ueBLOG

#! /usr/bin/env python
# vim: fileencoding=utf8

import os
import sys
import shutil
import logging as logger

def copyfile(path, to_path, ext):
    """
    pathの中の全ての拡張子がextのファイルを
    to_pathにコピー
    """
    logger.debug("args path %s to_path %s ext %s " % (path, to_path, ext))
    # os.walk で　path内のディレクトリを全部なめる
    # ディレクトリのなかにディレクトリがあっても一つ一つ取得してくれる
    for root, dirs, flist in os.walk(path):
        # flistはファイルの配列なので一つ一つ処理
        for fname in flist:
            logger.debug("root:%s fname:%s" % (root, fname))
            # hoge.JPG のようなファイル名から拡張子を取得
            # 大文字は小文字にする
            # ex)
            # os.path.splitext(fname) で ("hoge", ".JPG") に分解
            # [-1] で .JPG  (上の要素の一番最後を取得
            # .lower() で .jpg ( すべて小文字に変換
            # [1:] で jpg (2番目の文字から最後までを取得
            file_ext = os.path.splitext(fname)[-1].lower()[1:]
            logger.debug("file ext %s " % file_ext)
            # 拡張子が一致したら処理
            if ext.lower() == file_ext:
                # root(c:\temp\hogeのような親ディレクトリ)とhoge.JPG を 合体させる
                filepath = os.path.join(root, fname)
                # to_pathにコピー
                shutil.copy2(filepath, to_path)
                logger.info("copy %s to %s" % (filepath, to_path))


# if __name__ == "main": でバッチとして実行できる
if __name__ == "__main__":
    # sys.argv には　実行ファイル名　オプション ,,,, のリストが入っている
    if len(sys.argv) < 3:
        print r"usage copyfile.py c:\temp\hoge c:\temp\huga jpg"
        sys.exit()
    path = sys.argv[1]
    to_path = sys.argv[2]
    ext = sys.argv[3]
    copyfile(path, to_path, ext)

#! /usr/bin/env python
#! vim: fileencoding=utf8

import Image

def tobitmap(imagepath, outputfilename):
    im = Image.open(imagepath)
    im = im.convert("1") #白黒に変換
    im.tobitmap() #bitmapに変換
    wf = open(outputfilename, "wb")
    im.save(wf)
    wf.close()


if __name__ == "__main__":
    path = r"C:\temp\image"
    tobitmap(r"C:\temp\image\ruko.jpg", r"C:\temp\image\ruko.bmp")

#! /usr/bin/env python
#! vim: fileencoding=utf8

import datetime

def test_relativedelta():
    """
    datetime.date型の加算減算のテスト
    monthsなどsが付くと加算、
    monthなどsが付かない場合は指定した数字になる

    >>> from dateutil import relativedelta
    >>> d = datetime.date(2009, 11, 3)
    >>> d + relativedelta.relativedelta(months=+1)
    datetime.date(2009, 12, 3)
    >>> d + relativedelta.relativedelta(month=2, days=+3)
    datetime.date(2009, 2, 6)
    >>> d + relativedelta.relativedelta(months=+2, day=20)
    datetime.date(2010, 1, 20)

    月末のテスト
    >>> d + relativedelta.relativedelta(months=+3, day=99)
    datetime.date(2010, 2, 28)
    """


def test_parser():
    """
    文字列からdatetime.datetimeを生成するテスト
    ありがちなフォーマットなら指定なしでparseしてくれる

    >>> from dateutil import parser
    >>> parser.parse("2010-01-22")
    datetime.datetime(2010, 1, 22, 0, 0)
    >>> parser.parse("2010/01/22")
    datetime.datetime(2010, 1, 22, 0, 0)
    >>> parser.parse("20100122")
    datetime.datetime(2010, 1, 22, 0, 0)
    >>> parser.parse("2010 01 22")
    datetime.datetime(2010, 1, 22, 0, 0)
    """

def test_rrule():
    """
    特定の期間の月曜日を取得するテスト
    >>> from dateutil.rrule import rrule, MO, WEEKLY, MONTHLY
    >>> list(rrule(WEEKLY, byweekday=(MO),
    ...   dtstart=datetime.datetime(2010,2,21),
    ...   until=datetime.datetime(2010,4,20)))
    [datetime.datetime(2010, 2, 22, 0, 0), datetime.datetime(2010, 3, 1, 0, 0), datetime.datetime(2010, 3, 8, 0, 0), datetime.datetime(2010, 3, 15, 0, 0), datetime.datetime(2010, 3, 22, 0, 0), datetime.datetime(2010, 3, 29, 0, 0), datetime.datetime(2010, 4, 5, 0, 0), datetime.datetime(2010, 4, 12, 0, 0), datetime.datetime(2010, 4, 19, 0, 0)]

    特定期間の月最初の月曜日のリストを取得するテスト
    >>> list(rrule(MONTHLY, byweekday=(MO(1)), dtstart=datetime.datetime(2010,2,21), until=datetime.datetime(2010,7,20)))
    [datetime.datetime(2010, 3, 1, 0, 0), datetime.datetime(2010, 4, 5, 0, 0), datetime.datetime(2010, 5, 3, 0, 0), datetime.datetime(2010, 6, 7, 0, 0), datetime.datetime(2010, 7, 5, 0, 0)]

    特定期間の月最後の月曜日のリストを取得するテスト
    >>> list(rrule(MONTHLY, byweekday=(MO(-1)), dtstart=datetime.datetime(2010,2,21), until=datetime.datetime(2010,7,20)))
    [datetime.datetime(2010, 2, 22, 0, 0), datetime.datetime(2010, 3, 29, 0, 0), datetime.datetime(2010, 4, 26, 0, 0), datetime.datetime(2010, 5, 31, 0, 0), datetime.datetime(2010, 6, 28, 0, 0)]

    """

def _test():
    import doctest
    doctest.testmod()


if __name__ == "__main__":
    _test()

2章　推薦を行う
       2.1　協調フィルタリング
       2.2　嗜好の収集
       2.3　似ているユーザを探し出す
               2.3.1　ユークリッド距離によるスコア
               2.3.2　ピアソン相関によるスコア
               2.3.3　どちらの類似性尺度を利用すべきなのか？
               2.3.4　評者をランキングする
       2.4　アイテムを推薦する

critics={'Lisa Rose': {'Lady in the Water': 2.5, 'Snakes on a Plane': 3.5,
    'Just My Luck': 3.0, 'Superman Returns': 3.5, 'You, Me and Dupree': 2.5,
    'The Night Listener': 3.0},
    'Gene Seymour': {'Lady in the Water': 3.0, 'Snakes on a Plane': 3.5,
        'Just My Luck': 1.5, 'Superman Returns': 5.0, 'The Night Listener': 3.0,
        'You, Me and Dupree': 3.5},
    'Michael Phillips': {'Lady in the Water': 2.5, 'Snakes on a Plane': 3.0,
        'Superman Returns': 3.5, 'The Night Listener': 4.0},
    'Claudia Puig': {'Snakes on a Plane': 3.5, 'Just My Luck': 3.0,
        'The Night Listener': 4.5, 'Superman Returns': 4.0,
        'You, Me and Dupree': 2.5},
    'Mick LaSalle': {'Lady in the Water': 3.0, 'Snakes on a Plane': 4.0,
        'Just My Luck': 2.0, 'Superman Returns': 3.0, 'The Night Listener': 3.0,
        'You, Me and Dupree': 2.0},
    'Jack Matthews': {'Lady in the Water': 3.0, 'Snakes on a Plane': 4.0,
        'The Night Listener': 3.0, 'Superman Returns': 5.0, 'You, Me and Dupree': 3.5},
    'Toby': {'Snakes on a Plane':4.5,'You, Me and Dupree':1.0,'Superman Returns':4.0}}

#! /usr/bin/env python
#! vim: fileencoding=utf8

import os

from math import pow, sqrt

from sqlalchemy import create_engine
from sqlalchemy import Table, Column, Integer, String, ForeignKey, Float
from sqlalchemy.ext.declarative import declarative_base
from sqlalchemy.orm import relation, backref

#_Debug = True
_Debug = False

BASE_PATH = os.path.dirname(os.path.abspath(__file__))
DATA_PATH = os.path.join(BASE_PATH, 'data')
if not os.path.exists(DATA_PATH):
    os.makedirs(DATA_PATH)
DB_FILE = os.path.join(DATA_PATH, 'data.db')


engine = create_engine('sqlite:///%s' % DB_FILE, echo=_Debug)
Base = declarative_base()
metadata = Base.metadata

class User(Base):

    __tablename__ = "users_table"

    id = Column('id', Integer, primary_key=True)
    name = Column('name', String)

    def __init__(self, name):
        self.name = name

    def __repr__(self):
        return "<User<' %s'>" % self.name


class Movie(Base):

