Installation

Requirements

  • Python 3.10 or higher

  • IPython/Jupyter Notebook or JupyterLab

  • Operating System: Linux, macOS, or Windows

Dependencies

The following packages will be installed automatically:

  • ipython: For magic command integration

  • psutil: For I/O tracking on Linux and Windows

  • matplotlib: For histogram visualization

  • numpy: For data processing in histograms

Installing from PyPI

The easiest way to install iops-profiler is from PyPI using pip:

>> pip install iops-profiler

Installing from Source

To install the latest development version from GitHub:

>> git clone https://github.com/lincc-frameworks/iops-profiler.git
>> cd iops-profiler
>> pip install -e .

Development Installation

If you want to contribute to iops-profiler, install with development dependencies:

>> pip install -e '.[dev]'
>> pre-commit install

This installs additional tools for testing, linting, and pre-commit hooks.

Platform-Specific Setup

Linux: No additional setup required.

macOS: Each time you run %iops or %%iops, you will be prompted for your password (fs_usage requires elevated privileges).

Windows: No additional setup required.

For detailed platform information, see Platform-Specific Notes.

Verifying Installation

To verify that iops-profiler is installed correctly, start Python or a Jupyter notebook and run:

import iops_profiler
print(iops_profiler.__version__)

Then load the extension:

%load_ext iops_profiler

If there are no errors, the installation is successful!

Quick Test

Run a simple test to ensure everything works:

%load_ext iops_profiler

%%iops
with open('test.txt', 'w') as f:
    f.write('Hello, IOPS!')

You should see a results table with I/O metrics.

Next Steps

Now that you have iops-profiler installed: