Feluda is a multimodal multilingual search engine framework developed by Tattle.
Search is hard. There is no silver bullet that serves all search needs. Lets say you are building a search engine for images. If all you want is the ability to search for exact duplicates, hashing the images and storing them in a sqlite db might be the most optimum choice for you. Whereas, if you care to search for images for visual similarity along with semantic similarity, you might want to consider experimenting with generating the image's vector representation and storing it in database that allows vector search like PinoSearch or ElasticSearch.
Depending on your data, storage and memory limitations, different approaches might make sense.
Feluda was born out of our need to try out and compare various search approaches to keep up with the ever changing requirements of making Social Media Data on Indian Internet searchable.
Feluda is still in a pre-release stage. Expect APIs and core functionality to change.
To begin experimenting with Feluda, start by taking a tour of the Architecture and Operators