    __tablename__ = "movies_table"

    id = Column(Integer, primary_key=True)
    name = Column(String, nullable=False)

    def __init__(self, name):
        self.name = name

    def __repr__(self):
        return "<Movie<' %s'>" % self.name

class Score(Base):

    __tablename__ = "scores_table"

    id = Column(Integer, primary_key=True)
    user_id = Column(Integer, ForeignKey('users_table.id'))
    movie_id = Column(Integer, ForeignKey('movies_table.id'))
    score = Column(Float)

    def __init__(self, user, movie, score):
        self.user = user
        self.movie = movie
        self.score = score

    def get_score(self):
        return self.score
        #return self.score /10.0

    def __repr__(self):
        return "<Score<' %s' ' %s' ' %s'>" % (
                self.user.name,
                self.movie.name,
                self.score)

Score.user = relation(User, backref=backref('scores', lazy='dynamic'))
Score.movie = relation(Movie, backref=backref('scores', lazy='dynamic'))
users_table = User.__tablename__
movies_table = Movie.__tablename__
score_table = Score.__tablename__

metadata.create_all(engine)

from sqlalchemy.orm import sessionmaker
Session = sessionmaker(bind=engine)
session = Session()


def initial():
    '''
    初期データ挿入
    '''
    users_list = [
        User(i) for i in [
            'Lisa Rose', 'Gene Seymour', 'Michael Phillips', 'Claudia Puig', 'Mick LaSalle',
            'Jack Matthews', 'Toby'
        ]
    ]

    movies_list = [
        Movie(i) for i in [
            'Lady in the Water', 'Snakes on a Plane', 'Just My Luck', 'Superman Returns',
            'You, Me and Dupree', 'The Night Listener'
        ]
    ]

    for u in users_list:
        session.add(u)

    for m in movies_list:
        session.add(m)

    score_list = [
        #La    Sn    Ju    Su    Yo    Th
        [2.5,  3.5,  3.0,  3.5,  2.5,  3.0], #Lisa
        [3.0,  3.5,  1.5,  5.0,  3.5,  3.0], #Gene
        [2.5,  3.0,  None, 3.5,  None, 4.0], #Michael
        [None, 3.5,  3.0,  4.0,  2.5,  4.5], #Claudia
        [3.0,  4.0,  2.0,  3.0,  2.0,  3.0], #Mick
        [3.0,  4.0,  None, 5.0,  3.5,  3.0], #jack
        [None, 4.5,  None, 4.0,  1.0, None] #Toby
    ]

    for u_id, u in enumerate(score_list):
        for m_id, m in enumerate(u):
            if m:
                session.add(Score(
                    user=users_list[u_id],
                    movie=movies_list[m_id],
                    score=m)
                )

    session.commit()

def hoge():
    '''
    ためし
    '''
    #Tobyさんの評価を取得
    u = session.query(User).filter_by(name="Toby").first()
    print u.scores.all()
    #スーパーマンの映画から評価を取得してTobyさんのを表示
    m = session.query(Movie).filter_by(name="Superman Returns").first()
    print m.scores.filter_by(user=u).all()
    #スーパーマンの映画の評価を取得
    for i in m.scores.all():
        print "user %s score %s" % (i.user.name, i.score)

def sim_distance(session, person1, person2):
    """
    person1とperson2の距離を元にした類似性スコアを返す
    """
    from sqlalchemy import or_
    p1 = session.query(User).filter_by(name=person1).first()
    p2 = session.query(User).filter_by(name=person2).first()
    movies = [i.movie for i in session.query(Score).filter(Score.user==p1) if
        i.movie in (j.movie for j in session.query(Score).\
                filter(Score.user==p2))]
    if movies == 0:
        return 0
    sum_of_squares = sum([pow(
        session.query(Score).filter(Score.user==p1).filter(Score.movie==m).\
                first().get_score() -
        session.query(Score).filter(Score.user==p2).filter(Score.movie==m).\
                first().get_score()
        , 2) for m in movies])
    return 1/(1+sum_of_squares)

def sim_person(session, person1, person2):
    """
    person1とperson2のピアソン相関係数を返す
    """
    p1 = session.query(User).filter_by(name=person1).first()
    p2 = session.query(User).filter_by(name=person2).first()
    movies = [i.movie for i in session.query(Score).filter(Score.user==p1) if
                i.movie in (j.movie for j in session.query(Score).\
                        filter(Score.user==p2))]

