code_index.plain

---
import:
  - code_retrieval_reqs
requires:
  - embeddings
---

***implementation reqs***

- :Index: should be implemented in the data_index folder

- :Index: should be implemented with functions from the faiss library

***functional specs***

- Create the CodeIndex class that implements the :Index: 
  - It should contain an init function that initializes the :Index: of :Embedding: instances.
  - Use the faiss library to create a FAISS :Index: object.
  - initialize the metadata as an empty list

- Implement an add function to the CodeIndex class that takes an :Embedding: and :Chunk: metadata as input and adds them to the :Index: and metadata.

- Implement functionality in the CodeIndex class to load the FAISS :Index: and metadata taking file paths as input.

- Implement a save function that writes the FAISS :Index: and metadata to their respective files.

  ***acceptance tests***

  - The save function should result in a FAISS index file and a metadata file where both files should have the same number of entries and the metadata file should contain accurate metadata for every extracted Chunk.

- Implement a query_index function that takes a string as an input and returns the top K results from :Index:.
  - The embeddings module should be used to extract the :Embedding: instances from an input string to compare to the :Embedding: instances in :Index:.
  - The function should return the the top K results from :Index:.