Refactored to allow ConceptEvaluator

This commit is contained in:
2019-11-14 22:04:38 +01:00
parent 576ce77740
commit 9e10e77737
30 changed files with 2406 additions and 1007 deletions
+18 -6
View File
@@ -1,7 +1,10 @@
from core.sheerka import ReturnValue
from core.builtin_concepts import ReturnValueConcept, BuiltinConcepts
from parsers.BaseParser import BaseParser
from core.tokenizer import Tokenizer, Keywords, TokenKind
from core.concept import Concept
import logging
log = logging.getLogger(__name__)
class ExactConceptParser(BaseParser):
@@ -15,11 +18,17 @@ class ExactConceptParser(BaseParser):
BaseParser.__init__(self, "ConceptParser")
def parse(self, context, text):
"""
text can be string, but text can also be an list of tokens
:param context:
:param text:
:return:
"""
res = []
sheerka = context.sheerka
words = self.get_words(text)
if len(words) > self.MAX_WORDS_SIZE:
return ReturnValue(self.name, False, sheerka.new(sheerka.CONCEPT_TOO_LONG_CONCEPT_NAME))
return ReturnValueConcept(self.name, False, sheerka.new(BuiltinConcepts.CONCEPT_TOO_LONG, obj=text))
recognized = False
for combination in self.combinations(words):
@@ -30,24 +39,27 @@ class ExactConceptParser(BaseParser):
# That will depend on the context
# Let's return a new one for now and see if it works
concept = sheerka.new(concept_key)
if not sheerka.isinstance(concept, sheerka.UNKNOWN_CONCEPT_NAME):
if not sheerka.isinstance(concept, BuiltinConcepts.UNKNOWN_CONCEPT):
# update the properties if needed
for i, token in enumerate(combination):
if token.startswith(Concept.PROPERTY_PREFIX):
index = int(token[len(Concept.PROPERTY_PREFIX):])
concept.set_prop_by_index(index, words[i])
res.append(ReturnValue(self.name, True, concept))
res.append(ReturnValueConcept(self.name, True, concept))
log.debug(f"Recognized '{text}' as '{concept}'")
recognized = True
if recognized:
return res
return ReturnValue(self.name, False, sheerka.new(sheerka.UNKNOWN_CONCEPT_NAME, body=text))
log.debug(f"Failed to recognize {words}")
return ReturnValueConcept(self.name, False, sheerka.new(BuiltinConcepts.UNKNOWN_CONCEPT, obj=text))
@staticmethod
def get_words(text):
tokens = iter(Tokenizer(text)) if isinstance(text, str) else text
res = []
for t in iter(Tokenizer(text)):
for t in tokens:
if t.type == TokenKind.EOF:
break
if t.type == TokenKind.NEWLINE or t.type == TokenKind.WHITESPACE: