Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Allow Customization of File Recommendation Scoring in Copilot #1382

Open
fgielow opened this issue Mar 21, 2025 · 1 comment
Open

Allow Customization of File Recommendation Scoring in Copilot #1382

fgielow opened this issue Mar 21, 2025 · 1 comment

Comments

@fgielow
Copy link

fgielow commented Mar 21, 2025

Is your feature request related to a problem? Please describe.

It's not something breaking, but definetely something that could be improved, as better detailed below.

Currently, Copilot Plus relies on a default relevance scoring method that doesn’t capture the nuances of file-based tagging practices, leading to suboptimal recommendations for users with unique organizational structures. Allowing customization—potentially even at the functional level with advanced methods like structured graph analysis—would greatly enhance the tool's flexibility and effectiveness in managing diverse vault content.

The Problem: Currently, my experience with Copilot Plus is that the default file recommendation scoring system doesn’t fully capture my personalized file-based tagging approach. I use [[file-links]] instead of #tags; then, when I run queries like "List [[fiction]] books by [[Tolkien]] in my @vault", the results often lack the accuracy and relevance I expect—returning files that aren’t even linked by these file-links, while overlooking many that are. Note: the underlying reason, I believe, could be that individual files link the [[tag]] file, but that tag file itself does not link back to them. So these back-links are extremely important for me and I'd like to score them higher if I could.

Describe the solution you'd like

To address this issue, I propose that Copilot Plus provide a customizable file recommendation scoring system. This feature would let users adjust relevance weights based on their unique vault configurations, ensuring that the algorithm accurately prioritizes file-based tags. This flexibility would significantly improve search accuracy and usability for those using tailored tagging methods.

Moreover, interpreting the relationships between files and tags as a graph could unlock significant potential in custom scoring systems. This approach would enable the use of advanced graph analytics to better understand and optimize individual vault structures enabling users to implement their own functions according to some provided meta-data or even graph representation.

Describe alternatives you've considered

No viable alternatives exist at this time. I explored the possibility of adjusting folder structures to try improve scoring against the current function, but that approach conflicts with my established and well-organized system. The solution should enable me to maximize the effectiveness of my current structure without requiring major structural changes.

@logancyang
Copy link
Owner

Thanks for the details. Graph-based relevance is one of the major directions for improvement for sure. Will come up with a generalizable design for this.

cc @zeroliu @wenzhengjiang

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
Status: Backlog
Development

No branches or pull requests

2 participants