New text grammar

textworld/textgen/__init__.py

@@ -1,5 +1,5 @@
 from tatsu.model import NodeWalker
-from typing import Iterable, Optional, Tuple
+from typing import Iterable, Optional, Tuple, List
 
 from textworld.textgen.model import TextGrammarModelBuilderSemantics
 from textworld.textgen.parser import TextGrammarParser
@@ -18,32 +18,75 @@ def full_form(self, include_adj=True) -> str:
             return adj + "|" + noun
 
 
-class LiteralAlternative(Alternative):
+class LiteralChunk:
     """
-    An alternative from a literal string.
+    It creates an object with a [str] value for every single literal.
+    literal is defined as any string which is not a symbol, i.e. it is not bounded by hashtags.
     """
+    def __init__(self,  value: str):
+        self._value = value
+
 
-    def __init__(self, value: str):
+class SymbolChunk:
+    """
+    It creates an object with a [str] value for every single symbol.
+    symbol is defined as any string in between two consecutive hashtags, e.g. #it_is_a_symbol#.
+    """
+    def __init__(self,  value: str):
         self._value = value
 
+
+class LiteralAlternative(Alternative):
+    """
+        An alternative from a literal string and represents it as a chunk of literal and symbol objects.
+    """
+    def __init__(self, node: str):
+        self._node = node
+        # self._val_chunk contains the objects which make the string.
+        # It is equivalent to self._value in LiteralAlternative.
+        self._val_chunk = self._symbol_finder(self._node)
+
+    def _symbol_finder(self, node):
+        self.chunks = []
+        while node:
+            is_has_tag = [i for i, ltr in enumerate(node) if ltr == '#']
+            if is_has_tag:
+                if node[:is_has_tag[0]]:
+                    self.chunks.append(LiteralChunk(node[:is_has_tag[0]]))
+                    self.chunks.append(SymbolChunk(node[is_has_tag[0]:is_has_tag[1] + 1]))
+                else:
+                    self.chunks.append(SymbolChunk(node[is_has_tag[0]:is_has_tag[1] + 1]))
+
+                node = node[is_has_tag[1] + 1:]
+            else:
+                if node:
+                    self.chunks.append(LiteralChunk(node))
+                break
+        return self.chunks
+
     def split_form(self, include_adj=True) -> Tuple[Optional[str], str]:
-        return None, self._value
+        return None, self._node
 
 
-class AdjectiveNounAlternative(Alternative):
+class AdjectiveNounAlternative(LiteralAlternative):
     """
     An alternative that specifies an adjective and a noun.
     """
 
-    def __init__(self, adjective: str, noun: str):
-        self._adjective = adjective
-        self._noun = noun
+    def __init__(self, adj_node: str, n_node: str):
+        self._adj_node = adj_node
+        self._n_node = n_node
+        # self._adj_chunk contains the objects which make the adjective string.
+        # self._noun_chunk contains the objects which make the noun string.
+        # These are equivalent to self._adjective and self._noun in AdjectiveNounAlternative.
+        self._adj_chunk = self._symbol_finder(self._adj_node)
+        self._noun_chunk = self._symbol_finder(self._n_node)
 
     def split_form(self, include_adj=True) -> Tuple[Optional[str], str]:
         if include_adj:
-            return self._adjective, self._noun
+            return self._adj_node, self._n_node
         else:
-            return None, self._noun
+            return None, self._n_node
 
 
 class MatchAlternative(Alternative):
@@ -70,6 +113,7 @@ def __init__(self, symbol: str, alternatives: Iterable[Alternative]):
 
 
 class _Converter(NodeWalker):
+
     def walk_list(self, node):
         return [self.walk(child) for child in node]