Update embeddings and Prompts

[zach-blumenfeld]

Update Workshop notebooks for the following:

1. Streamline text embedding by using genai.vector.* Cypher operations instead of external workflow Use Native Cypher Encode to generate and store embeddings #29
2. Simplify prompt used in LLM chain (No need for multi-part chat template)
3. Updated from "gpt-4" to "gpt-4o" to return results much faster
4. Updated prompt for better experience:
   • Updated prompt template langauge and inputs so model better accounts for customer interests and seasonality/timeOfYear when making fashion recommendations
   • more "summer friendly" search prompt examples Update Search Prompt Examples for Summer #31
5. Updated markdown text in introduction with "GraphRAG" and aligned to abstract.....cleaned markdown in a few other places

Things not done due to time constraints (open to feedback if we still need them)

1. Remove dependency on GDS client in the beginning along with the custom neo4j_tools package I made Remove CUstom Python Package Dependency - GDS Tools #32
2. Some other weird GDS things: Switch out GDS Debug Call for Testing Connection #33 GDS deprecated field warning #35
3. Use Vector Index for Graph Embeddings Rather then KNN Relationship Drawing #30 - for retrieval...this would be nice, but also would require changing a lot of language in the notebook and slides, and arguably may be worse, especially on sandbox instances with limited memory.