Innovative Users Conference - IUG 2024
IUG 2024 - Python Power-Up 🐍🍄⚡
Supercharging Sierra Integration with sierra-ils-utils
Links
Description:
Integration of information systems, particularly the Sierra Integrated Library System (ILS), presents a set of unique, ongoing challenges. This presentation introduces “sierra-ils-utils,” an unofficial yet innovative Python library to help simplify and enhance this integration process.
The “sierra-ils-utils” library’s primary goal is to improve flexibility, readability, and efficiency in terms of interacting with Sierra’s REST APIs. Key features of “sierra-ils-utils” include: automated authentication and token refresh mechanisms, robust request retries upon failures, and efficient mapping of JSON responses from Sierra REST API endpoints to predefined data models. Additionally, the library’s availability on PyPi and easy installation via PIP significantly simplifies “sierra-ils-utils” integration into various projects. Suggestions and pull requests on the public repository are welcome! https://github.com/chimpy-me/sierra-ils-utils
Semantic Search Swiftly 🗣️🔎🚀: Thirty-Six minutes to Transform Sierra Data for Enhanced Data Discovery
Links
-
Introduction
- Extract
- Transform
- Load / Search
Description:
There’s no avoiding AI, so you may as well understand it a bit better! This presentation will cover basic concepts of Natural Language Processing (NLP) – A subset of artificial intelligence, and machine learning – and will briefly explore how semantic search differs from traditional keyword searching. The presentation will also cover a rapid process that extracts, transforms and loads (ETL) data for the purpose of creating a semantic search platform.
In this session, we’ll utilize the the ‘sierra-ils-utils’ Python library to extract ILS data, the Google Colab platform and a transformer model from the Huggingface platform to transform the data, and Qdrant – an easy-to-use, free, and open source vector database – to load the resulting vector embeddings. This combination of a vector database and data modeling enables a powerful semantic search capability. All this within just 36 minutes (hopefully)!