I have a CherryPy server running on a BeagleBone Black. Server generates a simple webpage and does local SPI reads / writes (hardware interface). The application is going to be used on a local network with 1-2 clients at a time.
I need to prevent a CherryPy class function being called twice, two or more instances before it completes.
Thoughts?
As saaj commented, a simple threading.Lock() will prevent the handler from being run at the same time by another client. I might also add, using cherrypy.session.acquire_lock() will prevent the same client from the running two handlers simultaneously.
Refreshing article on Python locks and stuff: http://effbot.org/zone/thread-synchronization.htm
Although I would make saaj's solution much simpler by using a "with" statement in Python, to hide all those fancy lock acquisitions/releases and try/except block.
lock = threading.Lock()
#cherrypy.expose
def index(self):
with lock:
# do stuff in the handler.
# this code will only be run by one client at a time
return '<html></html>'
It is general synchronization question, though CherryPy side has a subtlety. CherryPy is a threaded-server so it is sufficient to have an application level lock, e.g. threading.Lock.
The subtlety is that you can't see the run-or-fail behaviour from within a single browser because of pipelining, Keep-Alive or caching. Which one it is is hard to guess as the behaviour varies in Chromium and Firefox. As far as I can see CherryPy will try to serialize processing of request coming from single TCP connection, which effectively results in subsequent requests waiting for active request in a queue. With some trial-and-error I've found that adding cache-prevention token leads to the desired behaviour (even though Chromium still sends Connection: keep-alive for XHR where Firefox does not).
If run-or-fail in single browser isn't important to you you can safely ignore the previous paragraph and JavaScript code in the following example.
Update
The cause of request serialisation coming from one browser to the same URL doesn't lie in server-side. It's an implementation detail of a browser cache (details). Though, the solution of adding random query string parameter, nc, is correct.
#!/usr/bin/env python
# -*- coding: utf-8 -*-
import threading
import time
import cherrypy
config = {
'global' : {
'server.socket_host' : '127.0.0.1',
'server.socket_port' : 8080,
'server.thread_pool' : 8
}
}
class App:
lock = threading.Lock()
#cherrypy.expose
def index(self):
return '''<!DOCTYPE html>
<html>
<head>
<title>Lock demo</title>
<script type='text/javascript' src='http://cdnjs.cloudflare.com/ajax/libs/qooxdoo/3.5.1/q.min.js'></script>
<script type='text/javascript'>
function runTask(wait)
{
var url = (wait ? '/runOrWait' : '/runOrFail') + '?nc=' + Date.now();
var xhr = q.io.xhr(url);
xhr.on('loadend', function(xhr)
{
if(xhr.status == 200)
{
console.log('success', xhr.responseText)
}
else if(xhr.status == 503)
{
console.log('busy');
}
});
xhr.send();
}
q.ready(function()
{
q('p a').on('click', function(event)
{
event.preventDefault();
var wait = parseInt(q(event.getTarget()).getData('wait'));
runTask(wait);
});
});
</script>
</head>
<body>
<p><a href='#' data-wait='0'>Run or fail</a></p>
<p><a href='#' data-wait='1'>Run or wait</a></p>
</body>
</html>
'''
def calculate(self):
time.sleep(8)
return 'Long task result'
#cherrypy.expose
def runOrWait(self, **kwargs):
self.lock.acquire()
try:
return self.calculate()
finally:
self.lock.release()
#cherrypy.expose
def runOrFail(self, **kwargs):
locked = self.lock.acquire(False)
if not locked:
raise cherrypy.HTTPError(503, 'Task is already running')
else:
try:
return self.calculate()
finally:
self.lock.release()
if __name__ == '__main__':
cherrypy.quickstart(App(), '/', config)
Related
I have a view that generates data and streams it in real time. I can't figure out how to send this data to a variable that I can use in my HTML template. My current solution just outputs the data to a blank page as it arrives, which works, but I want to include it in a larger page with formatting. How do I update, format, and display the data as it is streamed to the page?
import flask
import time, math
app = flask.Flask(__name__)
#app.route('/')
def index():
def inner():
# simulate a long process to watch
for i in range(500):
j = math.sqrt(i)
time.sleep(1)
# this value should be inserted into an HTML template
yield str(i) + '<br/>\n'
return flask.Response(inner(), mimetype='text/html')
app.run(debug=True)
You can stream data in a response, but you can't dynamically update a template the way you describe. The template is rendered once on the server side, then sent to the client.
One solution is to use JavaScript to read the streamed response and output the data on the client side. Use XMLHttpRequest to make a request to the endpoint that will stream the data. Then periodically read from the stream until it's done.
This introduces complexity, but allows updating the page directly and gives complete control over what the output looks like. The following example demonstrates that by displaying both the current value and the log of all values.
This example assumes a very simple message format: a single line of data, followed by a newline. This can be as complex as needed, as long as there's a way to identify each message. For example, each loop could return a JSON object which the client decodes.
from math import sqrt
from time import sleep
from flask import Flask, render_template
app = Flask(__name__)
#app.route("/")
def index():
return render_template("index.html")
#app.route("/stream")
def stream():
def generate():
for i in range(500):
yield "{}\n".format(sqrt(i))
sleep(1)
return app.response_class(generate(), mimetype="text/plain")
<p>This is the latest output: <span id="latest"></span></p>
<p>This is all the output:</p>
<ul id="output"></ul>
<script>
var latest = document.getElementById('latest');
var output = document.getElementById('output');
var xhr = new XMLHttpRequest();
xhr.open('GET', '{{ url_for('stream') }}');
xhr.send();
var position = 0;
function handleNewData() {
// the response text include the entire response so far
// split the messages, then take the messages that haven't been handled yet
// position tracks how many messages have been handled
// messages end with a newline, so split will always show one extra empty message at the end
var messages = xhr.responseText.split('\n');
messages.slice(position, -1).forEach(function(value) {
latest.textContent = value; // update the latest value in place
// build and append a new item to a list to log all output
var item = document.createElement('li');
item.textContent = value;
output.appendChild(item);
});
position = messages.length - 1;
}
var timer;
timer = setInterval(function() {
// check the response for new data
handleNewData();
// stop checking once the response has ended
if (xhr.readyState == XMLHttpRequest.DONE) {
clearInterval(timer);
latest.textContent = 'Done';
}
}, 1000);
</script>
An <iframe> can be used to display streamed HTML output, but it has some downsides. The frame is a separate document, which increases resource usage. Since it's only displaying the streamed data, it might not be easy to style it like the rest of the page. It can only append data, so long output will render below the visible scroll area. It can't modify other parts of the page in response to each event.
index.html renders the page with a frame pointed at the stream endpoint. The frame has fairly small default dimensions, so you may want to to style it further. Use render_template_string, which knows to escape variables, to render the HTML for each item (or use render_template with a more complex template file). An initial line can be yielded to load CSS in the frame first.
