The HDF5_BLS_treat package

This package is a part of the HDF5_BLS project. It's purpose is to unify the extraction of relevant information from power spectra data.

Installation

To install the package, you can use pip:

pip install HDF5_BLS_treat

To install the package from source for local development, please refer to the documentation.

Documentation

You can access the documentation of the project at this link or on the dedicated ReadTheDocs page at this link.

Example of usage

from HDF5_BLS_treat import Treat, Models
import numpy as np
import sys

def progress(current, total):
    """
    Function to display the progress of the treatment.
    """
    percent = (current / total) * 100
    sys.stdout.write(f"\rProgress: sample {current} of {total} - {percent:.2f}% treated")
    sys.stdout.flush()

# We first recover the model we are going to use to simulate the data
models = Models()
models = models.models
lorentzian = models["DHO"]

# We then create a frequency axis and a PSD based on the model
frequency = np.linspace(-10,10,512)
psd = lorentzian(frequency, 0, 1, -5, 1) + lorentzian(frequency, 0, 1, 5, 1) + 0.1 * np.random.normal(size = frequency.shape)

# We set the list parameters for the analysis
points = [-5, 5] # An initial guess for the center of the peak
windows_find = [3, 3] # The size of the windows around the peaks given to refine their position
windows_fit = [[3, 1.5], [1.5, 3]] # The windows around the peaks to fit them
point_type = ["Anti-Stokes", "Stokes"] # The nature of the peaks to fit
max_width_guess = 2 # The maximal estimation allowed for the linewidth before the fitting
bound_shift = [[-7, -3], [3, 7]] # The bounds for the fitting of the shift
bound_linewidth = [[0, 5], [0, 5]] # The bounds for the fitting of the linewidth

# Initialising the Treat object on the a doublet of frequency and PSD
treat = Treat(frequency = frequency, PSD = psd)

# Setting the progress callback (optional, this allows us to display the progress of the treatment on large datasets)
treat._progress_callback = progress

# Adding the points to normalize the data
for point, window in zip(points, windows_find):
    treat.add_point(position_center_window = point, type_pnt = "Other", window_width = window)
treat.normalize_data(threshold_noise = 0.05)

# Adding the peaks to fit
for point, window, tpe in zip(points, windows_fit, point_type):
    treat.add_point(position_center_window = point, type_pnt = tpe, window_width = window)

# Defining the model for fitting the peaks - you can choose between the models defined in the Models class
treat.define_model(model = "Lorentzian", elastic_correction = True)

# Estimating the linewidth from selected peaks
treat.estimate_width_inelastic_peaks(max_width_guess = max_width_guess)

# Fitting all the selected inelastic peaks with multiple peaks fitting
treat.single_fit_all_inelastic(guess_offset = True, 
                                update_point_position = True, 
                                bound_shift = bound_shift, 
                                bound_linewidth = bound_linewidth)

# Applying the algorithm to all the spectra (in the case where PSD is a 2D array)
treat.apply_algorithm_on_all()

# Combining the two fitted peaks together here weighing the result on the standard deviation of the shift
treat.combine_results_FSR(FSR = 60, keep_max_amplitude = False, amplitude_weight = False, shift_err_weight= True)

# Printing the results
print("\nFinal results:")
print(f"Shfit: {treat.shift}")
print(f"Linewidth: {treat.linewidth}")
print(f"Amplitude: {treat.amplitude}")