Added ExactConceptParser

parsers/ExactConceptParser.py

@@ -0,0 +1,107 @@
+from core.sheerka import ReturnValue
+from parsers.BaseParser import BaseParser
+from core.tokenizer import Tokenizer, Keywords, TokenKind
+from core.concept import Concept
+
+
+class ExactConceptParser(BaseParser):
+    """
+    Tries to recognize a single concept
+    """
+
+    MAX_WORDS_SIZE = 10
+
+    def __init__(self):
+        BaseParser.__init__(self, "ConceptParser")
+
+    def parse(self, context, text):
+        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))
+
+        recognized = False
+        for combination in self.combinations(words):
+            concept_key = " ".join(combination)
+
+            # Very important question to think about later
+            # Must we return a new instance or the existing one
+            # 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):
+                # 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))
+                recognized = True
+
+        if recognized:
+            return res
+
+        return ReturnValue(self.name, False, sheerka.new(sheerka.UNKNOWN_CONCEPT_NAME, body=text))
+
+    @staticmethod
+    def get_words(text):
+        res = []
+        for t in iter(Tokenizer(text)):
+            if t.type == TokenKind.EOF:
+                break
+            if t.type == TokenKind.NEWLINE or t.type == TokenKind.WHITESPACE:
+                continue
+            res.append(t.value.value if isinstance(t.value, Keywords) else t.value)
+        return res
+
+    def combinations(self, iterable):
+        # combinations('foo', 'bar', 'baz') -->
+        # ('foo', 'bar', 'baz'),
+        # ('__var__0', 'bar', 'baz'),
+        # ('foo', '__var__0', 'baz'),
+        # ('foo', 'bar', '__var__0'),
+        # ('__var__0', '__var__1', 'baz'),
+        # ('__var__0', 'bar', '__var__1'),
+        # ('foo', '__var__0', '__var__1'),
+        # ('__var__0', '__var__1', '__var__2')]
+
+        pool = tuple(iterable)
+        n = len(pool)
+
+        res = set()
+
+        for r in range(0, n + 1):
+            indices = list(range(r))
+            res.add(self.get_tuple(pool, indices))
+            while True:
+                for i in reversed(range(r)):
+                    if indices[i] != i + n - r:
+                        break
+                else:
+                    break
+                indices[i] += 1
+                for j in range(i + 1, r):
+                    indices[j] = indices[j - 1] + 1
+                res.add(self.get_tuple(pool, indices))
+
+        return res
+
+    @staticmethod
+    def get_tuple(pool, indices):
+        res = []
+        vars = {}
+        k = 0
+
+        # init vars
+        for i in indices:
+            value = pool[i]
+            if value not in vars:
+                vars[pool[i]] = f"{Concept.PROPERTY_PREFIX}{k}"
+                k += 1
+
+        # create tuple
+        for i in range(len(pool)):
+            value = pool[i]
+            res.append(vars[value] if value in vars else value)
+        return tuple(res)