Neo4j GenAI Workshop

Please see genai-workshop.ipynb which serves as the self-contained workshop.

The other companion notebooks contain code for staging data, building the Neo4j Graph, and providing easy access to demos:

1. data-prep.ipynb stages the workshop data, sampling and formatting data sourced from the H&M Personalized Fashion Recommendations Dataset.
2. data-load.ipynb loads the staged data into Neo4j, performs text embedding, and creates a vector index.
3. genai-workshop-w-outputs.ipynb is a reference notebook for following along and checking outputs. It is simply genai-workshop.ipynbwith cell outputs intact.
4. genai-example-app-only.ipynb is a copy of genai-workshop.ipynb that contains only the final section: the demo LLM GraphRAG app for content generation. It assumes you have already run genai-workshop.ipynb and exists only for instructor demo purposes.

Changelog

v4 (Sep 2nd, 2024 - Present)

---

• Transition from using Neo4j Sandbox to AuraDS

• Split out data loading

  • Split out data load into a separate notebook
  • Live workshops now begin with the dataset pre-loaded to cut down on time and spend more of the course walking through GraphRAG. the data-load.ipynb is kept for reference and replication.
  • Remove neo4j_tools Python package. the functions/utilities are now included in data-load.ipynb

• Updated Workshop Slides

• Added more query exploration & improved explainer queries

  • Add Browser-based graph exploration in beginning of workshop
  • Include database tips & more Cypher queries in multiple steps
  • Update explainer markdown and code cells for graph patterns and GDS for clarity
  • Various other minor adjustments to markdown and code to improve course quality

• Added genai-workshop-w-outputs.ipynb and GitHub Actions Workflow

  • genai-workshp.ipynb is now maintained with cleared outputs for better workshop experience and easier PR review
  • A GitHub actions workflow automatically tests data loading, workshop, and example-only notebooks and auto-commits genai-workshop-w-outputs.ipynb file for each PR.

v3 (June 25th, 2024 - Sep 1st, 2024)

---

• improve LLM response quality and cleaned up code for LLM chains and vector stores

  • parameterizing customer id so don't need to recreate chains & stores for each customer
  • updated prompts to better account for seasonality and use all retrieved data
  • update to use gpt-4o

• Improve text embedding speed and reduce code by transitioning to native genai.vector Cypher functions

• Updated slides

• Various other minor adjustments to markdown and code to improve course quality

v2 (Feb 20th, 2024 - June 24th, 2024)

---

• (fix) Add langchain_community to the libraries that are pip installed in the notebooks

• Simplify and Shorten Course

  • Shortened GDS section to just three cells to run
  • Condensed Vector Search Section
  • Condensed Loading to Single Notebook Cell
  • Switched Recommendation Retriever to a Simple KG Query
  • Adding neo4j_tools Package to hold convenience functions for loading data and reduce code footprint in main workshop notebook
  • Updated to GPT-4 throughout
  • General Notebook Cleaning - Removed duplicate load statements, updating to newest llm packages, etc.

• Provide Better Explainers & Examples

  • Add A Chain for Printing Final Prompt to LLM with retrieval data to better explain process.
  • Added Differentiated Names for Customer Examples in Demo App.

• Added Additional Resources

  • Added workshop slides
  • Added "demo only" notebook

v1 (Nov 13th, 2023 - Feb 19th, 2024)

---

• Initial 5-part course with
  • Building the knowledge graph
  • Vector search & text embedding
  • Graph patterns to improve semantic search
  • knowledge graph inference & ML
  • Building the LLM chain and demo app for generating content

Contributing

Contributions are welcome! To contribute please:

1. Make a PR with a descriptive name
2. If you are updating genai-workshop.ipynb please ensure to clear all outputs before committing.
3. Do not alter the genai-workshop-w-outputs.ipynb file. This file is autogenerated upon creating/updating PRs.