Embeddings and Vector Databases With ChromaDB
Now that you understand the mechanisms behind ChromaDB, you’re ready to tackle a real-world scenario. Say you have a library of thousands of documents, and you need a way to search through them.
In particular, you want to be able to make queries that point you to relevant documents. For example, if your query is find me documents containing financial information, then you want whatever system you use to point you to a financial document in your library.
How would you design this system? With your knowledge of vectors and embeddings, your first inclination might be to run all of the documents through an embedding algorithm and store the documents and embeddings together. You’d then convert a new query to an embedding