from flask import render_template_string, stream_with_context
#app.route("/stream")
def stream():
#stream_with_context
def generate():
yield render_template_string('<link rel=stylesheet href="{{ url_for("static", filename="stream.css") }}">')
for i in range(500):
yield render_template_string("<p>{{ i }}: {{ s }}</p>\n", i=i, s=sqrt(i))
sleep(1)
return app.response_class(generate())
<p>This is all the output:</p>
<iframe src="{{ url_for("stream") }}"></iframe>
5 years late, but this actually can be done the way you were initially trying to do it, javascript is totally unnecessary (Edit: the author of the accepted answer added the iframe section after I wrote this). You just have to include embed the output as an <iframe>:
from flask import Flask, render_template, Response
import time, math
app = Flask(__name__)
#app.route('/content')
def content():
"""
Render the content a url different from index
"""
def inner():
# simulate a long process to watch
for i in range(500):
j = math.sqrt(i)
time.sleep(1)
# this value should be inserted into an HTML template
yield str(i) + '<br/>\n'
return Response(inner(), mimetype='text/html')
#app.route('/')
def index():
"""
Render a template at the index. The content will be embedded in this template
"""
return render_template('index.html.jinja')
app.run(debug=True)
Then the 'index.html.jinja' file will include an <iframe> with the content url as the src, which would something like:
<!doctype html>
<head>
<title>Title</title>
</head>
<body>
<div>
<iframe frameborder="0"
onresize="noresize"
style='background: transparent; width: 100%; height:100%;'
src="{{ url_for('content')}}">
</iframe>
</div>
</body>
When rendering user-provided data render_template_string() should be used to render the content to avoid injection attacks. However, I left this out of the example because it adds additional complexity, is outside the scope of the question, isn't relevant to the OP since he isn't streaming user-provided data, and won't be relevant for the vast majority of people seeing this post since streaming user-provided data is a far edge case that few if any people will ever have to do.
Originally I had a similar problem to the one posted here where a model is being trained and the update should be stationary and formatted in Html. The following answer is for future reference or people trying to solve the same problem and need inspiration.
A good solution to achieve this is to use an EventSource in Javascript, as described here. This listener can be started using a context variable, such as from a form or other source. The listener is stopped by sending a stop command. A sleep command is used for visualization without doing any real work in this example. Lastly, Html formatting can be achieved using Javascript DOM-Manipulation.
Flask Application
import flask
import time
app = flask.Flask(__name__)
#app.route('/learn')
def learn():
def update():
yield 'data: Prepare for learning\n\n'
# Preapre model
time.sleep(1.0)
for i in range(1, 101):
# Perform update
time.sleep(0.1)
yield f'data: {i}%\n\n'
yield 'data: close\n\n'
return flask.Response(update(), mimetype='text/event-stream')
#app.route('/', methods=['GET', 'POST'])
def index():
train_model = False
if flask.request.method == 'POST':
if 'train_model' in list(flask.request.form):
train_model = True
return flask.render_template('index.html', train_model=train_model)
app.run(threaded=True)
HTML Template
<form action="/" method="post">
<input name="train_model" type="submit" value="Train Model" />
</form>
<p id="learn_output"></p>
{% if train_model %}
<script>
var target_output = document.getElementById("learn_output");
var learn_update = new EventSource("/learn");
learn_update.onmessage = function (e) {
if (e.data == "close") {
learn_update.close();
} else {
target_output.innerHTML = "Status: " + e.data;
}
};
</script>
{% endif %}
I am setting up a web application using Flask and Flask-socketio for Websockets, using eventlet as recommended in the documentation.
The purpose is to create a coding platform where users can subscribe, join the current challenge and for each exercise can submit their own solution (a piece of code) which will then be assessed for correctness on the server via test cases (they write a function, given some input it must give a specific output, if it doesn't count zero score).
The issue is that some server-side computation in executing the uploaded code may take several seconds to complete before returning the output to the client. In this case, the client disconnects upon receiving the data from the server, after the long wait. You can see below some code which reproduces the issue, the time.sleep represents the long server side computation, which will cause the client to disconnect and reconnect.
Note: in the code I put 40 seconds of sleep, that is to have the client disconnect every time, for a smaller time (say between 10 and 20), sometimes it works fine and sometimes it disconnects.
