This implementation is meant to demonstrate how to integrate a document manager with retrieval in your existing infrastructure. In a nutshell, it serves as a document DB leveraging fulltext indexing.
To integrate into your existing production infrastructure, you have 2 options:
- Faster to implement, DB independent : Implement the interface provided by
Retrieverin document_manager_core.py - Slower to implement, optimized for your storage : Implement the interface provided by annotation and document stores in annotation_store.py and document_store.py