Build a graph to process streams
Emit is a Python library for realtime data processing. It can distribute work with Celery or RQ, coordinate execution in other languages, and let you swing from the trees of your graph with Tarzan-like precision.
from emit import router router = Router() @router.node(['word'], entry_point=True) def parse_document(msg): for word in msg.document.strip().split(' '): yield word @router.node(['word', 'count'], 'parse_document') def count_word(msg): return msg.word, redis.zincrby('word_counts', msg.word, 1) import random document = 'the words in this document will be counted and emitted by count_word'.split(' ') router(document=' '.join(random.choice(document) for i in range(20)))
So how do you get it? On PyPI!
pip install emit
- CPython 2.6
- CPython 2.7
- CPython 3.2
- CPython 3.3 (except for Celery, currently)
- PyPy 1.9
Installing from source/development branch:
Emit’s releases are fairly frequent, so you should be good to use a released version. However, if you need something currently not in a version, try the develop branch:
pip install git+git://github.com/BrianHicks/emit.git@develop
Just do be warned that things will break. And be sure to check Travis to see if the platform you want is currently passing. It should be, but maybe not.