ai-agents-node-instantiation-when-using-llama-index-v0-10-with-a-vector-store-t-7852d0e4

THEN Ensure that the vector store's query method returns concrete node types like TextNode or Document, which implement all abstract methods of BaseNode. If using a custom vector store, subclass BaseNode appropriately and implement get_content, get_metadata_str, get_type, hash, set_content methods, or directly return TextNode as returned by the underlying vector database.

from llama_index.core.schema import TextNode

class MyVectorStore(BasePydanticVectorStore):
    def query(self, ...):
        # Return TextNode instances
        return [TextNode(text='...', id_='...')]