So you are storing some key-values in a dict but your data became huge than your memory or you want to persist it on the disk? Then mongodict is for you!
As it uses MongoDB to store the data, you get all cool MongoDB things, like shardings and replicas. It uses the pickle module available on Python standard library to serialize/deserialize data and store everything as bson.Binary in MongoDB. You can also provide another codec (serializer/deserializer).
mongodict is tested under Python 2.7.5 and Python 3.3.2.
As simple as:
pip install mongodictAs it uses collections.MutableMapping as its base, you just need to change the line which creates your dict. For instace, just replace:
>>> my_dict = {}with:
>>> from mongodict import MongoDict
>>> my_dict = MongoDict(host='localhost', port=27017, database='my_dict',
collection='store')and then use it like a normal dict:
>>> my_dict['python'] = 'rules'
>>> print my_dict['python']
rules
>>> del my_dict['python']
>>> print my_dict['python']
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
File "mongodict.py", line 82, in __getitem__
raise KeyError(key)
KeyError: u'python'
>>> my_dict['spam'] = 'eggs'
>>> my_dict['ham'] = 'damn'
>>> for key, value in my_dict.items():
... print '{} = {}'.format(key, value)
...
spam = eggs
ham = damnIf you want to use another codec, you should pass serialize and deserialize functions to the class during the initialization. For example, to use JSON:
>>> import json
>>> json_dict = MongoDict(host='localhost', port=27017,
database='json_dict', collection='store',
codec=(json.dumps, json.loads))
>>> # use json_dict as usualEnjoy! :-)
Note
There is no kind of in-memory cache, so all key lookups will be translated in a MongoDB query but as MongoDB's server put everything it can in memory, probably it'll not be a problem (if your working set is always entire in memory).
If you want to use MongoDB's authentication to the database MongoDict is connecting to, you just need to provide an auth parameter, as in this example:
from mongodict import MongoDict
my_dict = MongoDict(host='localhost', port=27017, database='mydb',
collection='mongodict',
auth=('my username', 'my password'))Redis is "remote directory server" - it's a great piece of software and can do the job if all your data fit on memory. By other side, MongoDB already have mature sharding and replica set features. So, if you need to store lots of key-value pairs that don't fit on memory, mongodict can solve your problem.
Note
mongodict does not have the same API other key-value software have (like memcached). Some features are missing to compete directly with these kind of software (by now we have only the dict-like behaviour), but I have plans to add it soon.
- This software uses Semantic Versioning.
- This software uses Test-Driven Development.
You can run the tests either with or without tox.
This is the simplest approach: you'll test only for one Python version. To do it, just execute:
mkvirtualenv --no-site-packages mongodict-without-tox
pip install -r requirements/develop.txt
make testWith tox you can test for more than one Python version (currently for 2.7 and 3.2). You just need to create a virtualenv, install and run it:
mkvirtualenv --no-site-packages tox-for-mongodict
pip install tox
toxtox will create one virtualenv for each Python version, install requirements and then run the tests for each of them. Note that you need the python binaries available in your system (2.7 and 3.2) to run the tests.
This software was written and is maintained by Álvaro Justen (aka Turicas). Please contact me at alvarojusten at gmail dot com.
It's licensed under GPL version 3.
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent