@krishmoodbidri this is a Flask question, how is Socket.IO related to this? If Flask does not see a remote user in your incoming requests, then this extension is going to also not have this information.

@Abbyyan without monkey patching it will not work. With monkey-patching I believe eventlet still has some issues/bugs that haven't been solved in the implementation of multiprocessing.

Gevent is exactly the same as eventlet with regards to blocking and patching. As I said a few times already, you need to learn how to work with these platforms. You are not using standard Python anymore, these are asynchronous frameworks. The rules are different.

@Abbyyan I really don't know how to help, this isn't a problem with this package, you need to work around the limitations eventlet imposes in your application. Monkey patching helps in most cases, but not always.

You can't use threading and queues with eventlet, they are incompatible. See https://eventlet.net/doc/patching.html for how to monkey patch them to become compatible.

Dear miguelgrinberg, When I have checked out 17f, and completed docker compose running, there was logs :

flasky_1 | INFO [alembic.runtime.migration] Context impl MySQLImpl. flasky_1 | INFO [alembic.runtime.migration] Will assume non-transactional DDL. flasky_1 | INFO [alembic.runtime.migration] Running upgrade -> 38c4e85512a9, initial migration flasky_1 | INFO [alembic.runtime.migration] Running upgrade 38c4e85512a9 -> 456a945560f6, login support flasky_1 | INFO [alembic.runtime.migration] Running upgrade 456a945560f6 -> 190163627111, account confirmation flasky_1 | INFO [alembic.runtime.migration] Running upgrade 190163627111 -> 56ed7d33de8d, user roles flasky_1 | INFO [alembic.runtime.migration] Running upgrade 56ed7d33de8d -> d66f086b258, user information flasky_1 | INFO [alembic.runtime.migration] Running upgrade d66f086b258 -> 198b0eebcf9, caching of avatar hashes flasky_1 | INFO [alembic.runtime.migration] Running upgrade 198b0eebcf9 -> 1b966e7f4b9e, post model flasky_1 | INFO [alembic.runtime.migration] Running upgrade 1b966e7f4b9e -> 288cd3dc5a8, rich text posts flasky_1 | INFO [alembic.runtime.migration] Running upgrade 288cd3dc5a8 -> 2356a38169ea, followers flasky_1 | INFO [alembic.runtime.migration] Running upgrade 2356a38169ea -> 51f5ccfba190, comments flasky_1 | [2020-09-21 13:43:55 +0000] [1] [INFO] Starting gunicorn 19.7.1 flasky_1 | [2020-09-21 13:43:55 +0000] [1] [INFO] Listening at: http://0.0.0.0:5000 (1) flasky_1 | [2020-09-21 13:43:55 +0000] [1] [INFO] Using worker: sync flasky_1 | [2020-09-21 13:43:55 +0000] [28] [INFO] Booting worker with pid: 28 flasky_1 | 172.18.0.1 - - [21/Sep/2020:13:55:59 +0000] "GET / HTTP/1.1" 200 3133 "-" "Mozilla/5.0 (Macintosh; Intel Mac OS X 10_14_6) AppleWebKit/605.1.15 (KHTML, like Gecko) Version/13.1.2 Safari/605.1.15" flasky_1 | 172.18.0.1 - - [21/Sep/2020:13:55:59 +0000] "GET /static/styles.css HTTP/1.1" 200 - "http://localhost:8000/" "Mozilla/5.0 (Macintosh; Intel Mac OS X 10_14_6) AppleWebKit/605.1.15 (KHTML, like Gecko) Version/13.1.2 Safari/605.1.15" flasky_1 | 172.18.0.1 - - [21/Sep/2020:13:56:00 +0000] "GET /static/favicon.ico HTTP/1.1" 200 - "http://localhost:8000/" "Mozilla/5.0 (Macintosh; Intel Mac OS X 10_14_6) AppleWebKit/605.1.15 (KHTML, like Gecko) Version/13.1.2 Safari/605.1.15"

Home page works. But I have not received confirm email after registering. I remembered I have input correct mail username and password. What's the matter?

@DavidLeibs I don't know what is it specifically that your Safari does not like. When I request the socket.io.js URL it comes back with the correct CORS header that allows all origins.

Do you get a different CORS header there?

@DavidLeibs Can you provide more details on this error? What is the attempted method? What URL? This information should be available from the developer console.

No, this extension works with Flask-SQLAlchemy only. Sorry.

You may want to run your application in debug mode, so that Flask-Mail generates logging.

How to force disconnect the client form server side in python?

Thanks!

Just as title says, add calpaterson's original repository link. By the way, thanks for your excellent work.