    count = len(movies)
    if count == 0:
        return 0

    sum1 = sum([s.get_score() for s in session.query(Score).filter(Score.user==p1)
        if s.movie in movies])
    sum2 = sum([s.get_score() for s in session.query(Score).filter(Score.user==p2)
        if s.movie in movies])

    sum1sq = sum([pow(s.get_score(), 2) for s in session.query(Score).filter(Score.user==p1) if s.movie in movies])
    sum2sq = sum([pow(s.get_score(), 2) for s in session.query(Score).filter(Score.user==p2) if s.movie in movies])

    psum = sum([
        session.query(Score).filter(Score.user==p1).filter(Score.movie==m).first().get_score() *
        session.query(Score).filter(Score.user==p2).filter(Score.movie==m).first().get_score()
        for m in movies])

    num = psum-(sum1*sum2/count)
    den = sqrt((sum1sq-pow(sum1, 2)/count)*(sum2sq-pow(sum2, 2)/count))
    if den == 0:
        return 0
    """
    print "num %s" % num
    print "den %s" % den
    print "sum1 %s" % sum1
    print "sum2 %s" % sum2
    print "sum1sq %s" % sum1sq
    print "sum2sq %s" % sum2sq
    print "psum %s" % psum
    """
    return num/den

def topMaches(session, person, n=5, similarity=sim_person):
    scores = [(similarity(session, person, other.name), other.name)
        for other in session.query(User).filter(User.name!=person)]
    scores.sort()
    scores.reverse()
    return scores[:n]


def getRecommendations(session, person, similarity=sim_person):
    from sqlalchemy.sql import exists

    totals = {}
    simSums = {}
    for other in session.query(User).filter(User.name!=person):
        sim = similarity(session, person, other.name)
        #print "sim %s" % sim
        if sim <= 0:
            continue

        for s in other.scores:
            #まだ評価していなものを計算
            if not session.query(Score).\
                    filter(Score.user==session.query(User).\
                    filter(User.name==person).first()).\
                    filter(Score.movie==s.movie).count():

                totals.setdefault(s.movie, 0)
                totals[s.movie] += s.get_score() * sim
                simSums.setdefault(s.movie, 0)
                simSums[s.movie] += sim
    rankings = [(total/simSums[movie], movie.name) for movie, total in totals.iteritems()]
    rankings.sort()
    rankings.reverse()
    return rankings

if __name__ == "__main__":
    #initial() #最初にDBを作ってデータを入れる関数、最初以外はコメントアウト
    #hoge()
    print sim_distance(session, "Lisa Rose", "Gene Seymour")
    print sim_person(session, "Lisa Rose", "Gene Seymour")
    print topMaches(session, 'Toby', n=3)
    print getRecommendations(session, 'Toby')
    print getRecommendations(session, 'Toby', sim_distance)

0.148148148148
0.396059017191
[(0.99124070716192991, u'Lisa Rose'), (0.92447345164190486, u'Mick LaSalle'), (0.89340514744156474, u'Claudia Puig')]
[(3.3477895267131017, u'The Night Listener'), (2.8325499182641614, u'Lady in the Water'), (2.5309807037655649, u'Just My Luck')]
[(3.5002478401415877, u'The Night Listener'), (2.7561242939959363, u'Lady in the Water'), (2.4619884860743739, u'Just My Luck')]

#! /usr/bin/env python
#! vim: fileencoding=utf8

import os
import sqlite3

from math import sqrt
import logging as logger

PATH = os.path.dirname(os.path.abspath(__file__))
DATA_PATH = os.path.join(PATH, 'data')
DB_FILE = os.path.join(DATA_PATH, 'data.db')

def sim_distance(cur, table, col1, col2, score_col, a_value, b_value):
    """
    a_valueとb_valueの距離を元にした類似性スコアを返す
    """
    sql = """
    SELECT a.%(score_col)s AS a_score, b.%(score_col)s AS b_score
    FROM %(table)s a, %(table)s b
    WHERE a.%(col1)s = %(a_value)s AND a.%(col2)s = b.%(col2)s
    AND b.%(col1)s = %(b_value)s""" % locals()
    logger.debug(sql)
    cur.execute(sql)
    sum_of_sequence = count = 0
    for a_s, b_s in cur:
        count += 1
        sum_of_sequence += pow(a_s * 1.0 - b_s * 1.0, 2)
    if count == 0:
        return 0
    return 1 / (1 + sum_of_sequence)

