Jupyter Qt console

Warnings

• gotcha When installing via `pip`, Qt bindings (e.g., PyQt5, PySide2) are *not* automatically installed. These must be installed separately. Using `conda install qtconsole` typically handles this dependency automatically.
  Fix: After `pip install qtconsole`, also run `pip install PyQt5` (or PySide2/6/PyQt6) or use `conda install qtconsole`.
• breaking In version 4.3, the `ConsoleWidget.width` and `ConsoleWidget.height` traits were renamed to `console_width` and `console_height` respectively to avoid clashes with `QWidget` properties.
  Fix: Update any code referencing `ConsoleWidget.width` or `ConsoleWidget.height` to use `console_width` or `console_height`.
• gotcha When embedding `qtconsole` in a custom Qt application, user code executed in the kernel runs in a separate process from the frontend. This means you cannot directly access the `QtApplication` instance of the `QtConsole` itself from within the kernel's executed code.
  Fix: Design your embedded application to handle the separation of processes for kernel execution and frontend GUI. Avoid attempting direct cross-process Qt object manipulation.
• gotcha Users have reported issues with the `qtconsole` freezing and throwing `QTextCursor::setPosition: out of range` errors after calling `input()`, and crashes when using the 'auto' backend with Matplotlib.
  Fix: Check for official patches or workarounds on the `qtconsole` GitHub issues page. For `input()` issues, consider alternative input methods if possible within your application context. For Matplotlib, explicitly set a non-'auto' backend.
• gotcha There are reports of `Segfault on import` when running `qtconsole` against Qt 6.7 on 64-bit Windows 11 with Python 3.12, indicating potential compatibility issues with newer Qt or Python versions.
  Fix: If encountering segfaults, consider using a slightly older, known-compatible combination of Python, Qt, and `qtconsole` versions, or monitor the GitHub repository for official fixes.

Install

• pip install qtconsole Via pip (Qt bindings need separate installation)
• conda install qtconsole Via Conda (includes Qt bindings)

Imports

• RichJupyterWidget

  from qtconsole.rich_jupyter_widget import RichJupyterWidget

• QtInProcessKernelManager

  from qtconsole.inprocess import QtInProcessKernelManager

Quickstart

The simplest way to start the Qt console is by running `jupyter qtconsole` from your terminal. For embedding it into a custom Qt application, you can use `RichJupyterWidget` to host a Jupyter kernel in-process. This example demonstrates launching an in-process kernel and executing commands.

import sys
from qtconsole.rich_jupyter_widget import RichJupyterWidget
from qtconsole.inprocess import QtInProcessKernelManager
from PyQt5.QtWidgets import QApplication

def main():
    app = QApplication([])
    
    # Create a kernel manager and start a kernel
    kernel_manager = QtInProcessKernelManager()
    kernel_manager.start_kernel(show_banner=False)
    kernel_client = kernel_manager.client()
    kernel_client.start_channels()

    # Create the Qt console widget
    console = RichJupyterWidget()
    console.kernel_manager = kernel_manager
    console.kernel_client = kernel_client
    
    # Execute some code programmatically
    console.execute('print("Hello from embedded QtConsole!")')
    console.execute('import matplotlib.pyplot as plt')
    console.execute('plt.plot([1,2,3], [4,5,6])')
    console.execute('plt.show()')

    console.show()
    sys.exit(app.exec_())

if __name__ == '__main__':
    # To run this example, ensure you have:
    # pip install qtconsole ipykernel PyQt5
    main()