eyepea / api-hour Goto Github PK

Just kicking the tyres on your project...which seems really interesting btw.

To try things out I ran the all_in_one.py script in the examples directory but was getting errors from gunicorn

....
File "/Users/leon/envs/midi-sequencer/lib/python3.4/site-packages/gunicorn/config.py", line 58, in __getattr__
raise AttributeError("No configuration setting for: %s" % name)      
AttributeError: No configuration setting for: debug

Providing a default for Gunicorn's debug setting seems to fix things for me.
This is what I added
leonh@259e8ee

Is there any performance advantage to using this over just normal aiohttp + gunicorn? Maybe the readme could provide benchmarks comparing it to a normal aiohttp stack as well

https://github.com/Eyepea/API-Hour/blob/master/benchmarks/flask/application.py

Is there any reason specifically why we are using psycopg directly instead of using something like Flask-SQLAlchemy (http://pythonhosted.org/Flask-SQLAlchemy/)?

It's pretty much the standard for interacting with PostgreSQL and Flask-SQLAlchemy is almost always the preferred method when in the wild. I can whip one up really quickly if we want to add another benchmark into the mix.

Spent 2 days trying to run api_hour with project updated to aiohttp==3.0.9

I'm running it with command:
/usr/local/bin/api_hour --chdir /path/to/project server:Container -b 127.0.0.7:8000 --log-level=DEBUG

And it starts great but when I try to connect I have: "no data received" or "connection reset by peer" message. No logs in terminal, no errors. It seems like api-hour even can't handle request.

The same config works great with aiohttp==2.2.5 but now fails.

probably using uvloop you can get much better performance, its interesting how this compare to flask/gunicorn/uwsgi.

Hi.

Apologize if I open an issue but I cannot post on google groups (don't know why)

I would like to use motor with api-hour, but I can't understand how to implement it.

I'm stucked here:
@asyncio.coroutine
def start(self):
yield from super().start()
LOG.info('Starting engines...')
'''
# Add your custom engines here, example with PostgreSQL:
self.engines['pg'] = self.loop.create_task(aiopg.create_pool(host=self.config['engines']['pg']['host'],
port=int(self.config['engines']['pg']['port']),
sslmode='disable',
dbname=self.config['engines']['pg']['dbname'],
user=self.config['engines']['pg']['user'],
password=self.config['engines']['pg']['password'],
cursor_factory=psycopg2.extras.RealDictCursor,
minsize=int(self.config['engines']['pg']['minsize']),
maxsize=int(self.config['engines']['pg']['maxsize']),
loop=self.loop))
'''
self.engines['mongo'] = motor_asyncio.AsyncIOMotorClient()
# yield from asyncio.wait([self.engines['pg']], return_when=asyncio.ALL_COMPLETED)
yield from asyncio.wait([self.engines['mongo']], return_when=asyncio.ALL_COMPLETED)

I can't understand how to initialize motor.

Many thanks,
giovanni