If the prompt I want to give is an instruction about tone/delivery style, does it still make sense to pass it in via index.query?

How would we go about defining a custom method for creating nodes from a given document?

Yeah, that's what I was planning on doing—but how significant a performance/monetary cost would that incur?

How would we go about defining a custom method for creating nodes from a given document? Is there some way of defining a custom/alternate TextSplitter (https://github.com/jerryjliu/gpt_index/blob/main/gpt_index/langchain_helpers/text_splitter.py)?

Upon attempting to run the notebook provided for the Qdrant vector store, and actually query the index, I came across this error:

AttributeError Traceback (most recent call last)
Cell In[6], line 1
----> 1 response = index.query('What does the author think about lawyers?')

File ~/Code/benipers/env/lib/python3.10/site-packages/gpt_index/indices/base.py:399, in BaseGPTIndex.query(self, query_str, verbose, mode, query_kwargs) 384 query_config = QueryConfig( 385 index_struct_type=self._index_struct.get_type(), 386 query_mode=mode_enum, 387 query_kwargs=query_kwargs, 388 ) 389 query_runner = QueryRunner( 390 self._llm_predictor, 391 self._prompt_helper, (...) 397 recursive=False, 398 )--> 399 return query_runner.query(query_str, self._index_struct)File ~/Code/benipers/env/lib/python3.10/site-packages/gpt_index/indices/query/query_runner.py:112, in QueryRunner.query(self, query_str, index_struct) 104 query_kwargs = self._get_query_kwargs(config) 105 query_obj = query_cls( 106 index_struct, 107 query_kwargs,
108 query_runner=query_runner,
...
89 )
90 nodes.append(node)
92 if similarity_tracker is not None:

AttributeError: 'ScoredPoint' object has no attribute 'get'