notebook-terminal 0.3.5

A cross-platform PTY/ConPTY terminal for Jupyter with bundled xterm.js

Notebook Terminal 0.3.4

A real PTY/ConPTY terminal embedded in Jupyter using bundled xterm.js and only standard ipywidgets channels. It needs no anywidget, JupyterLab extension, administrator installation, HTTP server, WebSocket, CDN, or internet connection.

Installation

python -m pip install notebook_terminal-0.3.4-py3-none-any.whl

Restart the kernel after replacing an older version.

Usage

from notebook_terminal import terminal
term = terminal(height=450, interactive=True)

The shell prompt, username, host, and working directory come from the actual PowerShell, CMD, Bash, or Zsh process. They are not hardcoded.

0.5.0 transport changes

• lossless output packets with sequence numbers and acknowledgements;
• every frontend state contains all unacknowledged packets, so coalesced widget updates cannot drop output;
• exact byte transport preserves split UTF-8 and ANSI sequences;
• output batching every 25 ms and maximum 64 KiB packets;
• ordered, acknowledged input queue prevents lost keystrokes;
• xterm writes are serialized and acknowledged only after rendering;
• resize events and terminal control messages remain deduplicated.

This release specifically addresses missing characters, overwritten output, invisible fast commands such as ls, and severe lag caused by sending every PTY fragment as an independent widget trait update.

Repository

GitHub: ZawadzkiR/notebook-terminal

Passing arguments to Python files

term.run_python_file(
    "script.py",
    args=["input.csv", "value with spaces", 123],
)

The script receives:

import sys
print(sys.argv)
# ['script.py', 'input.csv', 'value with spaces', '123']

Named options can be generated with kwargs:

term.run_python_file(
    "train.py",
    args=["dataset.csv"],
    kwargs={
        "epochs": 100,
        "learning_rate": 0.001,
        "verbose": True,
        "dry_run": False,
        "tag": ["baseline", "test"],
    },
    rich_output=True,
)

This produces arguments equivalent to:

python train.py dataset.csv --epochs 100 --learning-rate 0.001 --verbose --tag baseline --tag test

Rules:

• values in args are converted to strings and passed separately;
• underscores in kwargs keys become hyphens;
• keys already beginning with - are preserved;
• True creates a flag;
• False and None omit the option;
• list and tuple values repeat the option.