Yes, there are actually two ways. One is with the "external process" technique (docs: https://flask-socketio.readthedocs.io/en/latest/#emitting-from-an-external-process). The other way is by using a Socket.IO client and connecting like the browser would do (docs: https://python-socketio.readthedocs.io/en/latest/client.html)

Well, you seem to have a problem with some low-level networking code, not with this package, right? Why is this package part of the problem?

As for other ways to communicate, two that come to mind are sending HTTP requests, and using a Socket.IO client.

I'm not sure what you mean by "pipe" in this context. You are not showing all your code, but it seems you are just connecting the two processes with a low-level TCP connection? I'm guessing, because I don't know what your Client and Listener classes do.

Are you using eventlet or gevent in these applications? If yes, did you monkey patch the standard library so that socket operations become non-blocking?

I want to use yolo to call multiple cameras at the same time for identification. Please help me.

That has nothing to do with this project, sorry but I cannot help you with yolo, not my thing.

import upip upip.install('microdot') Installing to: /lib/ Warning: micropython.org SSL certificate is not validated Error installing 'microdot': syntax error in JSON, packages may be partially installed

The upip is not very robust. Sometimes it reads partial information from the package repository, making it think the data is invalid, other times it just runs out of memory. I suggest you just get the .py file(s) that you are interested in from this repository and copy them into your project.

i am middle of the project and i was connecting django with gmail and i got this type of error.....

settings.py code for :

EMAIL_BACKEND = 'django.core.mail.backends.smtp.EmailBackend' EMAIL_HOST = 'smtp.gmail.com' EMAIL_HOST_USER = os.environ.get('EMAIL_USER') EMAIL_HOST_PASSWORD = os.environ.get('EMAIL_PASS') EMAIL_PORT = 587 EMAIL_USE_TLS = True DEFAULT_FROM_EMAIL = EMAIL_HOST_USER

what should is in this !!!?

You have a problem with a Django project. This is a Flask project, you've come to the wrong place to ask your question.

Uninstall the socketio package. That's something else, nothing to do with this package which is called python-socketio.

This is not a notebook app, it is a standalone web server.

Run the app by calling socketio.run(app) in your main script instead of flask run.

Hi i need to add authorization for socket to avoid others to access to our socket. How colud i add this feature. can some one help me out.

This isn't a WebSocket client, you cannot connect to WS servers with this package. You can only connect to Socket.IO servers.

As far as concurrent connections you can use threads with the standard client, or async tasks with the asyncio client. Both approaches work.

@krishmoodbidri You've got it backwards. uWSGI has native support for WebSocket, gunicorn does not. When you use gunicorn the WS functionality comes from eventlet or gevent-websocket.

@kirankumarind of course, everything is possible, it's just software. You are not asking any specific questions, so I'm really confused about why you wrote an issue here.

I am not familiar with the Java client, so I really don't know. The only potential issue I can think if is that if the server does too much CPU work then it is going to become unresponsive under high load. Not sure if that is your case or not, but with asynchronous servers you always have to watch for CPU use.

I'm not familiar with elastic beanstalk, so I cannot help you with specific questions about that service. The log shows that the upgrade was carried, but you are using a sqlite database, that seems like strange choice for a cloud deployment. How would that work when the service is scaled to multiple instances? Would they all be able to see the same database file?

Yes, you can read about that in the docs: https://flask-socketio.readthedocs.io/en/latest/#emitting-from-an-external-process

The 0.9 tag points to a8800271547e5cd961817b3b4f347c9f863b0835, which does have EmptyForm. I think you are looking at an older version of this repository.

@harrypotter90 wouldn't it make more sense to fix the client to send pings at the correct interval? You can set the interval to anything you want when you create the SocketIO() object, but I'm not sure that is the problem.

On running it is showing connection failed.is the backend working?

I would think so. I don't really have experience with this issue, I just Googled the error and found that article.

@NamanChuriwala Thanks, now I understand. Please review this blog post, which explains your problem: https://virtualandy.wordpress.com/2019/09/04/a-fix-for-operationalerror-psycopg2-operationalerror-ssl-error-decryption-failed-or-bad-record-mac/.

@kirilgeorgiev82 yes, if you can post your working code here when you have it that would be awesome, as others will find it when they face the same problem. Thanks!

You can call db.session.remove() when you are done with your session. I'm not sure what you mean regarding concurrent requests in Celery. First of all, Celery runs jobs, not requests, and second, it runs one job at a time in each worker process. So how can there be concurrent database accesses?

Did you read the SQLAlchemy docs? There is a page specifically about this issue: https://docs.sqlalchemy.org/en/13/orm/extensions/mutable.html.

What is the type of the dictionary attribute in your model?

I recorded a short video in which I go through all the steps in installing and running this application: https://www.youtube.com/watch?v=5hdrW6yEwro.

Hope this helps clarify the process.

