Fix: Issue of failure to initialize input_converter in Seq2SeqGenerator when model_file_path is given as folder path on local disk after manual model download (#4213)

* test

* test documentation commit:

* added original return statement for linting

* removed empty lines

* formatted code using black

* made changes based on suggestions

Author: Kshitijpawar

haystack/nodes/answer_generator/transformers.py

@@ -413,6 +413,11 @@ def _register_converters(cls, model_name_or_path: str, custom_converter: Optiona
 
     @classmethod
     def _get_converter(cls, model_name_or_path: str) -> Optional[Callable]:
+        # using dictionary key based on model_name_or_path endswith
+        if model_name_or_path.endswith("bart_lfqa"):
+            model_name_or_path = "vblagoje/bart_lfqa"
+        elif model_name_or_path.endswith("bart_eli5"):
+            model_name_or_path = "yjernite/bart_eli5"
         return cls._model_input_converters.get(model_name_or_path)
 
     def predict(self, query: str, documents: List[Document], top_k: Optional[int] = None) -> Dict: