search

smallcloudai / refact

WebUI for Fine-Tuning and Self-hosting of Open-Source Large Language Models for Coding
https://refact.ai
BSD 3-Clause "New" or "Revised" License
1.56k stars 105 forks source link

Consider integrating with atom #96

Open prabhu opened 1 year ago

prabhu commented 1 year ago

https://github.com/AppThreat/atom

We created an open-sourced atom for the precise identification of usages and dataflows across large code bases. This approach is better for summarizing and identifying context than the primitive tokenization commonly found everywhere.

We have a dedicated community support channel should you require any help with using atom or code analysis in general.

olegklimov commented 1 year ago

Interesting!

This approach is better for summarizing and identifying context

Are you saying Atom might help to fill model context, to help it to come up with a better suggestions or edits? Or do you mean something more broad?

primitive tokenization commonly found everywhere

I think you are describing the process of looking for tokens in a sliding window, and matching that vs other files in the project.

I've found these supported languages in joern-lib:

Name Based on Maturity
C/C++ Fuzzy Parser High
C/C++ Eclipse CDT High
x86/x64 Ghidra Medium
JVM Soot Medium
Javascript GraalVM Medium
Javasrc javaparser Medium
Kotlin IntelliJ PSI Medium
Python Custom Medium

Does it mean Atom can't work outside of these languages, such as Rust?