List view
MVP of the JetBrains & VSCode Plugins. This entails: - Code-completion works, and they are suggested at appropriate times. - Basic settings menu allowing the user to opt-in to data collection, and to select which languages they want code-completion on. - Completion feedback; e.g. whether the completion is still present in the code after 30s.
Overdue by 1 year(s)•Due by July 27, 2024•0/2 issues closedThe goal is to have an MVP of the server which is basically able to serve any user with meaningful completions. This entails many things, namely: 1. Having a work database implementation that is able to handle read, write, and update operations through the provided interface. 2. A plug-&-play system for having LLMs generate completions for a given request in the expected format (also through a provided interface -> generation as a service) 3. A server that can take a request, and provide a meaningful response to it in the expected format. This component can then also handle the lifecycles of existing sessions per user, handle the lifecycle of a request (i.e. updates to whether the user accepted the generated response from the LLM, how many times a generation was shown, etc.), and facilitate seamless interaction between the server and the user. ~4. The central logic of the server manages the overall interactions between the different services and the incoming requests from the user.
Overdue by 1 year(s)•Due by July 12, 2024•13/15 issues closed