File "/Users/vinodvarma/Documents/GitHub/talktodata-api/ai.py", line 138, in parse_sql_results
program = GuidancePydanticProgram(
File "/Library/Frameworks/Python.framework/Versions/3.10/lib/python3.10/site-packages/llama_index/program/guidance_program.py", line 42, in __init__
output_str = pydantic_to_guidance_output_template_markdown(output_cls)
File "/Library/Frameworks/Python.framework/Versions/3.10/lib/python3.10/site-packages/llama_index/prompts/guidance_utils.py", line 54, in pydantic_to_guidance_output_template_markdown
output = json_schema_to_guidance_output_template(cls.schema(), root=cls.schema())
File "/Library/Frameworks/Python.framework/Versions/3.10/lib/python3.10/site-packages/llama_index/prompts/guidance_utils.py", line 89, in json_schema_to_guidance_output_template
+ json_schema_to_guidance_output_template(v, k, indent + 1, root)
File "/Library/Frameworks/Python.framework/Versions/3.10/lib/python3.10/site-packages/llama_index/prompts/guidance_utils.py", line 107, in json_schema_to_guidance_output_template
+ json_schema_to_guidance_output_template(schema["items"], "this", 0, root)
File "/Library/Frameworks/Python.framework/Versions/3.10/lib/python3.10/site-packages/llama_index/prompts/guidance_utils.py", line 79, in json_schema_to_guidance_output_template
root["definitions"][model], key, indent, root
KeyError: 'definitions' -----------Any clue why this this is happening?How to use the latest 'ObjectIndex' for working with Vector Store(PGVector) instead of default 'VectorStoreIndex'. to retrive data from already created index in my PGVector. self.vector_store = PGVectorStore.from_params(
database=DATABASE_NAME,
host=DATABASE_HOST,
password=DATABASE_PASSWORD,
port=DATABASE_PORT,
user=DATABASE_USER,
table_name=self.collection_table_name)
self.obj_index = ObjectIndex.from_objects(self.table_schema_objs, self.table_node_mapping, VectorStoreIndex)
self.db_query_engine = SQLTableRetrieverQueryEngine(
self.sql_database,
self.obj_index.as_retriever(similarity_top_k=3)) So far i have my code like this, But i'm not able to use my 'vector_store' inside the ObjectIndex.from_objects(). How to approach this. My usecase is I'm creating embeddings on entire databse(100+ tables) and storing them in a pgvector table. Now i just want to use that table to retrive the top 3 results. Any help here would be of a great help. I'm just following this example here https://gpt-index.readthedocs.io/en/latest/examples/index_structs/struct_indices/SQLIndexDemo-ManyTables.html