[anotherpaulg]

Super interesting, thanks for sharing. How large a corpus of domain specific text do you need to obtain a useful knowledge graph?

Aider has been doing PageRank on the call graph of code repos since forever. All non trivial code has lots of graph structure to support PageRank. So it works really well to find the most relevant context in the project related to the currently active task.

https://aider.chat/docs/repomap.html#optimizing-the-map

[liukidar]

We have tried from small novels to full documentations of some milion tokens and both seem to create interesting graphs, it would be great to hear some feedback as more people start using it :)

[ukuina]

I enjoy Aider, but it has never successfully created a repo map, regardless of whether the codebase is Python, JS, or TS. Are there any plans to allow force-creation and inspection of a repo map?

[LiveTheDream]

In the chat, you can[0]:

- View the current repository map using `/map`

- Force a refresh of the repository map using `/map-refresh`

If you want to save the repository map to a file for inspection, you can use [1]

    aider --show-repo-map

[0] https://aider.chat/docs/usage/commands.html

[1] https://aider.chat/docs/config/options.html#--show-repo-map

[ukuina]

Thanks, I have tried this, but it simply shows no elements in the repo map with no accompanying errors. Inspecting the file also shows an empty repo map. I'm wondering if Aider needs additional logging when invoking tree-sitter.

[LiveTheDream]

I’m on a slightly modified version of 0.52.1 which is getting a bit dated but it works well for me even with not officially supported source, like svelte.

In case this thread helps someone else, some errors with —show-repo-map can be solved by setting environment variable PYTHONIOENCODING=utf-8

[LiveTheDream]

Is it possible that you the files you expect mapped at not tracked in a git repo?

That would explain the empty output.