utils.py

LINUX下

    res = []
    for (key, value) in dictionary.iteritems():
        if element is value:
            res.append(key)
    return res

def dictincr(dictionary, element):
    """
    Increments `element` in `dictionary`, 
    setting it to one if it doesn't exist.
    
        >>> d = {1:2, 3:4}
        >>> dictincr(d, 1)
        3
        >>> d[1]
        3
        >>> dictincr(d, 5)
        1
        >>> d[5]
        1
    """
    dictionary.setdefault(element, 0)
    dictionary[element] += 1
    return dictionary[element]

def dictadd(*dicts):
    """
    Returns a dictionary consisting of the keys in the argument dictionaries.
    If they share a key, the value from the last argument is used.
    
        >>> dictadd({1: 0, 2: 0}, {2: 1, 3: 1})
        {1: 0, 2: 1, 3: 1}
    """
    result = {}
    for dct in dicts:
        result.update(dct)
    return result

def listget(lst, ind, default=None):
    """
    Returns `lst[ind]` if it exists, `default` otherwise.
    
        >>> listget(['a'], 0)
        'a'
        >>> listget(['a'], 1)
        >>> listget(['a'], 1, 'b')
        'b'
    """
    if len(lst)-1 < ind: 
        return default
    return lst[ind]

def intget(integer, default=None):
    """
    Returns `integer` as an int or `default` if it can't.
    
        >>> intget('3')
        3
        >>> intget('3a')
        >>> intget('3a', 0)
        0
    """
    try:
        return int(integer)
    except (TypeError, ValueError):
        return default

def datestr(then, now=None):
    """
    Converts a (UTC) datetime object to a nice string representation.
    
        >>> from datetime import datetime, timedelta
        >>> d = datetime(1970, 5, 1)
        >>> datestr(d, now=d)
        '0 microseconds ago'
        >>> for t, v in {
        ...   timedelta(microseconds=1): '1 microsecond ago',
        ...   timedelta(microseconds=2): '2 microseconds ago',
        ...   -timedelta(microseconds=1): '1 microsecond from now',
        ...   -timedelta(microseconds=2): '2 microseconds from now',
        ...   timedelta(microseconds=2000): '2 milliseconds ago',
        ...   timedelta(seconds=2): '2 seconds ago',
        ...   timedelta(seconds=2*60): '2 minutes ago',
        ...   timedelta(seconds=2*60*60): '2 hours ago',
        ...   timedelta(days=2): '2 days ago',
        ... }.iteritems():
        ...     assert datestr(d, now=d+t) == v
        >>> datestr(datetime(1970, 1, 1), now=d)
        'January  1'
        >>> datestr(datetime(1969, 1, 1), now=d)
        'January  1, 1969'
        >>> datestr(datetime(1970, 6, 1), now=d)
        'June  1, 1970'
    """
    def agohence(n, what, divisor=None):
        if divisor: n = n // divisor

        out = str(abs(n)) + ' ' + what       # '2 day'
        if abs(n) != 1: out += 's'           # '2 days'
        out += ' '                           # '2 days '
        if n < 0:
            out += 'from now'
        else:
            out += 'ago'
        return out                           # '2 days ago'

    oneday = 24 * 60 * 60

    if not now: now = datetime.datetime.utcnow()
    if type(now).__name__ == "DateTime":
        now = datetime.datetime.fromtimestamp(now)
    if type(then).__name__ == "DateTime":
        then = datetime.datetime.fromtimestamp(then)
    delta = now - then
    deltaseconds = int(delta.days * oneday + delta.seconds + delta.microseconds * 1e-06)
    deltadays = abs(deltaseconds) // oneday
    if deltaseconds < 0: deltadays *= -1 # fix for oddity of floor

    if deltadays:
        if abs(deltadays) < 4:
            return agohence(deltadays, 'day')

        out = then.strftime('%B %e') # e.g. 'June 13'
        if then.year != now.year or deltadays < 0:
            out += ', %s' % then.year
        return out

    if int(deltaseconds):
        if abs(deltaseconds) > (60 * 60):
            return agohence(deltaseconds, 'hour', 60 * 60)
        elif abs(deltaseconds) > 60:
            return agohence(deltaseconds, 'minute', 60)
        else:
            return agohence(deltaseconds, 'second')

    deltamicroseconds = delta.microseconds
    if delta.days: deltamicroseconds = int(delta.microseconds - 1e6) # datetime oddity
    if abs(deltamicroseconds) > 1000:
        return agohence(deltamicroseconds, 'millisecond', 1000)

    return agohence(deltamicroseconds, 'microsecond')

def numify(string):
    """
    Removes all non-digit characters from `string`.
    
        >>> numify('800-555-1212')
        '8005551212'
        >>> numify('800.555.1212')
        '8005551212'
    
    """
    return ''.join([c for c in str(string) if c.isdigit()])

def denumify(string, pattern):
    """
    Formats `string` according to `pattern`, where the letter X gets replaced
    by characters from `string`.
    
        >>> denumify("8005551212", "(XXX) XXX-XXXX")
        '[phone]'
    
    """
    out = []
    for c in pattern:
        if c == "X":
            out.append(string[0])
            string = string[1:]
        else:
            out.append(c)
    return ''.join(out)

def dateify(datestring):
    """
    Formats a numified `datestring` properly.
    """
    return denumify(datestring, "XXXX-XX-XX XX:XX:XX")

class CaptureStdout:
    """
    Captures everything `func` prints to stdout and returns it instead.
    
