pulsarfitpy is a tool intended to streamline analysis of pulsars, a rapidly rotating neutron star. pulsarfitpy offers an intuitive PyTorch framework using Physics Informed Neural Networks (PINNs) to analyze differential equations in comparison to observed ATNF data. The application of the PINN framework provides an additional resource in pulsar analysis to test derived theoretical physics expressions and determine their validity.
- 1D PINN Support
- 2D PINN Support
- Python Library, Command Line Interface, and Graphical User Interface Options
- Real Time Error Metrics
- Customizable Neural Networks
- Python 3.12 or higher
- Go 1.21 or higher (for CLI)
To use all features of pulsarfitpy, install the following Python modules.
- psrqpy
- numpy
- torch
- sympy
- typing
- logging
- matplotlib
- argparse
- scikit-learn
- deepxde
- dataclasses
- pathlib
- pandas
You may install these by running either one of the following commands:
pip install -r requirements.txtpip install psrqpy numpy torch sympy typing logging matplotlib argparse scikit-learn deepxde dataclasses pathlib pandasFor the former, please ensure that you are in the correct directory containing the requirements text file.
pip install pulsarfitpygit clone https://github.com/jfk-astro/pulsarfitpy.git
cd pulsarfitpy
pip install -e .# Clone the repository
git clone https://github.com/jfk-astro/pulsarfitpy.git
cd pulsarfitpy
# Build the CLI
make buildThe binary will be available at bin/pulsar-cli.exe (Windows) or bin/pulsar-cli (Unix).
You may find a compilation of more detailed information about pulsarfitpy at this website.
pulsarfitpy is under the GNU GPL v3.0 license, a free, copyleft license published by the Free Software Foundation.
pulsarfitpy is made by jfk-astro consisting of Om Kasar, Saumil Sharma, Kason Lai, and Jonathan Sorenson.