Add Language Model Decoder

Closes #142 (closed)

Taking over from !222 (closed)

dan/datasets/extract/extract.py

@@ -242,7 +242,7 @@ class ArkindexExtractor:
         """
         return " ".join(
             [
-                self.mapping.encode[token] if token in self.mapping else token
+                self.mapping.encode[token] if token in self.mapping.encode else token
                 for token in list(text.strip())
             ]
         )