Manages Retrieval-Augmented Generation (RAG) indices to enable semantic search capabilities over BigQuery datasets.
Simplifies the management of RAG indices within the BTDP infrastructure by providing a suite of tools to list, create, search, and delete indices. It bridges the gap between natural language queries and BigQuery data by utilizing semantic embeddings, allowing for efficient fuzzy matching and fast lookups without complex SQL. Designed for data environments with C1 and C2 classification, it automates the setup of underlying Cloud Workflows and Schedulers to ensure indices remain synchronized with source tables while providing high-performance retrieval for conversational queries.
Key Features
01Automated index creation for C1 and C2 BigQuery tables
02Automated orchestration of Cloud Workflows and Schedulers for embedding refreshes
03Semantic search execution with ranked results and fuzzy matching
04Shell script wrappers for reliable API interaction and error handling
05Comprehensive index lifecycle management including listing and deletion
060 GitHub stars
Use Cases
01Finding specific cloud identity groups or project IDs using conversational keywords
02Setting up automated embedding pipelines for semantic similarity tasks
03Performing natural language searches on large BigQuery datasets without writing SQL
