We Keep Coding

Python asyncio in a thread for migrating existing codebase

We have a rather big project that is doing a lot of networking (API calls, Websocket messages) and that also has a lot of internal jobs running in intervals in threads. Our current architecture involves spawning a lot of threads and the app is not working very well when the system is under a big load, so we've decided to give asyncio a try.
I know that the best way would be to migrate the whole codebase to async code, but that is not realistic in the very near future because of the size of the codebase and the limited development resources. However, we would like to start migrating parts of our codebase to use asyncio event loop and hopefully, we will be able to convert the whole project at some point.
The problem we have encountered so far is that the whole codebase has sync code and in order to add non-blocking asyncio code inside, the code needs to be run in different thread since you can't really run async and sync code in the same thread.
In order to combine async and sync code, I came up with this approach of running the asyncio code in a separate thread that is created on app start. Other parts of the code add jobs to this loop simply by calling add_asyncio_task.
import threading
import asyncio
_tasks = []
def threaded_loop(loop):
asyncio.set_event_loop(loop)
global _tasks
while True:
if len(_tasks) > 0:
# create a copy of needed tasks
needed_tasks = _tasks.copy()
# flush current tasks so that next tasks can be easily added
_tasks = []
# run tasks
task_group = asyncio.gather(*needed_tasks)
loop.run_until_complete(task_group)
def add_asyncio_task(task):
_tasks.append(task)
def start_asyncio_loop():
loop = asyncio.get_event_loop()
t = threading.Thread(target=threaded_loop, args=(loop,))
t.start()
and somewhere in app.py:
start_asyncio_loop()
and anywhere else in the code:
add_asyncio_task(some_coroutine)
Since I am new to asyncio, I am wondering if this is a good approach in our situation or if this approach is considered an anti-pattern and has some problems that will hit us later down the road? Or maybe asyncio already has some solution for this and I'm just trying to invent the wheel here?
Thanks for your inputs!

The approach is fine in general. You have some issues though:
(1) Almost all asyncio objects are not thread safe
(2) Your code is not thread safe on its own. What if a task appears after needed_tasks = _tasks.copy() but before _tasks = []? You need a lock here. Btw making a copy is pointless. Simple needed_tasks = _tasks will do.
(3) Some asyncio constructs are thread safe. Use them:
import threading
import asyncio
# asyncio.get_event_loop() creates a new loop per thread. Keep
# a single reference to the main loop. You can even try
# _loop = asyncio.new_event_loop()
_loop = asyncio.get_event_loop()
def get_app_loop():
return _loop
def asyncio_thread():
loop = get_app_loop()
asyncio.set_event_loop(loop)
loop.run_forever()
def add_asyncio_task(task):
asyncio.run_coroutine_threadsafe(task, get_app_loop())
def start_asyncio_loop():
t = threading.Thread(target=asyncio_thread)
t.start()

How do you understand the ioloop in tornado?

I am looking for a way to understand ioloop in tornado, since I read the official doc several times, but can't understand it. Specifically, why it exists.
from tornado.concurrent import Future
from tornado.httpclient import AsyncHTTPClient
from tornado.ioloop import IOLoop
def async_fetch_future():
http_client = AsyncHTTPClient()
future = Future()
fetch_future = http_client.fetch(
"http://mock.kite.com/text")
fetch_future.add_done_callback(
lambda f: future.set_result(f.result()))
return future
response = IOLoop.current().run_sync(async_fetch_future)
# why get current IO of this thread? display IO, hard drive IO, or network IO?
print response.body
I know what is IO, input and output, e.g. read a hard drive, display graph on the screen, get keyboard input.
by definition, IOLoop.current() returns the current io loop of this thread.
There are many IO device on my laptop running this python code. Which IO does this IOLoop.current() return? I never heard of IO loop in javascript nodejs.
Furthermore, why do I care this low level thing if I just want to do a database query, read a file?

I never heard of IO loop in javascript nodejs.
In node.js, the equivalent concept is the event loop. The node event loop is mostly invisible because all programs use it - it's what's running in between your callbacks.
In Python, most programs don't use an event loop, so when you want one, you have to run it yourself. This can be a Tornado IOLoop, a Twisted Reactor, or an asyncio event loop (all of these are specific types of event loops).
Tornado's IOLoop is perhaps confusingly named - it doesn't do any IO directly. Instead, it coordinates all the different IO (mainly network IO) that may be happening in the program. It may help you to think of it as an "event loop" or "callback runner".

