The spec

Depends on:

• RAG - Add models #1645
• RAG - Add Embedding Provider and Services #1647

Requirement

The user needs to be able to choose which index backend to use when creating a new Collection.

• OpenAI
• OCS

When OCS is selected, the user need to be able to select which embedding provider and model to use.

When OCS is selected and the user upload new files:

• Chunk the files according to the configured strategy
• Create embeddings by sending each chunk to the selected provider. Save the embedding in the newly added FileChunkEmbedding model (which are added in this ticket)