Using LAMMPS for HiC
Guido Polles <guido.polles@gmail.com>

This is my module for performing Population Modeling from HiC data using LAMMPS.
The lammps_hic module is based on ipyparallel. The UI -if ever- will be based
on Tornado (as a web interface).
Needs:
numpy
h5py
cloudpickle
ipyparallel
tornado


Sub-modules:

1. lammps
Contains scripts to generate lammps inputs, call lammps and extract coordinates
and info from the lammps output.
The main function here is 
 lammps.lammps_minimize(crd, 
                        radius, 
                        chrom, 
                        crd_id, 
                        tmp_files_dir='/dev/shm', 
                        log_dir='.', 
                        **kwargs)
    crd: numpy array of coordinates
    radius: float single radius  # TODO: handle multiple radiuses
    chrom: list of chromosome string for each bead (not only the haploid)
    crd_id: a string to identify the run, used in filenames
    tmp_files_dir: directory where to store temporary lammps files (they are 
                   removed before function returns)
    log_dir: directory where to store log files in case of error
    **kwargs: dictionary with arguments for the minimization, default values
              are in lammps.ARG_DEFAULT

returns a pair (crd, info):
    crd: numpy array of minimized coordinates
    info: dictionary with information about the run

2. actdist
Contains the parallel version of the functions to compute the activation distances.
 actdist.get_actdists(parallel_client,
                      crd_fname, 
                      probability_matrix, 
                      theta, 
                      last_ad, 
                      save_to=None)
    parallel_client: an instance to an ipyparallel.Client()
    crd_fname: the h5py file of the population
    probability_matrix: 2D HiC input matrix
    theta: the ptobability parameter
    last_ad: the last activation distance structure or an empty iterable. Activation
             distances are considered here as numpy recarrays with the following
             items
              ('i', int),
              ('j', int),
              ('pwish', float),
              ('actdist', float),
              ('pclean', float),
              ('pnow', float)] 
    save_to: filename where to save a text file of the computed activation distances

Sub-packages:

1. webui