Rather to say it is IOLoop, maybe EventLoop is clearer for you to understand.
IOLoop.current() doesn't really return an IO device but just a pure python event loop which is basically the same as asyncio.get_event_loop() or the underlying event loop in nodejs.
The reason why you need event loop to just do a database query is that you are using event-driven structure to do databse query(In your example, you are doing http request).
Most of time you do not need to care about this low level structure. Instead you just need to use async&await keywords.
Let's say there is a lib which supports asynchronous database access:
async def get_user(user_id):
user = await async_cursor.execute("select * from user where user_id = %s" % user_id)
return user
Then you just need to use this function in your handler:
class YourHandler(tornado.web.RequestHandler):
async def get():
user = await get_user(self.get_cookie("user_id"))
if user is None:
return self.finish("No such user")
return self.finish("Your are %s" % user.user_name)

Python Requests: Don't wait for request to finish

In Bash, it is possible to execute a command in the background by appending &. How can I do it in Python?
while True:
data = raw_input('Enter something: ')
requests.post(url, data=data) # Don't wait for it to finish.
print('Sending POST request...') # This should appear immediately.

Here's a hacky way to do it:
try:
requests.get("http://127.0.0.1:8000/test/",timeout=0.0000000001)
except requests.exceptions.ReadTimeout:
pass
Edit: for those of you that observed that this will not await a response - that is my understanding of the question "fire and forget... do not wait for it to finish". There are much more thorough and complete ways to do it with threads or async if you need response context, error handling, etc.

I use multiprocessing.dummy.Pool. I create a singleton thread pool at the module level, and then use pool.apply_async(requests.get, [params]) to launch the task.
This command gives me a future, which I can add to a list with other futures indefinitely until I'd like to collect all or some of the results.
multiprocessing.dummy.Pool is, against all logic and reason, a THREAD pool and not a process pool.
Example (works in both Python 2 and 3, as long as requests is installed):
from multiprocessing.dummy import Pool
import requests
pool = Pool(10) # Creates a pool with ten threads; more threads = more concurrency.
# "pool" is a module attribute; you can be sure there will only
# be one of them in your application
# as modules are cached after initialization.
if __name__ == '__main__':
futures = []
for x in range(10):
futures.append(pool.apply_async(requests.get, ['http://example.com/']))
# futures is now a list of 10 futures.
for future in futures:
print(future.get()) # For each future, wait until the request is
# finished and then print the response object.
The requests will be executed concurrently, so running all ten of these requests should take no longer than the longest one. This strategy will only use one CPU core, but that shouldn't be an issue because almost all of the time will be spent waiting for I/O.

Elegant solution from Andrew Gorcester. In addition, without using futures, it is possible to use the callback and error_callback attributes (see
doc) in order to perform asynchronous processing:
def on_success(r: Response):
if r.status_code == 200:
print(f'Post succeed: {r}')
else:
print(f'Post failed: {r}')
def on_error(ex: Exception):
print(f'Post requests failed: {ex}')
pool.apply_async(requests.post, args=['http://server.host'], kwargs={'json': {'key':'value'},
callback=on_success, error_callback=on_error))

According to the doc, you should move to another library :
Blocking Or Non-Blocking?
With the default Transport Adapter in place, Requests does not provide
any kind of non-blocking IO. The Response.content property will block
until the entire response has been downloaded. If you require more
granularity, the streaming features of the library (see Streaming
Requests) allow you to retrieve smaller quantities of the response at
a time. However, these calls will still block.
If you are concerned about the use of blocking IO, there are lots of
projects out there that combine Requests with one of Python’s
asynchronicity frameworks.
Two excellent examples are
grequests and
requests-futures.

Simplest and Most Pythonic Solution using threading
A Simple way to go ahead and send POST/GET or to execute any other function without waiting for it to finish is using the built-in Python Module threading.
import threading
import requests
def send_req():
requests.get("http://127.0.0.1:8000/test/")
for x in range(100):
threading.Thread(target=send_req).start() # start's a new thread and continues.
Other Important Features of threading
You can turn these threads into daemons using thread_obj.daemon = True
You can go ahead and wait for one to complete executing and then continue using thread_obj.join()
You can check if a thread is alive using thread_obj.is_alive() bool: True/False
You can even check the active thread count as well by threading.active_count()
Official Documentation

If you can write the code to be executed separately in a separate python program, here is a possible solution based on subprocessing.
Otherwise you may find useful this question and related answer: the trick is to use the threading library to start a separate thread that will execute the separated task.
A caveat with both approach could be the number of items (that's to say the number of threads) you have to manage. If the items in parent are too many, you may consider halting every batch of items till at least some threads have finished, but I think this kind of management is non-trivial.
For more sophisticated approach you can use an actor based approach, I have not used this library myself but I think it could help in that case.

from multiprocessing.dummy import Pool
import requests
pool = Pool()
def on_success(r):
print('Post succeed')
def on_error(ex):
print('Post requests failed')
def call_api(url, data, headers):
requests.post(url=url, data=data, headers=headers)
def pool_processing_create(url, data, headers):
pool.apply_async(call_api, args=[url, data, headers],
callback=on_success, error_callback=on_error)

What kind of problems (if any) would there be combining asyncio with multiprocessing?

