Is there an option to recreate the MultiModalVectorStoreIndex once the collections are in a Qdrant Cluster? from_vector_store does not work as Qdrant only takes 1 collection. Which approach is doable The only solution I see is creating a local client, but I want to do it from a cluster

Initialize Qdrant client

client = QdrantClient(
url=QDRANT_URL,
api_key=QDRANT_API_KEY
)


text_store = QdrantVectorStore(
client=client, collection_name="text_collection"
)
image_store = QdrantVectorStore(
client=client, collection_name="image_collection"
)
storage_context = StorageContext.from_defaults(
vector_store=text_store, image_store=image_store
)

Create the MultiModal index

documents = SimpleDirectoryReader("/content/output_folder").load_data()

index = MultiModalVectorStoreIndex.from_documents(
documents,
storage_context=storage_context,
embed_model=embed_model
)

Question: Agents and routing

Which options would you choose from LlamaIndex to generate synthetic Exam question? I want the model first to check sample questions and then go to the knowledge base to generate similar ones.

I am not yet familiar with the new workflows, so would this work better or would it be a better option to use RouterQueryEngine with Agents?

@kapa.ai Could anyone point out which llamaindex library can extract text from a pdf in the fastest possible way? Lets say the pdf has also images but i just need the text in the fastest possible way. Any hint or any better alternatives?