Vectorize Markdown files (or snippets) in the current workspace in some way and make them accessible from Markdown Copilot.
Relying on a nice vector database to store index files under a directory (e.g. .markdown_copilot) would be advisable.
There are two main uses, with the first being the priority:
Retrieve: Implement a feature to vectorize and index Markdown files within the workspace. When a user requests related content, the feature searches the index and inserts relevant text snippets into the current buffer.
RAG: Allow the user to ask questions and receive answers generated based on the indexed content of the workspace.
Proposal
Vectorize Markdown files (or snippets) in the current workspace in some way and make them accessible from Markdown Copilot. Relying on a nice vector database to store index files under a directory (e.g.
.markdown_copilot
) would be advisable.There are two main uses, with the first being the priority:
RAG can be split into another issue.