
])
def test_i_can_detect_error_in_declaration(text):
context = get_context()
sheerka = context.sheerka
parser = DefaultParser()
res = parser.parse(context, text)
return_value = res.value
assert not res.status
assert isinstance(return_value, ParserResultConcept)
assert sheerka.isinstance(return_value.value[0], BuiltinConcepts.TOO_MANY_ERRORS)
def test_new_line_is_not_allowed_in_the_name():
text = "def concept hello \n my friend as 'hello'"
parser = DefaultParser()
res = parser.parse(get_context(), text)
return_value = res.value
assert not res.status
assert return_value.value == [SyntaxErrorNode([], "Newline are not allowed in name.")]
def test_i_can_parse_def_concept_from_regex():
text = "def concept name from bnf a_concept | 'a_string' as __definition[0]"
parser = DefaultParser()
res = parser.parse(get_context(), text)
node = res.value.value
definition = OrderedChoice(ConceptMatch("a_concept"), StrMatch("a_string"))
parser_result = ParserResultConcept(RegexParser(), "a_concept | 'a_string'", definition, definition)
expected = get_concept(name="name", body="__definition[0]", definition=parser_result)
assert res.status
assert res.who == parser.name
assert res.value.source == text
assert isinstance(res.value, ParserResultConcept)
assert node == expected
def test_i_can_detect_empty_bnf_declaration():
text = "def concept name from bnf as __definition[0]"
parser = DefaultParser()
res = parser.parse(get_context(), text)
assert not res.status
assert res.value.value[0] == SyntaxErrorNode([], "Empty declaration")