skip to navigation
skip to content

quick-server 0.4.6

QuickServer is a quick to use and easy to set up server implementation.

QuickServer is a quick to use and easy to set up server implementation. It has the following goals / features and is primarily meant to speed up back end implementation / iteration:

  • serve local files as is with basic black- and white-listing
  • provide functionality for dynamic requests
  • provide easy access to worker threads (and caching)
  • provide a basic command interpret loop for server commands


You can install quick_server with pip:

pip install --user quick-server

Import it in python via:

from quick_server import create_server, msg, setup_restart

Setting up a basic file server

Following we will set up a basic quick_server. Please refer to the inline documentation of the methods for full information.

setup_restart() # sets up restart functionality (if not called the `restart` command of the server needs external help to work)
# should be the first real executed command in the script
# some services, like heroku, don't play well with this command and it should not be called if in such an environment

addr = '' # empty address is equivalent to 'localhost'
port = 8080
server = create_server((addr, port), parallel=True) # parallel is recommended unless your code is not thread-safe
server.bind_path('/', 'www') # binds the 'www' directory to the server's root
server.add_default_white_list() # adds typical file types to the list of files that will be served; you can use server.add_file_patterns to add more file types
server.favicon_fallback = 'favicon.ico' # sets the default favicon file to the given file on disk (you'll need a file called 'favicon.ico')
# you can also use server.link_empty_favicon_fallback()

server.suppress_noise = True # don't report successful requests (turn off if you want to measure performance)
server.report_slow_requests = True # reports requests that take longer than 5s

Starting the actual server:

msg("{0}", " ".join(sys.argv)) # prints how the script was started
msg("starting server at {0}:{1}", addr if addr else 'localhost', port)
server.serve_forever() # starts the server -- only returns when the server stops (e.g., by typing `quit`, `restart`, or `CTRL-C`)
msg("shutting down..")
server.server_close() # make sure to clean up all resources

Adding dynamic requests

Dynamic requests can be set up by annotating a function. The annotation consists of return-type and http-method.

A POST request in JSON format:

@server.json_post('/json_request', 0) # creates a request at http://localhost:8080/json_request
def json_request(req, args):
    return {
        "post": args["post"],

A GET request as plain text:

@server.text_get('/text_request', 0) # creates a request at http://localhost:8080/text_request
def text_request(req, args):
    return "plain text"

Worker threads and caching

Worker threads require support from the client side.

First, provide the necessary JavaScript file via


and load it on the client side:

<script src="js/worker.js" charset="utf-8"></script>

A worker request can be set up on the server side with

def json_worker(args):
    # ...
    # long, slow computation
    return myresult # myresult must be JSON convertible

and accessed from the client. An instance of the Worker class is needed:

var work = new quick_server.Worker();
work.status((req) => {
  // req contains the number of currently active requests (-1 indicates an error state)
  // it can be used to tell the user that something is happening

Accessing the worker:

// the first argument identifies worker jobs
// jobs with the same name get replaced when a new one has been started
// the second argument is the URL"worker_name", "json_worker", {
  // this object will appear as args on the server side
}, (data) => {
  // data is the result of the worker function of the server side
  // this function is only called if the request was successful

A worker can be cancelled using its name:


Note that all running workers are cancelled when the page is unloaded.

Workers can automatically cache the server response using quick_cache. The server needs to be set up for this:

cache = QuickCache(base_file, quota=500, ram_quota=100, warnings=msg)
server.cache = cache

Then caching can be used for workers:

@server.json_worker('/json_worker', cache_id=lambda args: {
        # uniquely identify the task from its arguments (must be JSON convertible)
def json_worker(args):
    # ...
    # long, slow computation
    return myresult # myresult must be JSON convertible

Custom server commands

By default quick_server provides the commands help (list of available commands), restart (restart the server), and quit (terminates the server). You can add own commands via

def name(args): # creates the command name
    if not args:
        msg("hi {0}", " ".join(args)) # words typed after name are printed here

A common command to add when having caching functionality (e.g., provided by quick_cache) is to clear caches. This show-cases also auto-complete functionality:

def complete_cache_clear(args, text): # args contains already completed arguments; text the currently started one
    if args: # we only allow up to one argument
        return []
    return [ section for section in cache.list_sections() if section.startswith(text) ] # cache is the quick_cache object

def cache_clear(args):
    if len(args) > 1: # we only allow up to one argument
      msg("too many extra arguments! expected one got {0}", ' '.join(args))
    msg("clear {0}cache{1}{2}", "" if args else "all ", " " if args else "s", args[0] if args else "")
    cache.clean_cache(args[0] if args else None)

Server without command loop

The easiest way to start the server without a command loop (e.g., when started as service) is to stop the loop with an EOF by calling the script like this:

cat /dev/null | python

or use the no_command_loop flag and run the script normally.


You can wrap the server socket to support HTTPS:

import ssl

addr = '' # empty address is equivalent to 'localhost'
port = 443 # the HTTPS default port 443 might require root privileges
server = create_server((addr, port), parallel=True)
server.socket = ssl.wrap_socket(server.socket, certfile='path/to/localhost.pem', server_side=True)

# setup your server


More examples and also contain minimal example servers. You can run them with ./ and ./ respectively from the examples directory. Then you can browse to http://localhost:8000/example/.


Pull requests are highly appreciated :) Also, feel free to open issues for any questions or bugs you may encounter.

File Type Py Version Uploaded on Size
quick_server-0.4.6-py2.py3-none-any.whl (md5) Python Wheel py2.py3 2017-07-20 30KB
quick_server-0.4.6.tar.gz (md5) Source 2017-07-20 29KB