utils.py

from collections import OrderedDict
import datetime
import klepto
from os.path import dirname, abspath, join, expanduser, isfile, exists
from os import environ, makedirs
import pytz
import re
from socket import gethostname



def get_root_path():
    return dirname(abspath(__file__))


def get_data_path():
    return join(get_root_path(), 'data')


def get_corpus_path():
    return join(get_data_path(), 'corpus')


def get_save_path():
    return join(get_root_path(), 'save')


def load(filepath, print_msg=True):
    fp = proc_filepath(filepath)
    if isfile(fp):
        return load_klepto(fp, print_msg)
    elif print_msg:
        print('Trying to load but no file {}'.format(fp))


def load_klepto(filepath, print_msg):
    rtn = klepto.archives.file_archive(filepath)
    rtn.load()
    if print_msg:
        print('Loaded from {}'.format(filepath))
    return rtn


def save(obj, filepath, print_msg=True):
    if type(obj) is not dict and type(obj) is not OrderedDict:
        raise ValueError('Can only save a dict or OrderedDict'
                         ' NOT {}'.format(type(obj)))
    fp = proc_filepath(filepath, ext='.klepto')
    create_dir_if_not_exists(dirname(filepath))
    save_klepto(obj, fp, print_msg)


def create_dir_if_not_exists(dir):
    if not exists(dir):
        makedirs(dir)


def save_klepto(dic, filepath, print_msg):
    if print_msg:
        print('Saving to {}'.format(filepath))
    klepto.archives.file_archive(filepath, dict=dic).dump()


def proc_filepath(filepath, ext='.klepto'):
    if type(filepath) is not str:
        raise RuntimeError('Did you pass a file path to this function?')
    return append_ext_to_filepath(ext, filepath)


def append_ext_to_filepath(ext, fp):
    if not fp.endswith(ext):
        fp += ext
    return fp


def parse_as_int_list(il):
    rtn = []
    for x in il.split('_'):
        x = int(x)
        rtn.append(x)
    return rtn


def get_user():
    try:
        home_user = expanduser("~").split('/')[-1]
    except:
        home_user = 'user'
    return home_user


def get_host():
    host = environ.get('HOSTNAME')
    if host is not None:
        return host
    return gethostname()

tstamp = None


def get_ts():
    global tstamp
    if not tstamp:
        tstamp = get_current_ts()
    return tstamp


def get_current_ts(zone='US/Pacific'):
    return datetime.datetime.now(pytz.timezone(zone)).strftime(
        '%Y-%m-%dT%H-%M-%S.%f')


def sorted_nicely(l, reverse=False):
    def tryint(s):
        try:
            return int(s)
        except:
            return s

    def alphanum_key(s):
        if type(s) is not str:
            raise ValueError('{} must be a string in l: {}'.format(s, l))
        return [tryint(c) for c in re.split('([0-9]+)', s)]

    rtn = sorted(l, key=alphanum_key)
    if reverse:
        rtn = reversed(rtn)
    return rtn