A PCA Trainer and transformer, train pca from csv file

ipca.py

#!/bin/python
"""CCIPCA Trainer and transformer.

Usage:
  ipca.py train <model_file_name> <csvfile>... [--ncomp=<nc>] [--ntrainiter=<nti>] [--itersave]
  ipca.py trans <model_file_name> <csvfile>... [--new_prefix=<prefix>]
  ipca.py (-h|--help)

Options:
  --ncomp=<nc>                      Number of Component [default: 30].
  -h --help                         Show this screen.
  --new_prefix=<prefix>             Prefix of new filename [default: trans_].
  --ntrainiter=<nti>                Number of Train Iterator [default: 1].
  --itersave                        Save model when iterator complete.
"""

try:
    import cPickle as pickle
except ImportError:
    import pickle
import csv
import numpy as np
from docopt import docopt
from pyIPCA import CCIPCA


def load_pca(args):
    try:
        with open(args['<model_file_name>'], 'rb') as f:
            pca = pickle.load(f)
            return pca
    except IOError, pickle.PickleError:
        pca = CCIPCA(n_components=int(args['--ncomp']))
        return pca


def train_pca(pca, fn):
    data = np.genfromtxt(fn, delimiter=',')
    pca.fit(data)


def trans_pca(pca, fn, pre):
    data = np.genfromtxt(fn, delimiter=',')
    result = pca.transform(data)
    nsamp, ncomp = result.shape
    with open(pre + fn, 'w') as f:
        writer = csv.writer(f)
        for i in range(0, nsamp):
            row = [result[i, j] for j in range(0, ncomp)]
            writer.writerow(row)


def save_pca(pca, args):
    with open(args['<model_file_name>'], 'wb') as f:
        pickle.dump(pca, f)


def main():
    args = docopt(__doc__, version='CCIPCA Trainer V0.1.0')
    if args['train']:
        # Load PCA Model
        pca = load_pca(args)
        # Train PCA By CSVFile
        nit = int(args['--ntrainiter'])
        for _ in range(0, nit):
            for fn in args['<csvfile>']:
                try:
                    train_pca(pca, fn)
                except:
                    pass
                if args['--itersave']:
                    save_pca(pca, args)
        save_pca(pca, args)
    elif args['trans']:
        pca = load_pca(args)
        for fn in args['<csvfile>']:
            trans_pca(pca, fn, args['--new_prefix'])


if __name__ == '__main__':
    main()