Use Semantic Similarity to group requests

[gbrls]

I've developed a library to group binary data based on their similarity (https://github.com/gbrls/zip-dist). It has shown very interesting results and I wish to implement it. It'll show up as an extra column with the ID of the group that has been selected.

There are a few approaches that I'm thinking about implementing it. We could group responses based on:

• Just the headers.
• Just the body.
• The whole response with the headers.

Using just the headers in my experience, grouped the responses almost by tech stack, such as, Shopify shops, Nginx servers with Azion on the front, Microsoft IIS servers, requests from those groups were grouped in their own groups.

I haven't tried doing so using the whole body, so I'll try experimenting with it before releasing Kurl with it as a default.