Separate core and cli into separate modules (directories) and only allow cli to import from core

[AntonOsika]

The idea is to separate the core logic and CLI UX specific things. To make it easier to take decisions on what makes sense from UX perspective, and how the core building blocks should work.

Would look something like:

gpt_engineer
├── core
│   ├── ai.py
│   ├── domain.py
│   ├── chat_to_files.py
│   ├── steps.py
│   └── db.py
├── cli
│   ├── main.py
│   ├── file_selector.py
│   ├── learning.py
│   └── collect.py
├── api
│   └── main.py
└── preprompts
    └── ...

One could use either:

• PyCharm "move" automagic functionality
• Or! gpt-engineer by adding new steps and configs, or somehow the existing -i flag

[fabhed]

We should probably also add an __init__.py and lift the core functionality, like this:

from gpt_engineer.core import ai, domain, chat_to_files, steps, db

then it can be consumed without specifying core in imports:

from gpt_engineer import ai

[AntonOsika]

Sounds good. It would be great to enforce that cli can only import from core, not vice versa, and that "internal" functions cannot be imported.

…

On Wed, Sep 20, 2023 at 3:31 PM Fabian Hedin ***@***.***> wrote: We should probably also add an __init__.py and lift the core functionality, like this: from gpt_engineer.core import ai, domain, chat_to_files, steps, db then it can be consumed without specifying core in imports: from gpt_engineer import ai — Reply to this email directly, view it on GitHub <#718 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ABCCSUKHYZTWEXMEGLVC65LX3LV43ANCNFSM6AAAAAA44T4PUI> . You are receiving this because you authored the thread.Message ID: ***@***.***>

-- Anton Osika

[UmerHA]

Done in #766