Skip to content

Instantly share code, notes, and snippets.

@minrk
Created September 1, 2013 21:20
Show Gist options
  • Star 0 You must be signed in to star a gist
  • Fork 0 You must be signed in to fork a gist
  • Save minrk/6407393 to your computer and use it in GitHub Desktop.
Save minrk/6407393 to your computer and use it in GitHub Desktop.
"""Simple server for testing code against a slow-responding server.
requests to '/INTEGER' are replied to with 'INTEGER',
after an artificial delay.
"""
import time
from tornado import web, ioloop, options, httpserver
options.define("port", default=8888, help="run on the given port", type=int)
options.define("delay", default=0.2, help="delay builtin to each request", type=float)
class SlowHandler(web.RequestHandler):
@web.asynchronous
def get(self, page):
loop = ioloop.IOLoop.instance()
loop.add_timeout(loop.time() + self.settings['delay'],
lambda : self.finish(page),
)
if __name__ == '__main__':
options.parse_command_line()
app = web.Application([
(r'/(\d+)', SlowHandler)
], delay=options.options.delay,
)
http_server = httpserver.HTTPServer(app)
http_server.listen(options.options.port)
ioloop.IOLoop.instance().start()
Display the source blob
Display the rendered blob
Raw
{
"metadata": {
"name": ""
},
"nbformat": 3,
"nbformat_minor": 0,
"worksheets": [
{
"cells": [
{
"cell_type": "heading",
"level": 1,
"metadata": {},
"source": [
"Using ThreadPool for concurrent downloads"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"This is a simple illustration of using `ThreadPool` to parallelize downloads.\n",
"Assumes that bandwidth is not the limiting factor, in which case concurrency doesn't help."
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"import requests\n",
"from multiprocessing.pool import ThreadPool"
],
"language": "python",
"metadata": {},
"outputs": [],
"prompt_number": 1
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Test a simple request to my slow server\n",
"It just replies to any request for `/NUMBER` with the number requested,\n",
"but the server is artificially slow in its handling of requests."
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"%time r = requests.get(\"http://localhost:8888/10\")\n",
"r.content"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "stream",
"stream": "stdout",
"text": [
"CPU times: user 18.1 ms, sys: 4.54 ms, total: 22.6 ms\n",
"Wall time: 224 ms\n"
]
},
{
"metadata": {},
"output_type": "pyout",
"prompt_number": 2,
"text": [
"'10'"
]
}
],
"prompt_number": 2
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Our test function downloads the URL for a given ID,\n",
"and parses the result (casts str of int to int)."
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"def get_data(ID):\n",
" \"\"\"function for getting data from our slow server\"\"\"\n",
" r = requests.get(\"http://localhost:8888/%i\" % ID)\n",
" return int(r.content)"
],
"language": "python",
"metadata": {},
"outputs": [],
"prompt_number": 3
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Now test using a threadpool to get the data, using a varying number of concurrent threads"
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"IDs = range(128)\n",
"for nthreads in [1, 2, 4, 8, 16, 32]:\n",
" pool = ThreadPool(nthreads)\n",
" tic = time.time()\n",
" result = pool.map(get_data, IDs)\n",
" toc = time.time()\n",
" print \"%i threads: %3.1f seconds\" % (nthreads, toc-tic)"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "stream",
"stream": "stdout",
"text": [
"1 threads: 26.2 seconds\n",
"2 threads: 13.3 seconds"
]
},
{
"output_type": "stream",
"stream": "stdout",
"text": [
"\n",
"4 threads: 6.7 seconds"
]
},
{
"output_type": "stream",
"stream": "stdout",
"text": [
"\n",
"8 threads: 3.4 seconds"
]
},
{
"output_type": "stream",
"stream": "stdout",
"text": [
"\n",
"16 threads: 1.8 seconds"
]
},
{
"output_type": "stream",
"stream": "stdout",
"text": [
"\n",
"32 threads: 1.1 seconds"
]
},
{
"output_type": "stream",
"stream": "stdout",
"text": [
"\n"
]
}
],
"prompt_number": 4
}
],
"metadata": {}
}
]
}
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment