{"repo":"graphgeeks-lab/odsc-agentic-ai-summit-2025","free":true,"listed":false,"github":"https://github.com/graphgeeks-lab/odsc-agentic-ai-summit-2025","clone":"git clone https://github.com/graphgeeks-lab/odsc-agentic-ai-summit-2025.git","description":"Code from the ODSC Agentic Graph RAG workshop combining vector, FTS & graph retrieval for RAG. Includes observability and guardrails for evaluating outputs.","language":"Python","stars":19,"topics":["agent","baml","fts","full-text-search","graph-database","graph-rag","graphrag","hybrid-rag","hybrid-search","kuzu"],"license":null,"category":"analytics","readme_excerpt":"Agentic Workflows for Graph RAG ================ ODSC Agentic AI Summit 2025 Workshop By GraphGeeks ---- :spiral calendar: July 16 - 31, 2025 :world map: Virtual :writing hand: AI Summit Track ---- All you really need to know - Basics of Python - Basic familiarity with vector search - How to use the terminal and modern IDEs (like VS Code or Cursor) Tips to get started After you've attended the sessions, we recommend cloning this repo and following our tips to get started. Instructors This 3-part workshop is led by a team of 4 instructors. Amy Hodler Dennis Irorere Prashanth Rao David Hughes Dataset This workshop uses a dataset of 2,726 FHIR records of patients and their notes. The dataset is obtained from this Hugging Face dataset and preprocessed using the script create dataset.py . To create the dataset locally, run the following command: This creates the following two JSON files: - Raw data: 2,726 notes in unstructured text format, output to data/note.json - Evaluation data: 2,726 FHIR JSON records, output to data/fhir.json Setup Python environment It's recommended to install uv to manage the dependencies. Alternatively, you can install the dependencies manually via pip. Components 1. Information extraction Our first goal is to extract entities and relationships that can form a knowledge graph from the raw data (patient notes) in the data/note.json file. The information extraction pipeline is powered by BAML, a programming language for obtaining high-quality structured out","default_branch":null,"files":null,"tree":[],"storefront":"/r/graphgeeks-lab","claimed":false,"request_supported":{"post":"https://gitbuyer.com/r/graphgeeks-lab/odsc-agentic-ai-summit-2025/request-supported","requests":0},"note":"indexed from public GitHub; nothing is for sale on this page. Clone it from GitHub. Paid listings live at /search."}