Interaction Base

[ntindle]

Background

Right now, the interaction models for autogpt are tightly coupled to our autogpt.logs logger and input. Decoupling them is required for supporting other UX models and frontends.

Changes

Adds an abstraction (unused) for the console to interact with that can be hooked for other interactables. Open Questions:

• [ ] Can you register multiple frontends in the same instance of autogpt?
• [ ] What more do the frontend messages do we need to support?
• [ ] what type of messages should we support from a priority perspective?
• [ ] whats the first pretty frontend we want?

Documentation

None Yet

Test Plan

None yet

PR Quality Checklist

• [ ] My pull request is atomic and focuses on a single change.
• [ ] I have thoroughly tested my changes with multiple different prompts.
• [ ] I have considered potential risks and mitigations for my changes.
• [ ] I have documented my changes clearly and comprehensively.
• [ ] I have not snuck in any "extra" small tweaks changes

[vercel[bot]]

The latest updates on your projects. Learn more about Vercel for Git ↗︎

1 Ignored Deployment

Name	Status	Preview	Comments	Updated (UTC)
docs	⬜️ Ignored (Inspect)			Apr 30, 2023 7:44am

[codecov[bot]]

Codecov Report

Patch coverage has no change and project coverage change: -0.75 :warning:

Comparison is base (06ae468) 58.78% compared to head (a3db552) 58.04%.

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #3569      +/-   ##
==========================================
- Coverage   58.78%   58.04%   -0.75%     
==========================================
  Files          68       71       +3     
  Lines        3062     3120      +58     
  Branches      501      511      +10     
==========================================
+ Hits         1800     1811      +11     
- Misses       1131     1178      +47     
  Partials      131      131              

Impacted Files	Coverage Δ
autogpt/interaction/base.py	0.00% <0.00%> (ø)
autogpt/interaction/console.py	0.00% <0.00%> (ø)
autogpt/interaction/interaction_manager.py	0.00% <0.00%> (ø)

... and 1 file with indirect coverage changes

:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Do you have feedback about the report comment? Let us know in this issue.

[p-i-]

This is a mass message from the AutoGPT core team. Our apologies for the ongoing delay in processing PRs. This is because we are re-architecting the AutoGPT core!

For more details (and for infor on joining our Discord), please refer to: https://github.com/Significant-Gravitas/Auto-GPT/wiki/Architecting

[Boostrix]

Right now, the interaction models for autogpt are tightly coupled to our autogpt.logs logger and input. Decoupling them is required for supporting other UX models and frontends.

The potential use-case for his was today discussed as part of #3820 where the conclusion was to support some kind of wizard mode (probably plugin based, using JSON under the hood for serialization). Depending on the implementation, such a wizard could also be controlled using other front-ends (think AJAX or REST).

The intial purpose would be helping set up the config for Auto-GPT itself, but the wizard would be a part of the command system - so that agents could optionally generate wizards "on the fly".

Internally, the wizard would work interactively via stdin/input like you say, but could be mapped to other back-ends, as needed. That would make it possible to map the whole thing onto a REST API - or even let agents use the same mechanism for inter-agent messaging

[anonhostpi]

I agree. Decoupling is always a good practice. Good IoC