As almost everyone is aware when they first look at threading in Python, there is the GIL that makes life miserable for people who actually want to do processing in parallel - or at least give it a chance.
I am currently looking at implementing something like the Reactor pattern. Effectively I want to listen for incoming socket connections on one thread-like, and when someone tries to connect, accept that connection and pass it along to another thread-like for processing.
I'm not (yet) sure what kind of load I might be facing. I know there is currently setup a 2MB cap on incoming messages. Theoretically we could get thousands per second (though I don't know if practically we've seen anything like that). The amount of time spent processing a message isn't terribly important, though obviously quicker would be better.
I was looking into the Reactor pattern, and developed a small example using the multiprocessing library that (at least in testing) seems to work just fine. However, now/soon we'll have the asyncio library available, which would handle the event loop for me.
Is there anything that could bite me by combining asyncio and multiprocessing?

You should be able to safely combine asyncio and multiprocessing without too much trouble, though you shouldn't be using multiprocessing directly. The cardinal sin of asyncio (and any other event-loop based asynchronous framework) is blocking the event loop. If you try to use multiprocessing directly, any time you block to wait for a child process, you're going to block the event loop. Obviously, this is bad.
The simplest way to avoid this is to use BaseEventLoop.run_in_executor to execute a function in a concurrent.futures.ProcessPoolExecutor. ProcessPoolExecutor is a process pool implemented using multiprocessing.Process, but asyncio has built-in support for executing a function in it without blocking the event loop. Here's a simple example:
import time
import asyncio
from concurrent.futures import ProcessPoolExecutor
def blocking_func(x):
time.sleep(x) # Pretend this is expensive calculations
return x * 5
#asyncio.coroutine
def main():
#pool = multiprocessing.Pool()
#out = pool.apply(blocking_func, args=(10,)) # This blocks the event loop.
executor = ProcessPoolExecutor()
out = yield from loop.run_in_executor(executor, blocking_func, 10) # This does not
print(out)
if __name__ == "__main__":
loop = asyncio.get_event_loop()
loop.run_until_complete(main())
For the majority of cases, this is function alone is good enough. If you find yourself needing other constructs from multiprocessing, like Queue, Event, Manager, etc., there is a third-party library called aioprocessing (full disclosure: I wrote it), that provides asyncio-compatible versions of all the multiprocessing data structures. Here's an example demoing that:
import time
import asyncio
import aioprocessing
import multiprocessing
def func(queue, event, lock, items):
with lock:
event.set()
for item in items:
time.sleep(3)
queue.put(item+5)
queue.close()
#asyncio.coroutine
def example(queue, event, lock):
l = [1,2,3,4,5]
p = aioprocessing.AioProcess(target=func, args=(queue, event, lock, l))
p.start()
while True:
result = yield from queue.coro_get()
if result is None:
break
print("Got result {}".format(result))
yield from p.coro_join()
#asyncio.coroutine
def example2(queue, event, lock):
yield from event.coro_wait()
with (yield from lock):
yield from queue.coro_put(78)
yield from queue.coro_put(None) # Shut down the worker
if __name__ == "__main__":
loop = asyncio.get_event_loop()
queue = aioprocessing.AioQueue()
lock = aioprocessing.AioLock()
event = aioprocessing.AioEvent()
tasks = [
asyncio.async(example(queue, event, lock)),
asyncio.async(example2(queue, event, lock)),
]
loop.run_until_complete(asyncio.wait(tasks))
loop.close()

Yes, there are quite a few bits that may (or may not) bite you.
When you run something like asyncio it expects to run on one thread or process. This does not (by itself) work with parallel processing. You somehow have to distribute the work while leaving the IO operations (specifically those on sockets) in a single thread/process.
While your idea to hand off individual connections to a different handler process is nice, it is hard to implement. The first obstacle is that you need a way to pull the connection out of asyncio without closing it. The next obstacle is that you cannot simply send a file descriptor to a different process unless you use platform-specific (probably Linux) code from a C-extension.
Note that the multiprocessing module is known to create a number of threads for communication. Most of the time when you use communication structures (such as Queues), a thread is spawned. Unfortunately those threads are not completely invisible. For instance they can fail to tear down cleanly (when you intend to terminate your program), but depending on their number the resource usage may be noticeable on its own.
If you really intend to handle individual connections in individual processes, I suggest to examine different approaches. For instance you can put a socket into listen mode and then simultaneously accept connections from multiple worker processes in parallel. Once a worker is finished processing a request, it can go accept the next connection, so you still use less resources than forking a process for each connection. Spamassassin and Apache (mpm prefork) can use this worker model for instance. It might end up easier and more robust depending on your use case. Specifically you can make your workers die after serving a configured number of requests and be respawned by a master process thereby eliminating much of the negative effects of memory leaks.