        >>> def idiot():
        ...     print "foo"
        >>> capturestdout(idiot)()
        'foo\\n'
    
    **WARNING:** Not threadsafe!
    """
    def __init__(self, func): 
        self.func = func
    def __call__(self, *args, **keywords):
        from cStringIO import StringIO
        # Not threadsafe!
        out = StringIO()
        oldstdout = sys.stdout
        sys.stdout = out
        try: 
            self.func(*args, **keywords)
        finally: 
            sys.stdout = oldstdout
        return out.getvalue()

capturestdout = CaptureStdout

class Profile:
    """
    Profiles `func` and returns a tuple containing its output
    and a string with human-readable profiling information.
        
        >>> import time
        >>> out, inf = profile(time.sleep)(.001)
        >>> out
        >>> inf[:10].strip()
        'took 0.0'
    """
    def __init__(self, func): 
        self.func = func
    def __call__(self, *args): ##, **kw):   kw unused
        import hotshot, hotshot.stats, tempfile ##, time already imported
        temp = tempfile.NamedTemporaryFile()
        prof = hotshot.Profile(temp.name)

        stime = time.time()
        result = prof.runcall(self.func, *args)
        stime = time.time() - stime

        prof.close()
        stats = hotshot.stats.load(temp.name)
        stats.strip_dirs()
        stats.sort_stats('time', 'calls')
        x =  '\n\ntook '+ str(stime) + ' seconds\n'
        x += capturestdout(stats.print_stats)(40)
        x += capturestdout(stats.print_callers)()
        return result, x

profile = Profile


import traceback
# hack for compatibility with Python 2.3:
if not hasattr(traceback, 'format_exc'):
    from cStringIO import StringIO
    def format_exc(limit=None):
        strbuf = StringIO()
        traceback.print_exc(limit, strbuf)
        return strbuf.getvalue()
    traceback.format_exc = format_exc

def tryall(context, prefix=None):
    """
    Tries a series of functions and prints their results. 
    `context` is a dictionary mapping names to values; 
    the value will only be tried if it's callable.
    
        >>> tryall(dict(j=lambda: True))
        j: True
        ----------------------------------------
        results:
           True: 1

    For example, you might have a file `test/stuff.py` 
    with a series of functions testing various things in it. 
    At the bottom, have a line:

        if __name__ == "__main__": tryall(globals())

    Then you can run `python test/stuff.py` and get the results of 
    all the tests.
    """
    context = context.copy() # vars() would update
    results = {}
    for (key, value) in context.iteritems():
        if not hasattr(value, '__call__'): 
            continue
        if prefix and not key.startswith(prefix): 
            continue
        print key + ':',
        try:
            r = value()
            dictincr(results, r)
            print r
        except:
            print 'ERROR'
            dictincr(results, 'ERROR')
            print '   ' + '\n   '.join(traceback.format_exc().split('\n'))
        
    print '-'*40
    print 'results:'
    for (key, value) in results.iteritems():
        print ' '*2, str(key)+':', value

class ThreadedDict:
    """
    Takes a dictionary that maps threads to objects. 
    When a thread tries to get or set an attribute or item 
    of the threadeddict, it passes it on to the object 
    for that thread in dictionary.
    """
    def __init__(self, dictionary): 
        self.__dict__['_ThreadedDict__d'] = dictionary

    def __getattr__(self, attr): 
        return getattr(self.__d[threading.currentThread()], attr)

    def __getitem__(self, item): 
        return self.__d[threading.currentThread()][item]

    def __setattr__(self, attr, value):
        if attr == '__doc__':
            self.__dict__[attr] = value
        else:
            return setattr(self.__d[threading.currentThread()], attr, value)

    def __delattr__(self, item):
        try:
            del self.__d[threading.currentThread()][item]
        except KeyError, k:
            raise AttributeError, k

    def __delitem__(self, item):
        del self.__d[threading.currentThread()][item]

    def __setitem__(self, item, value): 
        self.__d[threading.currentThread()][item] = value

    def __hash__(self): 
        return hash(self.__d[threading.currentThread()])

threadeddict = ThreadedDict

def autoassign(self, locals):
    """
    Automatically assigns local variables to `self`.
    
        >>> self = storage()
        >>> autoassign(self, dict(a=1, b=2))
        >>> self
        <Storage {'a': 1, 'b': 2}>
    
    Generally used in `__init__` methods, as in:

        def __init__(self, foo, bar, baz=1): autoassign(self, locals())
    """
    for (key, value) in locals.iteritems():
        if key == 'self': 
            continue
        setattr(self, key, value)

def to36(q):
    """
    Converts an integer to base 36 (a useful scheme for human-sayable IDs).
    
        >>> to36(35)
        'z'
        >>> to36(119292)
        '2k1o'
        >>> int(to36(939387374), 36)
        939387374
        >>> to36(0)
        '0'
        >>> to36(-393)
        Traceback (most recent call last):
            ... 
        ValueError: must supply a positive integer
    
    """
    if q < 0: raise ValueError, "must supply a positive integer"
    letters = "0123456789abcdefghijklmnopqrstuvwxyz"
    converted = []
    while q != 0:
        q, r = divmod(q, 36)
        converted.insert(0, letters[r])
    return "".join(converted) or '0'


r_url = re_compile('(?<!\()(http://(\S+))')
def safemarkdown(text):
    """
    Converts text to HTML following the rules of Markdown, but blocking any
    outside HTML input, so that only the things supported by Markdown
    can be used. Also converts raw URLs to links.

    (requires [markdown.py](http://webpy.org/markdown.py))
    """
    from markdown import markdown
    if text:
        text = text.replace('<', '&lt;')
        # TODO: automatically get page title?
        text = r_url.sub(r'<\1>', text)
        text = markdown(text)
        return text


if __name__ == "__main__":
    import doctest
    doctest.testmod()