Update the LlamaIndex notebook to current SDKs

[nicoalba10]

Ports the tested code from the LlamaIndex integration guide so the notebook matches the page step by step. It was pinned to pinecone-client[grpc]==3.0.0 and llama-index==0.9.34 and used imports that no longer exist.

• Pins pinecone==9.1.0 (the highest version llama-index-vector-stores-pinecone==0.9.0 allows, since it requires pinecone<10), llama-index==0.14.25, and the split llama-index-* packages.
• Uses text-embedding-3-small for embeddings and gpt-4o-mini for generation and evaluation.
• Adds the out-of-scope evaluation example from the page and a cleanup cell that deletes the index.
• Points the Colab and nbviewer badges at main.

Ran end to end on Python 3.12: 46 vectors upserted, the relevancy check passes for the in-scope query and fails for the out-of-scope one.

---

Note

Low Risk
Documentation-only notebook update with no production code or auth changes; risk is limited to example correctness and pinned dependency versions.

Overview
Brings using-llamaindex-with-pinecone.ipynb in line with the LlamaIndex integration guide so it runs on current Pinecone and LlamaIndex packages.

Dependency and API refresh: Replaces legacy pins (pinecone-client[grpc]==3.0.0, monolithic llama-index==0.9.34) with pinecone==9.1.0, llama-index==0.14.25, and split packages (llama-index-vector-stores-pinecone, llama-index-readers-file, etc.), and updates imports to the llama_index.core / integration module layout (PineconeGRPC, PineconeVectorStore, IngestionPipeline, and related APIs).

Models and flow: Uses text-embedding-3-small for splitting/embeddings and gpt-4o-mini for RAG answers and RelevancyEvaluator checks. Adds the guide’s out-of-scope query/eval example and a cleanup cell that deletes the Pinecone index. Colab and nbviewer badge URLs now target the main branch.

^{Reviewed by Cursor Bugbot for commit 2eb5102. Bugbot is set up for automated code reviews on this repo. Configure here.}

[cursor]

This PR is too large for Bugbot to review. It changes 1,516 lines and 1,926,351 characters. Split the change into smaller pull requests to get a review.