TRAGICS: TRajectory Analysis and Gauging In Chemical Space

A Python package for analyzing molecular dynamics trajectories, perform NEB calculations and more.

Features

• SOAP Analysis: Calculate SOAP descriptors and kernel similarities
• Geometric Analysis: Radius of gyration, atomic distances, RDF
• NEB Calculations: Nudged Elastic Band with climbing image support
• Frame Selection: Sequential similarity selection with SOAP from trajectories
• Visualization: Automated plotting of (some) analysis results
• Trajectory I/O: Filter and write trajectory subsets

Installation

# Install dependencies
pip install numpy matplotlib MDAnalysis ase dscribe

# For MACE support (optional)
pip install mace-torch

Place tragics folder in your working directory.

Quick Start

NEB Calculation

from tragics import TRAGICS
from mace.calculators import MACECalculator

traj = TRAGICS('trajectory.xyz', 'analysis.log')

# Basic NEB from trajectory frames
images, energies = traj.calculate_neb(
    calculator=MACECalculator(model_paths='model.model', device='cpu'),
    initial_frame=0,
    final_frame=100,
    n_images=11,
    fmax=0.05
)

# Advanced: climbing image + endpoint optimization + TS guess
images, energies = traj.calculate_neb(
    calculator=calculator,
    initial_file='initial.xyz',      # Or use initial_frame=0
    final_file='final.xyz',           # Or use final_frame=100
    ts_guess_file='guess.xyz',        # Optional TS guess
    n_images=15,
    optimize_endpoints=True,          # Pre-optimize endpoints
    use_climbing_image=True,          # Accurate TS location
    optimizer='FIRE',                 # Or 'BFGS'
    spring_constant=0.3               # Or None for auto-scale
)

print(f"Barrier: {max(energies):.3f} eV")

Key parameters:

• calculator: Any ASE calculator (MACE, EMT, etc.)
• Input: (initial_frame, final_frame) OR (initial_file, final_file)
• n_images=7: Images including endpoints
• fmax=0.05: Convergence criterion (eV/Å)
• use_climbing_image=False: Enable for accurate TS
• optimize_endpoints=False: Pre-optimize reactant/product

Outputs: XYZ path, ASE trajectory, energy plot/CSV, detailed log

Parallel: Automatic MPI support via ASE: mpirun -np 4 python script.py

Other TRAGICS Features

Geometric Analysis

frames, rg = traj.calculate_radius_of_gyration()
distances = traj.calculate_distance(atom1_idx=0, atom2_idx=1)

Trajectory Filtering

traj.filter_trajectory(
    output_file='subset.xyz',
    frames_to_write=[0, 10, 20, 30]
)

Version

Current version: 0.2.0 (with NEB support)

Authors

Gers

License

MIT License