get(indexable, *indices|keys, [default])

You had a get() method on dict, but not on lists or tuple. Now you do

>>> lst = range(10)
>>> get(lst, 1, default='whatever_you_want')
1
>>> get(lst, 100, default='whatever_you_want')
'whatever_you_want'

Plus, you can chain look ups :

This:

try:
    res = data['key'][0]['other key'][1]
except (KeyError, IndexError):
    res = "value"

Becomes:

get(data, 'key', 0, 'other key', 1, default="value")

attrs(object, *attributes, [default])

And for attributes…

devise = attr(car, 'insurance', 'expiration_date', 'timezone')

iget(iterable, index, [default])

You can also get values at indices on any iterable, including generators :

>>> iget(xrange(10), 0)
0
>>> iget(xrange(10), 5)
5
>>> iget(xrange(10), 10000, default='wololo')
u'wololo'

chunks(iterable, size) and window(iterable, size)

Iteration by chunk or with a sliding window:

>>> l = range(10)
>>> for chunk in chunks(l, 3):
...     print list(chunk)
...
[0, 1, 2]
[3, 4, 5]
[6, 7, 8]
[9]
>>> for slide in window(l, 3):
...     print list(slide)
...
[0, 1, 2]
[1, 2, 3]
[2, 3, 4]
[3, 4, 5]
[4, 5, 6]
[5, 6, 7]
[6, 7, 8]
[7, 8, 9]

flatten(deeply_nested_iterable)

Returns a generator that lazily yield the items and deals with up to hundred of levels of nesting (~800 on my machine, and you can control it with sys.setrecursionlimit)

a = []
for i in xrange(10):
    a = [a, i]
print(a)

[[[[[[[[[[[], 0], 1], 2], 3], 4], 5], 6], 7], 8], 9]

print(list(flatten(a)))

[0, 1, 2, 3, 4, 5, 6, 7, 8, 9]

It works with most iterables, even of inifinite or unknown size.

dmerge(dict, dict, [merge_function])

I wish ‘+’’ was overloaded for dicts:

>>> dmerge({"a": 1, "b": 2}, {"b": 2, "c": 3})
{'a': 1, 'c': 3, 'b': 2}

Sometimes you do not want to simply overwrite the values inside the original dict, but merge them in custom fashion:

>>> def my_merge(v1, v2):
...     if isinstance(v1, dict) and isinstance(v2, dict):
...         return dmerge(v1, v2)
...     return v2
>>> dmerge({"a": 1, "b": {'ok': 5}}, {"b": {'ko': 5 }, "c": 3}, my_merge)
{'a': 1, 'c': 3, 'b': {'ko': 5, 'ok': 5}}

subdict(dict, [include], [exclude])

And for lazy people like me

>>> subdict({'a': 1, 'b': 2, 'c': 3}, include=('a', 'b'))
{'a': 1, 'b': 2}
>>> subdict({'a': 1, 'b': 2, 'c': 3}, exclude=('c',))
{'a': 1, 'b': 2}

Which is quite nice when you want a dict of some local variables (like in web framework functions returning responses such as Django, Flask or Bottle)

>>> def test():
...     a, b, c, d, e = range(5)
...     return subdict(locals(), exclude=('d',))
...
>>> test()
{'a': 0, 'c': 2, 'b': 1, 'e': 4}

This works with any indexable, not just dicts.

normalize(string)

>>> normalize(u"Hélo Whorde")
'Helo Whorde'

slugify(string)

>>> slugify(u"Hélo Whorde")
helo-whorde

You get better slugification if you install the unidecode lib, but it’s optional. You can specify separator if you don’t like - or call directly normalize() (the underlying function) if you wish more control.

json_dumps(struct) and json_loads(string)

JSON helpers that handle date/time

>>> import datetime
>>> json_dumps({'test': datetime.datetime(2000, 1, 1, 1, 1, 1)})
'{"test": "2000-01-01 01:01:01.000000"}'
>>> json_dumps({'test': datetime.date(2000, 1, 1)})
'{"test": "2000-01-01"}'
>>> json_dumps({'test': datetime.time(1, 1, 1)})
'{"test": "01:01:01.000000"}'
>>> json_dumps({'test': datetime.timedelta(1, 1)})
'{"test": "timedelta(seconds=\'86401.0\')"}'
>>> json_dumps({u'test': datetime.timedelta(1, 1), u'a': [1, 2]})
'{"test": "timedelta(seconds=\'86401.0\')", "a": [1, 2]}'

>>> json_loads('{"test": "2000-01-01 01:01:01.000000"}')
{u'test': datetime.datetime(2000, 1, 1, 1, 1, 1)}
>>> json_loads('{"test": "2000-01-01"}')
{u'test': datetime.date(2000, 1, 1)}
>>> json_loads('{"test": "01:01:01.000000"}')
{u'test': datetime.time(1, 1, 1)}
>>> json_loads('{"test": "timedelta(seconds=\'86401.0\')"}')
{u'test': datetime.timedelta(1, 1)}
>>> json_loads('{"test": "timedelta(seconds=\'86401.0\')", "a": [1, 2]}')
{u'test': datetime.timedelta(1, 1), u'a': [1, 2]}

write(path, *args, encoding=’utf8’, mode=’w’, errors=’replace’)

Write anything in one row to a file

>>> s = '/tmp/test'
>>> write(s, 'test', 'é', 1, ['fdjskl'])
>>> print open(s).read()
test
é
1
['fdjskl']

It will attempt decoding / encoding and casting automatically each value to a string.

This is an utility function : its slow and doesn’t consider edge cases, but allow to do just what you want most of the time in one line.

You can optionally pass :

• mode : among ‘a’, ‘w’, which default to ‘w’. Binary mode is forced.

• encoding : which default to utf8 and will condition decoding AND encoding

• errorswhat to do when en encoding error occurs‘replace’ by default,

  which replace faulty caracters with ‘?’

You can pass string or unicode as *args, but if you pass strings, make sure you pass them with the same encoding you wish to write to the file.