Why is that happening? How can I fix it?
from flask import Flask
from flask_socketio import SocketIO
from flask_socketio import emit, disconnect
import time
import random
flask_app = Flask(__name__)
socketio = SocketIO(flask_app, async_mode='eventlet')
webpage = '''
<html>
<body>
<p id="demo">Some content</p>
<button type="button" onclick="do_on_server()">Do server-side computation</button>
<script src="//cdnjs.cloudflare.com/ajax/libs/socket.io/2.2.0/socket.io.js" integrity="sha256-yr4fRk/GU1ehYJPAs8P4JlTgu0Hdsp4ZKrx8bDEDC3I=" crossorigin="anonymous"></script>
<script type="text/javascript" charset="utf-8">
var socket = io();
function do_on_server(){
socket.emit('do_on_server', {});
}
socket.on('feedback', feedback);
function feedback(data){
document.getElementById("demo").innerHTML = data["msg"];
}
</script>
</body>
</html>
'''
#flask_app.route('/')
def index():
return webpage
#socketio.on('do_on_server')
def do_on_server(json):
print('starting computation')
#long computation
time.sleep(40)
print('done computing')
emit('feedback', {'msg': random.random()})
#socketio.on('connect')
def on_connect():
print('Connected')
#socketio.on('disconnect')
def on_disconnect():
print('Disconnecting')
if __name__ == '__main__':
socketio.run(flask_app)
Console output:
>python socket_timeout.py
Connected
starting computation
done computing
Disconnecting
Connected
I have the following Python script which is using Flask-socketio
from flask import Flask, render_template
from flask_socketio import SocketIO, emit
from time import sleep
app = Flask(__name__)
app.config['SECRET_KEY'] = 'P#ssw0rd'
socketio = SocketIO(app)
#app.route('/')
def index():
return render_template('index.html')
#socketio.on('connect')
def on_connect():
payload1 = 'Connected!!!'
payload2 = 'Doing thing 1'
payload3 = 'Doing thing 2'
emit('send_thing', payload1, broadcast=True)
sleep(2)
emit('send_thing', payload2, broadcast=True)
sleep(2)
emit('send_thing', payload3, broadcast=True)
if __name__ == '__main__':
socketio.run(app)
And here is the corresponding index.html
<!DOCTYPE html>
<html>
<head>
<meta charset="utf-8">
<title>SocketIO Python</title>
</head>
<body>
<div id="my-div"></div>
<script src="https://cdnjs.cloudflare.com/ajax/libs/socket.io/1.4.5/socket.io.js"></script>
<script>
(function init() {
var socket = io()
var divElement = document.getElementById('my-div')
socket.on('send_thing', function(payload) {
var dataElement = document.createElement('inner')
dataElement.innerHTML = payload
divElement.appendChild(dataElement)
})
})()
</script>
</body>
</html>
What I am trying to achieve is that when a client connects, it first says 'Connected!!!' and then 2 seconds later a new 'inner' element appears that says 'Doing thing 1' followed by 2 seconds later a new 'inner' element appears that says 'Doing thing 2' etc.
But what is happening is that when a client connects, it sends all 3 lines at the same time (after 4 seconds which is both sleep statements). This is the first time using SocketIO so I'm sure I've done something wrong.
When you use eventlet or gevent, the time.sleep() function is blocking, it does not allow any other tasks to run.
Three ways to address this problem:
Use socketio.sleep() instead of time.sleep().
Use eventlet.sleep() or gevent.sleep().
Monkey patch the Python standard library so that time.sleep() becomes async-friendly.