def sim_person(cur, table, col1, col2, score_col, a_value, b_value):
    """
    a_valueとb_valueのピアソン相関係数を返す
    """
    sql = """
    SELECT a.%(score_col)s AS a_score, b.%(score_col)s AS b_score
    FROM %(table)s a, %(table)s b
    WHERE a.%(col1)s = %(a_value)s AND a.%(col2)s = b.%(col2)s
    AND b.%(col1)s = %(b_value)s""" % locals()
    logger.debug(sql)
    cur.execute(sql)
    sum1 = sum2 = sum1sq = sum2sq = psum = count = 0
    for a_s, b_s in cur:
        count += 1
        sum1 += a_s * 1.0
        sum2 += b_s * 1.0
        sum1sq += pow(a_s * 1.0, 2)
        sum2sq += pow(b_s * 1.0, 2)
        psum += a_s * b_s * 1.0
    if count == 0:
        return 0
    num = psum - (sum1 * sum2 / count)
    den = sqrt((sum1sq-pow(sum1, 2)/count)*(sum2sq-pow(sum2, 2)/count))
    if den == 0:
        return 0
    """
    logger.debug("num %s" % num)
    logger.debug("den %s" % den)
    logger.debug("sum1 %s" % sum1)
    logger.debug("sum2 %s" % sum2)
    logger.debug("sum1sq %s" % sum1sq)
    logger.debug("sum2sq %s" % sum2sq)
    logger.debug("psum %s" % psum)
    """
    return num/den

def topMaches(cur, table, col1, col2, score_col, a_value, n=5, similarity=sim_person):
    """
    a_valueとそれ以外との関係性を計算してもっとも似ているもののリストを返す
    """
    # a_value 以外のcol1を抽出
    sql = "SELECT DISTINCT %(col1)s FROM %(table)s WHERE %(col1)s != %(a_value)s" % locals()
    cur.execute(sql)
    other_list = [i[0] for i in cur.fetchall()]
    scores = [(similarity(cur, table, col1, col2, score_col, a_value, b_value), b_value)
            for b_value in other_list]
    scores.sort()
    scores.reverse()
    return scores[:n]

def getRecommendations(cur, table, col1, col2, score_col, a_value, n=50, similarity=sim_person):
    """
    a_valueになく、他のものにある中で必要と思われるものをリストで返す
    """
    totals = {}
    simSums = {}
    sql = "SELECT DISTINCT %(col1)s FROM %(table)s WHERE %(col1)s != %(a_value)s" % locals()
    logger.debug(sql)
    cur.execute(sql)
    for other in (i[0] for i in cur.fetchall()):
        sim = similarity(cur, table, col1, col2, score_col, a_value, other)
        logger.debug("sim %s" % sim)
        if sim <= 0:
            continue
        #a_valueになくotherにあるcol2とscoreを抽出
        #MySQL4系統では副問合せを使えないのでLEFT JOINを使用
        sql = """
        SELECT a.%(col2)s, a.%(score_col)s FROM %(table)s a
            LEFT OUTER JOIN %(table)s b ON b.%(col1)s = %(a_value)s AND a.%(col2)s = b.%(col2)s
            WHERE a.%(col1)s = %(other)s AND b.%(score_col)s IS NULL""" % locals()
        logger.debug(sql)
        cur.execute(sql)

        for c2, score in cur:
            #logger.debug("%s %s" % (c2, score))
            totals.setdefault(c2, 0)
            totals[c2] += score * sim
            simSums.setdefault(c2, 0)
            simSums[c2] += sim
    rankings = [(total/simSums[c2], c2) for c2, total in totals.iteritems()]
    rankings.sort()
    rankings.reverse()
    return rankings[:n]


def main():
    con = sqlite3.connect(DB_FILE)
    cur = con.cursor()
    args = dict(
            cur=cur,
            table = "scores_table",
            col1 = "user_id",
            col2 = "movie_id",
            score_col = "score",
            a_value = 1,
            b_value = 2)
    print sim_distance(**args)
    print sim_person(**args)
    print topMaches(
            cur=cur,
            table = "scores_table",
            col1 = "user_id",
            col2 = "movie_id",
            score_col = "score",
            a_value = 7,
            n = 3)
    print getRecommendations(
            cur=cur,
            table = "scores_table",
            col1 = "user_id",
            col2 = "movie_id",
            score_col = "score",
            a_value = 7,
            n = 3)

    con.close()

if __name__ == "__main__":
   main()

