Updated November 26th 2024

Martini for HOOMD: 

My implementation to streamline initializing a simulation parameterized with the Martini CG force field.

This works as a python package. Export to your python path

Usage for Simulation:
1. Initialize contents
2. Add molecules to contents
3. Run init_cell with contents
4. Either run simulation, or, in a new file, use forces_from_gsd to run from the initialized gsd

Usage for Development:

Adding molecules:
1. Add _bead.csv and _bonds.csv in martini the same format as they are implemented (note: bead is 1 indexed and bonds in 0 indexed)
2. Write add_*/make_* molecule to molecules.py following the given format 
3. Ready to use!

Adding polymers:
1. Be smart. Polymers.py constructs the _bead.csv and _bonds.csv for several types of polymers which I have had to code. Either work from that or make your own script to do this.

A summary of the components:

force_fields.py
- Initializes or reinitializes hoomd force fields
- Currently supports bond harmonic, angle harmonic, OPLS dihedral, improper harmonic, and rigid body cholesterol 

init_cell.py
- Given a contents object, creates a gsd with all components having correct bonds angles and impropers
- Saves bonds.csv, angles.csv ... etc, for rerun

particles.py
- Create a particles type
- Initialize leonard jones interactions across particle types and masses

molecules.py
- Creates bond/angles/rigid bodies/dihedral/whatever types  -- dihedral types are not tested
- Create molecule type - everything hoomd needs to run
- Creates contents, which is a list of molecules that keeps track of how many angles and bonds were added and prevents duplicate bonds from being added.
- Has functions to add specific molecules to contents.

polymers.py
- For polymers that I have needed, create _bonds and _bead.csv's for polymers given number of repeats, sequences, block-copolymers, slabs, etc.

Some notes on issues you may be experiencing:
1. There is a race condition if you try to start multiple simulations at the same time that build the same polymer AND that polymer has not been previously defined, hell breaks loose

Things on my list for development:
1. Support for combined bending and torsion potentials
2. Support for arbitrary rotation of cholesterol molecules on initialization (and resulting updates of the orientation quaternion)
3. More options for initial configuration of polymers (as opposed to my helicies)

Things that will be fixed in the next version:
1. _bead.csv and _bonds.csv will be a more intuitive structure
2. Generalized rigid body support
3. Support for all HOOMD bond, angle, and dihedral potentials
4. Support for custom masses instead of Martini Default masses