mirai
=====
`mirai` is a port of [Twitter Futures][1] to python powered by a backport of
Python 3.2's [concurrent.futures][2]. `mirai` offers a simpler
operator chaining-based way for using promises and futures.
[1]: http://twitter.github.io/scala_school/finagle.html#futconcurrent
[2]: https://docs.python.org/dev/library/concurrent.futures.html
### Installation
`mirai` is available on [pypi](http://pypi.python.org/pypi/mirai). Installation
is as simple as using `pip`,
```bash
$ pip install mirai --upgrade
```
### Diving In
`mirai` is based around the concept of a Promises -- containers that
are populated with the result of an asynchronous computation. Promises provide
a more streamlined way for dealing with asynchronous code than callbacks.
```python
from mirai import Promise
import funcy as fu
(
Promise
# `call` is a way to execute a synchronous function asynchronously. It
# returns a Promise you can use to queue up callbacks
.call(twitter.get_followers, "duck")
# `map` means "create another Promise by applying the following function to
# my contents." If the previous Promise threw an exception, this callback
# won't execute, and the resulting Promise will contain the previous
# Promise's exception.
.map(lambda followers: [
Future.call(twitter.get_latest_posts, follow)
for follower in followers
])
# `Promise.collect` turns a sequence of promises into a Promise with a
# sequence of items. If you have many sources for Promises, you can use
# `Promise.collect` to bundle them together. Notice that we're using
# `Promise.flatmap` here -- that's because `Promise.collect` returns a
# Promise, not a normal value. If we didn't use flatmap, we'd end up with a
# Promise containing another Promise!
.flatmap(Promise.collect)
# Each follower gave me a list of tweets, so rather than dealing a list of
# lists, lets just turn it into a flat list of tweets
.map(fu.flatten)
# `Promise.onsuccess` and `Promise.onfailure` allow you to attach a callback
# that doesn't actually affect the Promise's value. Great for logging!
.onsuccess(lambda tweets: print("Retrieved {:,d} Tweets from my followers"))
# Notice how we turned what could have been a synchronous double-for loop
# into a series of asynchronous calls? Pretty awesome!
.map(lambda tweets: [
Future.call(twitter.retweet, tweet)
for tweet in tweets
])
.flatmap(Future.collect)
.onsuccess(lambda retweets: print("I love EVERYBODY!"))
# You can use `Promise.rescue` to turn a failing Promise into a successfuly
# one, too.
.rescue(lambda e: Future.value("Something went wrong...oh well"))
# All of the above isn't executed until you request it's data, so don't
# forget to call `Promise.get`! You can also use `Promise.join` and
# `Promise.select` if you want to gather many Promises or just want the first
# that finishes.
.get()
)
```
### Documentation
Learn more about `mirai` at [http://mirai.readthedocs.org/](http://mirai.readthedocs.org/)
mirai-develop.zip （29个子文件） 
mirai-develop 
setup.py 1KB .gitignore 47B
requirements.txt 140B Makefile 577B .gitmodules 105B README.md 3KB mirai __init__.py 218B pool.py 3KB futures.py 28KB utils.py 840B tests test_futures.py 17KB test_pool.py 1KB __init__.py 0B _version.py 31B exceptions.py 3KB docs .gitignore 9B why.rst 6KB Makefile 7KB api.rst 1KB index.rst 2KB conf.py 9KB caveats.rst 3KB _tutorial tutorial04.py 1007B scraper.py 3KB commons.py 1KB tutorial03.py 533B tutorial05.py 1KB tutorial02.py 347B _themes tutorial.rst 3KB
评论3
最新资源