0.148148148148
0.396059017191
[(0.99124070716192991, 1), (0.92447345164190486, 5), (0.89340514744156474, 4)]
[(3.3477895267131017, 6), (2.8325499182641614, 1), (2.5309807037655649, 3)]

import math
from itertools import ifilter, count

limit = 100

def sieve(limit=None, start=2):
    g = count(2)
    g = ifilter(lambda x, start=start: x >= start, g) #start以上の素数を返すFliter
    while True:
        prime = g.next()
        if limit and prime > limit: # limitを超えたらStopIteration
            raise StopIteration
        yield prime
        g = ifilter(lambda x, prime=prime: x % prime, g) # primeで割り切れるとFalse、割り切れないとTrueを返すlambdaをフィルターに追加

for i in sieve(int(math.sqrt(limit))): # limit が100なら 10がlimit
    for k in sieve(limit/i, i):
        print "%s * %s = %s" % (i, k, (i*k))

2 * 2 = 4
2 * 3 = 6
2 * 5 = 10
2 * 7 = 14
2 * 11 = 22
2 * 13 = 26
2 * 17 = 34
2 * 19 = 38
2 * 23 = 46
2 * 29 = 58
2 * 31 = 62
2 * 37 = 74
2 * 41 = 82
2 * 43 = 86
2 * 47 = 94
3 * 3 = 9
3 * 4 = 12
3 * 5 = 15
3 * 7 = 21
3 * 11 = 33
3 * 13 = 39
3 * 17 = 51
3 * 19 = 57
3 * 23 = 69
3 * 29 = 87
3 * 31 = 93
5 * 5 = 25
5 * 6 = 30
5 * 7 = 35
5 * 8 = 40
5 * 9 = 45
5 * 11 = 55
5 * 13 = 65
5 * 17 = 85
5 * 19 = 95
7 * 7 = 49
7 * 8 = 56
7 * 9 = 63
7 * 10 = 70
7 * 11 = 77
7 * 12 = 84
7 * 13 = 91

>>> import datetime
>>> import time
>>> datetime.datetime(*time.strptime(value, '%Y-%m-%d')[:3])
datetime.datetime(2008, 3, 10, 0, 0)

>>> value = '2008-03-10'
>>> datetime(*map((lambda x: int(x)),value.split('-')))
datetime.datetime(2008, 3, 10, 0, 0)

>>> datetime.datetime(*[int(i) for i in value.split('-')])
datetime.datetime(2008, 3, 10, 0, 0)

>>> def str_to_date(datestring, format="%Y-%m-%d"):
....     return datetime.date(*time.strptime(datestring, format)[:3])
....
>>> str_to_date('2008-03-10')
datetime.date(2008, 3, 10)
>>> str_to_date('080310', '%y%m%d')
datetime.date(2008, 3, 10)

def partial(func, *args, **keywords):
    def newfunc(*fargs, **fkeywords):
        newkeywords = keywords.copy()
        newkeywords.update(fkeywords)
        return func(*(args + fargs), **newkeywords)
    newfunc.func = func
    newfunc.args = args
    newfunc.keywords = keywords
    return newfunc

>>> basic_str_to_date = partial(str_to_date, format="%Y-%m-%d")
>>> custom_str_to_date = partial(str_to_date, format="%y%m%d")
>>> basic_str_to_date('2008-10-01')
datetime.date(2008, 10, 1)
>>> basic_str_to_date('2008-11-01')
datetime.date(2008, 11, 1)
>>> custom_str_to_date("081001")
datetime.date(2008, 10, 1)
>>> custom_str_to_date("081101")
datetime.date(2008, 11, 1)

>>> b = "ﾅｲﾄｳあいう".decode("utf8").encode("sjis") #僕のipythonの内部コードがutf8のためutf8でdecode
>>> b
'ﾅｲﾄｳ\x82\xa0\x82\xa2\x82\xa4'
>>> print b
ﾅｲﾄｳあいう
>>> nkf.guess(b)
'Shift_JIS'

>>> a = "ﾅｲﾄｳ".decode("utf8").encode("sjis")
>>> nkf.guess(a)
'EUC-JP'

>>> print nkf.nkf("-sx", b)
ﾅｲﾄｳあいう
>>> print nkf.nkf("-sx", a)
轍蝶

>>> print nkf.nkf("-sxS", a)
ﾅｲﾄｳ