I'm new to asynchronous programming in python and I'm trying to write a script that starts a websocket server, listens for messages, and also sends messages when certain events (e.g. pressing the 's' key) are triggered in a gtk window. Here's what I have so far:
import gi
gi.require_version('Gtk', '3.0')
from gi.repository import Gtk
import asyncio
import websockets
import threading
ws = None
async def consumer_handler(websocket, path):
global ws
ws = websocket
await websocket.send("Hello client")
while True:
message = await websocket.recv()
print("Message from client: " + message)
def keypress(widget,event):
global ws
if event.keyval == 115 and ws: #s key pressed, connection open
asyncio.get_event_loop().create_task(ws.send("key press"))
print("Message to client: key press")
def quit(widget):
Gtk.main_quit()
window = Gtk.Window(Gtk.WindowType.TOPLEVEL)
window.connect("destroy", quit)
window.connect("key_press_event", keypress)
window.show()
start_server = websockets.serve(consumer_handler, 'localhost', 8765)
asyncio.get_event_loop().run_until_complete(start_server)
wst = threading.Thread(target=asyncio.get_event_loop().run_forever)
wst.daemon = True
wst.start()
Gtk.main()
And here's the client webpage:
<!DOCTYPE html>
<html>
<head>
<title>Websockets test page</title>
<meta charset="UTF-8" />
<script>
var exampleSocket = new WebSocket("ws://localhost:8765");
function mylog(msg) {
document.getElementById("log").innerHTML += msg + "<br/>";
}
function send() {
mylog("Message to server: Hello server");
exampleSocket.send("Hello server");
}
exampleSocket.onopen = function (event) {
mylog("Connection opened");
};
exampleSocket.onmessage = function (event) {
mylog("Message from server: " + event.data);
}
</script>
</head>
<body>
<p id="log"></p>
<input type="button" value="Send message" onclick="send()"/>
</body>
</html>
Run the python code, then load the webpage in a browser, now any messages the browser sends show up in the python stdout, so far so good. But if you hit the 's' key in the gtk window, python doesn't send the message until another message is received from the browser (from pressing the 'Send message' button). I thought that await websocket.recv() was meant to return control to the event loop until a message was received? How do I get it to send messages while it's waiting to receive?
But if you hit the 's' key in the gtk window, python doesn't send the message until another message is received from the browser
The problem is in this line:
asyncio.get_event_loop().create_task(ws.send("key press"))
Since the asyncio event loop and the GTK main loop are running in different threads, you need to use run_coroutine_threadsafe to submit the coroutine to asyncio. Something like:
asyncio.run_coroutine_threadsafe(ws.send("key press"), loop)
create_task adds the coroutine to the queue of runnable coroutines, but fails to wake up the event loop, which is why your coroutine is only run when something else happens in asyncio. Also, create_task is not thread-safe, so calling it while the event loop itself is modifying the run queue could corrupt its data structures. run_coroutine_threadsafe has neither of these problems, it arranges for the event loop to wake up as soon as possible, and it uses a mutex to protect the event loop's data structures.
I have a pyramid application running under apache using mod_wsgi.
I am planning to migrate from apache to cherrypy.
I am able to load static page of the existing web application with cherrypy. However for any AJAX request, I am getting resource not found (404) error.
Any clues??
Thanks
30-Mar-2016
Here is code structure
MyProject
|
cherry_wsgi.py (creates wsgi app object)
cherry_server.py (starts cherrypy server using app object from cherry_wsgi.py)
development.ini
myproject
|
__init__.py (Scans sub-folders recursively)
views.py
mydata
|
__init__.py
data
|
__init__.py (Added route for getdata)
views.py (implementation of getdata)
|
myclient
|
index.html (AJAX query)
Contents of myclient/index.html
<html>
<head>
<meta charset="utf-8">
<title>HOME UI</title>
</head>
<body>
<button id="submit">Give it now!</button>
<script src="./jquery-2.1.3.min.js"></script>
<script>$("#submit").on('click', function()
{
$.ajax(
{
type: "GET",
async: false,
url: "../myproject/data/getdata",
success: function (data)
{
console.log("LED On" );
},
error: function ()
{
console.error("ERROR");
}
});
});</script></body></html>
File myproject/__init__.py
from pyramid.config import Configurator
from pyramid.renderers import JSONP
import os
import logging
def includeme(config):
""" If include function exists, write this space.
