pyamares-xmris 0.5.0

Requirements

Installation

pip install pyamares-xmris

(The import name remains pyAMARES. To install the original upstream distribution instead, use pip install pyAMARES — note that it does not install on arm64.)

That is the fitting engine only: numpy, scipy, pandas, matplotlib, lmfit, sympy, nmrglue, jinja2 and tqdm. Everything else lives behind an extra, and each one is additive:

pip install 'pyamares-xmris[jupyter]'   # the notebooks and interactive workflows: Jupyter, requests, and the matlab/excel extras below
pip install 'pyamares-xmris[matlab]'    # mat73, for reading MATLAB v7.3 (HDF5) .mat files
pip install 'pyamares-xmris[excel]'     # openpyxl and xlrd, for .xlsx / .xls prior knowledge spreadsheets
pip install 'pyamares-xmris[hlsvd]'     # hlsvdpro, the optional native HSVD backend (x86_64 only; unused on numpy 2+)

Prior knowledge in CSV form and MATLAB files older than v7.3 need no extra. If an extra is missing, the function that needs it raises an ImportError naming the one to install. The HSVD path never needs [hlsvd]: without it pyAMARES uses its own bundled pure-Python implementation, which is what every numpy 2 install has used all along.

See the Installation Guide for detailed information.

Run pyAMARES in any web browser

New: PyAMARES now offers a user-friendly web interface for fitting AMARES models without writing any code. The web app provides a graphical interface to:

• Upload your FID data file

• Upload the prior knowledge spreadsheet (Excel or CSV)

• Edit the prior knowledge spreadsheet as needed

• Set MR parameters (MHz, spectrum width, deadtime, etc)

• Visualize results with interactive plots

• Download fitted results and figures

No installation required - just visit the link and start fittting your MRS data right away!

Run pyAMARES as standard-alone script

amaresFit -f ./pyAMARES/examples/fid.txt -p  ./pyAMARES/examples/example_human_brain_31P_7T.csv --MHz 120.0 --sw 10000 --deadtime 300e-6 --ifplot --xlim 10 -20 -o simple_example

Run pyAMARES in a Jupyter Notebook

Try Jupyter Notebook on Google Colab here

import pyAMARES
# Load FID from a 2-column ASCII file, and set the MR parameters
MHz = 120.0 # 31P nuclei at 7T
sw = 10000 # spectrum width in Hz
deadtime = 300e-6 # 300 us begin time for the FID signal acquisition

fid = pyAMARES.readmrs('./pyAMARES/examples/fid.txt')
# Load Prior Knowledge
FIDobj = pyAMARES.initialize_FID(fid=fid,
                                 priorknowledgefile='./pyAMARES/examples/example_human_brain_31P_7T.csv',
                                 MHz=MHz,
                                 sw=sw,
                                 deadtime=deadtime,
                                 preview=False,
                                 normalize_fid=False,
                                 xlim=(10, -20))# Region of Interest for visualization, -20 to 10 ppm

# Initialize the parameter using Levenberg-Marquard method
out1 = pyAMARES.fitAMARES(fid_parameters=FIDobj,
                           fitting_parameters=FIDobj.initialParams,
                           method='leastsq',
                           ifplot=False)

# Fitting the MRS data using the optimized parameter

out2 = pyAMARES.fitAMARES(fid_parameters=out1,
                          fitting_parameters=out1.fittedParams, # optimized parameter for last step
                          method='least_squares',
                          ifplot=False)

# Save the data
out2.styled_df.to_html('simple_example.html') # Save highlighted table to an HTML page
                                              # Python 3.6 does not support to_html.
out2.result_sum.to_csv('simple_example.csv') # Save table to CSV spreadsheet
out2.plotParameters.lb = 2.0 # Line Broadening factor for visualization
out2.plotParameters.ifphase = True # Phase the spectrum for visualization
pyAMARES.plotAMARES(fid_parameters=out1, filename='simple_example.svg') # Save plot to SVG

Fitting Result for Example 31P MRS data