fix(docs): modernize 00-fine-tuning.ipynb and chains.py to Pinecone SDK v8

[jhamon]

Summary

Modernizes learn/generation/openai/fine-tuning/gpt-3.5-agent-training/00-fine-tuning.ipynb and chains.py to use Pinecone SDK v8 (Linear SDK-174).

Changes

• Notebook: Replace pinecone-client==2.2.2 with pinecone>=7.0.0; add intro markdown, prerequisites, Colab/nbviewer badges; use getpass fallback for Pinecone API key per review template; clear outputs for valid nbformat.
• chains.py: Use Pinecone(api_key=...), ServerlessSpec(cloud="aws", region="us-east-1"), create_index with dimension=1536 and spec; align build_index metadata key with query (text).

Audience / use case

Developers fine-tuning GPT-3.5 with a Pinecone-backed retrieval tool.

Prerequisites

OpenAI API key, Pinecone API key; datasets, langchain, pinecone, openai (see notebook).

Related

• Linear: SDK-174

---

Note

Medium Risk
Moderate risk because it changes the Pinecone client initialization and index-creation parameters (serverless spec/region and embedding dimension), which could break the example if misconfigured or if users’ environments differ.

Overview
Modernizes the 00-fine-tuning.ipynb example by adding an intro/prereqs section and Colab/nbviewer badges, updating installs to pinecone==8.0.0 (plus requests), and switching Pinecone API key handling to an env-var-or-getpass flow; it also clears cell outputs/exec counts for a cleaner notebook.

Migrates chains.py to Pinecone SDK v8 by replacing legacy pinecone.init/Index usage with Pinecone(api_key=...) + ServerlessSpec index creation (explicit dimension=1536), and aligns stored metadata to use the text key so query() returns the expected field.

^{Written by Cursor Bugbot for commit 26fbf7a. This will update automatically on new commits. Configure here.}

[review-notebook-app]

Check out this pull request on 

See visual diffs & provide feedback on Jupyter Notebooks.

---

Powered by ReviewNB

[jhamon]

Addressed both Bugbot findings:

1. chains.py – environment is now passed to ServerlessSpec(region=environment) so the region is no longer hardcoded.
2. 00-fine-tuning.ipynb – Pip install runs in the first code cell; download + imports are in the second cell so Colab runs correctly.

Pushed in fa45214.

[jhamon]

Addressed Bugbot feedback:

1. Pip install before imports (High): Moved !pip install ... to the top of the first code cell so it runs before any imports. This fixes ModuleNotFoundError on a fresh Colab run.

2. Unused environment parameter (Medium): This was already fixed in a previous commit — chains.py now passes environment to ServerlessSpec(cloud="aws", region=environment), so the notebook’s PINECONE_ENV (or default us-east-1) is respected.

[cursor]

Cursor Bugbot has reviewed your changes and found 1 potential issue.

^{Bugbot Autofix is OFF. To automatically fix reported issues with Cloud Agents, enable Autofix in the Cursor dashboard.}

[cursor]

Duplicate placeholder cells accidentally committed

Low Severity

The notebook contains two identical code cells that only contain the comment # Setup complete (imports and chains.py in previous cell). These duplicate placeholder cells appear to be leftover from editing and serve no purpose. According to the review rules, notebooks should flow logically with clear structure, and these redundant cells clutter the notebook without adding value.