"""
pass
def main(global_config, **settings):
""" This function returns a Pyramid WSGI application."""
config = Configurator(settings=settings)
config.add_renderer('jsonp', JSONP(param_name='callback'))
config.include(includeme)
directory = "/home/redmine/Downloads/MyProject/myproject/mydata/"
for root,dir,files in os.walk(directory):
if root == directory:# Walk will return all sublevels.
for dirs in dir: #This is a tuple so we need to parse it
config.include('myproject.mydata.' + str(dirs), route_prefix='/' + str(dirs))
config.add_static_view('static', 'prototype', cache_max_age=3600)
config.scan()
return config.make_wsgi_app()
File myproject/views.py
from pyramid.view import view_config
File myproject/mydata/__init__.py
import data
File mproject/mydata/data/__init__.py
from pyramid.config import Configurator
def includeme(config):
config.add_route('get_data', 'getdata', xhr=True)
def main(global_config, **settings):
print 'hello'
config = Configurator(settings=settings)
config.include(includeme, route_prefix='/data')
config.add_static_view('static', 'prototype', cache_max_age=3600)
config.scan('data')
return config.make_wsgi_app()
File mproject/mydata/data/views.py
from pyramid.view import view_config
import json
#view_config(route_name='get_data', xhr=True, renderer='jsonp')
def get_data(request):
return "{'firstName' : 'John'}"
File cherry_wsgi.py
from pyramid.config import Configurator
from pyramid.response import Response
from pyramid.paster import get_app
config = Configurator()
app = get_app('development.ini', 'main')
File cherry_server.py
from cherry_wsgi import app
import cherrypy
conf = {
'/': {
'tools.sessions.on': True,
'tools.staticdir.root': '/home/redmine/Downloads/MyProject/'
},
'/myclient': {
'tools.staticdir.on': True,
'tools.staticdir.dir': './myclient'
}
}
if __name__ == '__main__':
cherrypy.tree.mount(app, "/", conf)
cherrypy.server.unsubscribe()
server = cherrypy._cpserver.Server()
server.socket_host = "0.0.0.0"
server.socket_port = 9090
server.thread_pool = 30
server.subscribe()
cherrypy.engine.start()
cherrypy.engine.block()
I'm not sure if I caught everything but I did see a couple of bugs. First off, your url was off in your ajax call. Next in your views.py you were using jsonp and not json as your renderer. Also, you were using "route_name" instead of "route" (ajax calls) for the #view_config. Finally, you were returning a string. I changed it to a dict.
Pyramid can be tricky if you don't set up your project structure in a straight forward way. I learned the hard way :)
Contents of myclient/index.html
<html>
<head>
<meta charset="utf-8">
<title>HOME UI</title>
</head>
<body>
<button id="submit">Give it now!</button>
<script src="./jquery-2.1.3.min.js"></script>
<script>$("#submit").on('click', function()
{
$.ajax(
{
type: "GET",
async: false,
url: "getdata",
success: function (data)
{
console.log("LED On" );
},
error: function ()
{
console.error("ERROR");
}
});
});
</script>
</body>
</html>
File mproject/mydata/data/views.py
from pyramid.view import view_config
#view_config(name='get_data', renderer='json')
def get_data(request):
return {'firstName' : 'John'}
Now after looking at you overall file structure it does not look like a standard pyramid app. You have a lot of things going and it looks like over programming to me. There is a lot of duplicate code. Maybe you are doing this for a reason but I don't know.
I included below a pyramid start git repo. I built it to help people get started putting there pyramid projects on Openshift. I would think your pyramid project should follow the same outline. No need for deep folders.
The file you want to pay particular close attention to is the "app.py.disabled" file. Don't mind the disabled part. There are two ways to start an Openshift pyramid app and this git repo is using the wsgi.py file. You can just switch the two.
Anyway, inside the app.py.disabled file you can see all the different ways that I have used to setup pyramid app using wsgi servers (simple, waitress, and cherrypy). Just uncomment/comment out the code you want.
I think you are mixing the cherrypy framework and pyramid framework. Just use cherrypy's wsgi server. Don't do any of the cherrypy configuration. The last I heard was that Cherrypy was separating out their wsgi server from their framework. It's been at least a year since I looked.
You might just try using Waitress. Very good and simple and works across all platforms.
Openshift-Pyramidstarter