Problem with near_text Query and Metadata Filtering in Weaviate

Hello @DudaNogueira
thank you for your response!

The insertion worked, and the search is also functioning correctly. What I don’t understand is why, when using:

batch_size = 100  # Batch size
for i in range(0, len(data_rows), batch_size):
    batch_data = data_rows[i : i + batch_size]
    try:
        rag.data.insert_many(objects=batch_data)
        print(f"Batch {i // batch_size + 1} successfully inserted!")
    except Exception as e:
        print(f"Error inserting batch {i // batch_size + 1}: {e}")

the chunks returned by the search with query.near_text are different from those generated when using:

db = WeaviateVectorStore.from_documents(documents, embedding=embeddings, client=client, index_name="RAG")

It seems that the embedding model is not being loaded correctly. The “distance” values are also inconsistent.