Repeating output

[qwertystars]

Describe the bug

When rendering Markdown content, particularly within code snippets or text areas, the text is being repeated multiple times, despite it being intended to display only once. This repetition occurs within the rendered output, making it difficult to read and understand the Markdown content properly. This maybe a issue of powershell. if Yes please create the gui fast.

Reproduce

Write this as prompt or any prompt that will make it generate output longer than the display size of monitor use. Make me a simple Pomodoro app.

Expected behavior

The Markdown content should be displayed exactly once, providing a clear and concise representation of the intended text.

Screenshots

Open Interpreter version

0.2.4

Python version

3.10.7

Operating System name and version

Windows 11

Additional context

No response

[rustom]

Reproduced on MacOS, I don't think it's a Powershell issue. +1

[Notnaton]

This is a known issue.

We rerender the output of the llm each time we get a token. This causes it to print a new "screen".

One possible fix would be to rerender everything once the llm is done. I will explore this.

Moving to textual would solve this, I have explored doing that. Just need some motivation 😅

[bars0um]

Not sure if this is an ok fix, but it appears that if I switch the Live attribute vertical_overflow to "elipses" on line 12 of terminal_interface/components/base_block.py the issue somewhat subsides...I've not tested this fully yet and it may end up causing a different issue if your file output is bigger than your console height..."crop" may help avoid issues with overrunning the console buffer.

With either of those settings I'm able to comfortably debug and review output when I put the --debug flag and things don't jitter and repeat.

[davidkan6666]

Have the similar issue, the output will be repeated many times:

[dandarm]

I'm experiencing the same problem still today, using gpt-4o-mini. Any solution up to now?