fix: support piping stdin into gptme -r (resume)

[bjsi]

Almost got it to do this autonomously, but I still had to do some cleanup :P

I added the msg.role == "user" checks because those functions contain asserts that the role should be user.

Fixes #202

---

Important

Add support for piping stdin into gptme -r (resume) by appending piped input to prompt messages in cli.py.

• Behavior:
  • Support piping stdin into gptme -r (resume) in cli.py.
  • Append piped input to prompt_msgs if resuming a conversation in main().
• Functions:
  • Modify main() in cli.py to handle piped input and append to initial_msgs for new conversations.
  • Update confirm_func() and execute_cmd() logic in chat.py to handle user role checks.
• Misc:
  • Refactor stdin reading logic in cli.py to use _read_stdin() function.

^{This description was created by for a69c42e. It will automatically update as commits are pushed.}

[ellipsis-dev]

👍 Looks good to me! Reviewed everything up to a69c42e in 11 seconds

More details

• Looked at 58 lines of code in 2 files
• Skipped 0 files when reviewing.
• Skipped posting 1 drafted comments based on config settings.

1. gptme/cli.py:240

• Draft comment:
  Consider refactoring the code to avoid redundancy when appending piped input as a system message. This pattern is repeated in multiple places.
• Reason this comment was not posted:
  Confidence changes required: 50%
  The code in cli.py and chat.py has a repeated pattern for appending piped input as a system message. This can be refactored to avoid redundancy.

Workflow ID: wflow_WjHUDUfAniVn5nUR

---

You can customize Ellipsis with 👍 / 👎 feedback, review rules, user-specific overrides, quiet mode, and more.

[ErikBjare]

Very nice! Thank you so much @bjsi ❤️

[ErikBjare]

Improved the behavior in #354