def preliminary_preps():
"""Set up logger and read command line arguments"""
# logger
logfile_name = os.path.join(tc.LOGDIR, "run_%s.log" % prep.find_run_id())
lgr, lfh = prep.set_log(__name__, logfile_name, False)
# cl options
clargs = set_option_parser()
if clargs.tag is not None and clargs.tag:
tc.TAG = True
elif clargs is not None and not clargs.tag:
tc.TAG = False
#TODO: options below don't seem to be able to affect tc other than for writing to the cumulog
elif clargs.maxdista is not None:
tc.maxdista = clargs.maxdista
elif clargs.distaw is not None:
tc.distaw = clargs.distaw
elif clargs.lmw is not None:
tc.lmw = clargs.lmw
return lgr, lfh, clargs
# -*- coding: utf-8 -*-
import codecs
from collections import defaultdict
import os
import re
import tnconfig as tc
import preparation as prep
# logging
logfile_name = os.path.join(tc.LOGDIR, "run_%s.log" % prep.find_run_id())
lgr, lfh = prep.set_log(__name__, logfile_name, False)
class EdScoreMatrix:
"""Methods to read cost matrix from module in arg cost_module
and to find costs for individual character-edits."""
def __init__(self, cost_module):
self.costm = cost_module
row_names = None
col_names = None
matrix_conts = None
accented_chars = [u'\xe1', u'\xe9', u'\xed', u'\xf1', u'\xf3', u'\xfa', u'\xfc']
matrix_stats = {"max" : None, "min" : None, "ave" : None}
def read_cost_matrix(self):
"""Read cost matrix into a hash. Set instance values for them"""
row_names = self.costm.row_names.strip().split("\t")
col_names = self.costm.col_names.strip().split("\t")
costs = self.costm.costs