Skip to content

Implementing and Testing Jacobian Functions in Stingray in models.py #898

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom
Open
wants to merge 6 commits into
base: main
Choose a base branch
from
Open

Implementing and Testing Jacobian Functions in Stingray in models.py #898

wants to merge 6 commits into from

Conversation

kashish2210
Copy link
Member

Contribution: Implementing and Testing Jacobian Functions in Stingray

Overview

This contribution implements and tests the Jacobian functions for GeneralizedLorentzianJacobian and SmoothBrokenPowerLawJacobian in stingray/simulator/models.py. These were listed as TODO and have now been completed and verified with test plots.

Implemented Functions

The following Jacobian functions were added: [file_change]

Testing and Verification

The functions were tested using the following code to ensure correctness and visualize the Jacobian components.

x_values = np.linspace(0.1, 10, 100)

# Test values for Generalized Lorentzian Jacobian
params_lorentz = (5.0, 1.0, 1.0, 2.0)  # (x_0, fwhm, value, power_coeff)
lorentzian_results = generalizedLorentzianJacobian(x_values, *params_lorentz)

# Test values for Smooth Broken Power Law Jacobian
params_powerlaw = (1.0, 1.5, 2.5, 3.0)  # (norm, gamma_low, gamma_high, break_freq)
powerlaw_results = SmoothBrokenPowerLawJacobian(x_values, *params_powerlaw)

# Plot results
fig, axes = plt.subplots(2, 1, figsize=(10, 8))

# Plot Lorentzian Jacobian components
axes[0].plot(x_values, lorentzian_results[:, 0], label='d_x0')
axes[0].plot(x_values, lorentzian_results[:, 1], label='d_fwhm')
axes[0].plot(x_values, lorentzian_results[:, 2], label='d_value')
axes[0].plot(x_values, lorentzian_results[:, 3], label='d_power_coeff')
axes[0].set_title("Generalized Lorentzian Jacobian Components")
axes[0].set_xlabel("x")
axes[0].set_ylabel("Jacobian Component Values")
axes[0].legend()
axes[0].grid()

# Plot Smooth Broken Power Law Jacobian components
axes[1].plot(x_values, powerlaw_results[:, 0], label='d_norm')
axes[1].plot(x_values, powerlaw_results[:, 1], label='d_gamma_low')
axes[1].plot(x_values, powerlaw_results[:, 2], label='d_gamma_high')
axes[1].plot(x_values, powerlaw_results[:, 3], label='d_break_freq')
axes[1].set_title("Smooth Broken Power Law Jacobian Components")
axes[1].set_xlabel("x")
axes[1].set_ylabel("Jacobian Component Values")
axes[1].legend()
axes[1].grid()

plt.tight_layout()
plt.show()

Results

Figure_1_jacobian

Helper_Notes:

  • I have tried implementing @custom_model from astropy.modeling.Model and found an error
self._initialize_parameters(args, kwargs)
raise TypeError(
    ...<2 lines>...
    )
  • then I found out
    Why Not Use @custom_model?
@custom_model expects a function returning a single array, while the Jacobian returns multiple derivatives.
@custom_model does not directly handle gradients.

This contribution addresses @matteobachetti's # TODO: Add Jacobian in stingray/simulator/models.py by implementing and testing the required Jacobian functions. The code has been validated with test plots demonstrating correct behavior.

guide me for my workflow:

@kashish2210
Copy link
Member Author

for further testing with some complications:

testing code:

# Define test values
x = np.logspace(-2, 2, 100)

# Parameters for GeneralizedLorentz1D
x_0 = 1.0
fwhm = 1.0
value = 1.0
power_coeff = 2.0

# Parameters for SmoothBrokenPowerLaw
norm = 1.0
gamma_low = 1.0
gamma_high = 2.0
break_freq = 1.0

# Compute Jacobians using provided functions
jac_lorentz = GeneralizedLorentz1DJacobian(x, x_0, fwhm, value, power_coeff)
jac_smooth_bkn_po = SmoothBrokenPowerLawJacobian(x, norm, gamma_low, gamma_high, break_freq)

