Lorentz Symmetry Violations of Superluminal Jets

Bayesian inference framework for constraining Lorentz violation parameters using superluminal astrophysical jet observations. Uses PyMC/PyTensor MCMC sampling to fit a custom log-likelihood model to transverse velocity data from the MOJAVE survey or synthetic datasets. The model constrains a scalar parameter (\(B_0\)) and a 3-vector (\(\vec{B}\)) that parameterize anisotropic Lorentz violation.

Pipeline

 ┌──────────────────────┐     ┌───────────────────────┐
 │  MOJAVE HTML table   │     │  Simulation generator  │
 │  Mojave_html_to_csv  │     │  aniso_simulated_data  │
 └─────────┬────────────┘     └───────────┬────────────┘
           │                              │
           ▼                              ▼
   mojave_cleaned.csv           generated_sources.csv
           │                              │
           └──────────┬───────────────────┘
                      │
                      ▼
            ┌─────────────────┐
            │ simulationImport │
            │   importCSV()   │
            └────────┬────────┘
                     │
          ┌──────────┴──────────┐
          │                     │
          ▼                     ▼
  ┌───────────────┐    ┌───────────────┐
  │  Anisotropic  │    │  Isotropic    │
  │  MCMC model   │    │  MCMC model   │
  │  (B0, B_vec)  │    │  (q)          │
  └───────┬───────┘    └───────┬───────┘
          │                    │
          ▼                    ▼
    trace plots           trace plots
    Mollweide map         posterior
    pair plot             summaries
    posterior

Quick Start

# Install dependencies (Python 3.13+, managed with uv)
uv sync

# Run the anisotropic MCMC analysis
python scripts/PyMC_Pytensor_noise_v_main_varsig_aniso.py

# Run the isotropic MCMC analysis
python scripts/PyMC_Pytensor_noise_v_main_varsig.py

# Generate synthetic data only
python scripts/aniso_simulated_data_gen_v.py

# Fetch MOJAVE data
python scripts/Mojave_html_to_csv.py

Scripts

Script	Purpose
Anisotropic Model	MCMC sampling of \(B_0\) and \(\vec{B}\) with direction-dependent \(w_c\)
Isotropic Model	MCMC sampling of a single scalar parameter \(q\) (\(w_c = q + 1\))
Data Generator	Produces synthetic velocity data with configurable LV parameters
Simulation Import	Shared CSV loader for both datasets
MOJAVE Scraper	Fetches and cleans the MOJAVE velocity table