Based on #dano's answer above I wrote this function to replace places where I used to use multiprocess pool + map.
def asyncio_friendly_multiproc_map(fn: Callable, l: list):
"""
This is designed to replace the use of this pattern:
with multiprocessing.Pool(5) as p:
results = p.map(analyze_day, list_of_days)
By letting caller drop in replace:
asyncio_friendly_multiproc_map(analyze_day, list_of_days)
"""
tasks = []
with ProcessPoolExecutor(5) as executor:
for e in l:
tasks.append(asyncio.get_event_loop().run_in_executor(executor, fn, e))
res = asyncio.get_event_loop().run_until_complete(asyncio.gather(*tasks))
return res

See PEP 3156, in particular the section on Thread interaction:
http://www.python.org/dev/peps/pep-3156/#thread-interaction
This documents clearly the new asyncio methods you might use, including run_in_executor(). Note that the Executor is defined in concurrent.futures, I suggest you also have a look there.

Assistance with Python multithreading

Currently, i have a list of url to grab contents from and is doing it serially. I would like to change it to grabbing them in parallel. This is a psuedocode. I will like to ask is the design sound? I understand that .start() starts the thread, however, my database is not updated. Do i need to use q.get() ? thanks
import threading
import Queue
q = Queue.Queue()
def do_database(url):
""" grab url then input to database """
webdata = grab_url(url)
try:
insert_data_into_database(webdata)
except:
....
else:
< do I need to do anything with the queue after each db operation is done?>
def put_queue(q, url ):
q.put( do_database(url) )
for myfiles in currentdir:
url = myfiles + some_other_string
t=threading.Thread(target=put_queue,args=(q,url))
t.daemon=True
t.start()

It's odd that you're putting stuff into q but never taking anything out of q. What is the purpose of q? In addition, since do_database() doesn't return anything, sure looks like the only thing q.put(do_database(url)) does is put None into q.
The usual way these things work, a description of work to do is added to a queue, and then a fixed number of threads take turns pulling things off the queue. You probably don't want to create an unbounded number of threads ;-)
Here's a pretty complete - but untested - sketch:
import threading
import Queue
NUM_THREADS = 5 # whatever
q = Queue.Queue()
END_OF_DATA = object() # a unique object
class Worker(threading.Thread):
def run(self):
while True:
url = q.get()
if url is END_OF_DATA:
break
webdata = grab_url(url)
try:
# Does your database support concurrent updates
# from multiple threads? If not, need to put
# this in a "with some_global_mutex:" block.
insert_data_into_database(webdata)
except:
#....
threads = [Worker() for _ in range(NUM_THREADS)]
for t in threads:
t.start()
for myfiles in currentdir:
url = myfiles + some_other_string
q.put(url)
# Give each thread an END_OF_DATA marker.
for _ in range(NUM_THREADS):
q.put(END_OF_DATA)
# Shut down cleanly. `daemon` is way overused.
for t in threads:
t.join()

You should do this with asynchronous programming rather than threads. Threading in Python is problematic (see: Global Interpreter Lock), and anyway you're not trying to achieve multicore performance here. You just need a way to multiplex potentially long-running I/O. For that you can use a single thread and an event-driven library such as Twisted.
Twisted comes with HTTP functionality, so you can issue many concurrent requests and react (by populating your database) when results come in. Be aware that this model of programming may take a little getting used to, but it will give you good performance if the number of requests you're making is not astronomical (i.e. if you can get it all done on one machine, which it seems is your intention).

For DB, You have to commit before your changes become effective. But, commit for every insert is not optimal. Commit after bulk changes gives much better performance.
For parallel, Python isn't born for this. For your use-case, i suppose using python with gevent would be a painless solution.
Here is a much more efficient pseudo implementation FYI:
import gevent
from gevent.monkey import patch_all
patch_all() # to use with urllib, etc
from gevent.queue import Queue
def web_worker(q, url):
grab_something
q.push(result)
def db_worker(q):
buf = []
while True:
buf.append(q.get())
if len(buf) > 20:
insert_stuff_in_buf_to_db
db_commit
buf = []
def run(urls):
q = Queue()
gevent.spawn(db_worker, q)
for url in urls:
gevent.spawn(web_worker, q, url)
run(urls)
plus, since this implementation is totally single threaded, you can safely manipulate shared data between workers like queue, db connection, global variables etc.