# Plot Jacobians
fig, ax = plt.subplots(figsize=(10, 5))
ax.loglog(x, np.abs(jac_lorentz[:, 0]), label='d/dx_0 (Lorentz)', linestyle='dashed')
ax.loglog(x, np.abs(jac_lorentz[:, 1]), label='d/dFWHM (Lorentz)', linestyle='dotted')
ax.loglog(x, np.abs(jac_lorentz[:, 2]), label='d/dValue (Lorentz)', linestyle='dashdot')
ax.loglog(x, np.abs(jac_lorentz[:, 3]), label='d/dPowerCoeff (Lorentz)', linestyle='solid')
ax.loglog(x, np.abs(jac_smooth_bkn_po[:, 0]), label='d/dNorm (Smooth BPL)', linestyle='dashed', color='red')
ax.loglog(x, np.abs(jac_smooth_bkn_po[:, 1]), label='d/dGammaLow (Smooth BPL)', linestyle='dotted', color='green')
ax.loglog(x, np.abs(jac_smooth_bkn_po[:, 2]), label='d/dGammaHigh (Smooth BPL)', linestyle='dashdot', color='purple')
ax.loglog(x, np.abs(jac_smooth_bkn_po[:, 3]), label='d/dBreakFreq (Smooth BPL)', linestyle='solid', color='brown')
ax.set_xlabel('Frequency (x)')
ax.set_ylabel('Jacobian Values (Absolute)')
ax.legend()
ax.set_title('Jacobian Values for Generalized Lorentzian and Smooth Broken Power Law')
ax.grid(True, which='both', linestyle='--', linewidth=0.5)

plt.show()

# Print Jacobian values for verification
print("Jacobian for Generalized Lorentzian (first few values):\n", jac_lorentz[:5])
print("Jacobian for Smooth Broken Power Law (first few values):\n", jac_smooth_bkn_po[:5])

output:

Jacobian for Generalized Lorentzian (first few values):
 [[ 0.327133    0.57168971  0.20323551 -0.22191759]
 [ 0.32783851  0.57284784  0.2035548  -0.22213299]
 [ 0.32861489  0.57412271  0.20390602 -0.22236968]
 [ 0.32946947  0.57552649  0.20429244 -0.22262979]
 [ 0.33041041  0.57707272  0.20471769 -0.22291567]]
Jacobian for Smooth Broken Power Law (first few values):
 [[ 1.00005000e+02  4.60540044e+02  5.00000000e-03 -9.99950004e-03]
 [ 9.11217629e+01  4.11153827e+02  5.48749382e-03 -1.09743267e-02]
 [ 8.30277791e+01  3.66908283e+02  6.02251770e-03 -1.20441617e-02]
 [ 7.56529422e+01  3.27279845e+02  6.60970573e-03 -1.32182566e-02]
 [ 6.89333748e+01  2.91797403e+02  7.25414388e-03 -1.45067611e-02]]

Figure_2_jacobian
let me know about bugs

@codecov Codecov
Copy link

codecov bot commented Mar 12, 2025 •
edited
Loading

Codecov Report

Attention: Patch coverage is 20.00000% with 16 lines in your changes missing coverage. Please review.

Project coverage is 94.09%. Comparing base (a91c031) to head (d513fe6).

Files with missing lines Patch % Lines
stingray/simulator/models.py 20.00% 16 Missing ⚠️
Additional details and impacted files 
@@            Coverage Diff             @@
##             main     #898      +/-   ##
==========================================
- Coverage   96.03%   94.09%   -1.95%     
==========================================
  Files          48       48              
  Lines        9770     9788      +18     
==========================================
- Hits         9383     9210     -173     
- Misses        387      578     +191

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:
  • ❄ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

@kashish2210 kashish2210 reopened this Mar 12, 2025
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers
No reviews
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.
1 participant
@kashish2210