I am not sure whether I receive the actual object

There is no concept of objects in the database. SQLAlchemy creates an object representation of the data stored in your database, so in a way all model instances are proxies that represent the real data.

What would be the purpose of doing these strange operations with copy on attributes of the model?

This isn't really enough information for me to tell you what the problem is. The problem is likely not in this code, but instead in the configuration or set up of the application or the database.

The error is (sqlite3.OperationalError) no such table: posts. You did not create/upgrade your database.

Creating a test request context doesn't help, a test context does not have information about the client.

What you need to do is pass all the arguments into the function, so that there is no need to fill in the gaps with information from the context. So in addition to the room name, pass the sid of the user in question and the namespace. For the sid you may need to pass it as an argument into your thread function.

So your function would be something like the following:

def foo(room_id, sid):
      with app.app_context():
          leave_handler(room_id, sid=sid, namespace='/')

I'm not familiar with this service, so I can't really tell you. The logs clearly show that the endpoint does not speak Socket.IO, it just accepts GET requests. Not sure if this is because you configured a WebSocket endpoint instead of a Socket.IO, but that could be one problem.

How did you set the /liveticker endpoint in the server?

Overall I think your idea is correct. You have to check the logs of your application to see the stack trace corresponding to the Internal Server Error. That's the only way to move forward.

Your mistake is thinking in terms of a main function. There is no such thing in an asynchronous system, everything is event-driven. Your main function should just start the connection and then wait.

client = socketio.AsyncClient(
    reconnection=False,
    logger=True,
    engineio_logger=True,
)

async def main():
    url = 'ws://localhost:8080'
    await client.connect(url, socketio_path='internal/ws', transports=['websocket'], namespaces=['/qwe'])
    await client.wait()

Then if you need to do something immediately after connection, start a task in the connect handler:

async def my_logic():
    resp = await client.call('foo', {'data': 'some data'}, namespace='/qwe', timeout=5)
    print('Resp is ', resp)

@client.event(namespace='/qwe')
async def connect(sid, environ):
    client.start_background_task(my_logic)

What do you think about adding an optional flag to connect() which will indicate if we need to wait all connections to be established?

It would be awfully complicated, and there are weird situations that would be hard to handle. For example, if you connect to two namespaces and only one of the connections is established.

Just import the HTTPBasicAuth object that you created into other files in which you need to use it. The "@" thing is called a decorator, it's a Python feature.

@kamranhossain are you asking me how to access your Mongo db from Tornado? I wouldn't know, that's not my area of expertise. I supposed any asyncio compatible Mongo library should work.

@calpaterson I just noticed that you quoted my tweet with results that more or less agree with you. This was before I've found the issues stated above. My results actually have shifted a bit after I fixed these problems, and I cannot say that they agree with yours anymore. Will publish a blog post with my investigation and results shortly.

Sleeping is a bad way of solving this. As I said above, there is a connect handler that is invoked on each namespace when the connection is established.

Why it was done this way? Because the Socket.IO protocol is asynchronous in nature, there are many messages that are exchanged between the server and the client until the point the connection is established, it's not a single thing.

Python has a query string parser as well: https://docs.python.org/3/library/urllib.parse.html#urllib.parse.parse_qs.

@JocoYo1 As I said above, your application is request a URL that the server hasn't implemented. This isn't an issue with this package.

Did you read the documentation?

its running now but on GPU I think it cant run on CPU.

You can't run CPU intensive tasks in the web server process, as that blocks it.

The background task is not associated with a client, so there is no request context. In particular, there is no way to modify the user session there, since the background task has no way to send the updated session cookie back to the client.

You have multiple db instances in your application. You should use only one, that is imported in all the places you need it.

@JocoYo1 I don't really follow what you are saying. The "not found" error is the 404, it means that the browser is asking for a URL that the server does not know about. It has nothing to do with what you were asking before, it's a different bug that you have in your application.

I think the problem is that you are not waiting for the connection to be fully established, which would be at the point where the connect handler for the /qwe namespace is called.

The correct way to do this is as follows:

• Start the timer as in your first example:

    threading.Timer(30, foo, [room_id])
  

• Create an app context when your foo function starts:

    def foo(room_id):
        with app.app_context():
            leave_handler(room_id)
  

@sr347012 As I asked the OP, please look in the network tab of your browser for errors. Something is obviously happening on your side, but I need some help in debugging this, since I can run the application just fine here.

I need to see the stack trace of the error. The problem is likely in your application or in a previous migration file.

When running my app locally in development, it's running under http. This means the js files referenced by Flask-Moment are retrieved over http. This is causing problems when I run some checks with the Lighthouse tool in Chrome Dev Tools. I'd like to be able to force the js references to be via https and not via http.

I propose making a change to include_moment with an optional variable of force_https. The default will be false and the function will work as at present and use whatever the application is running under. If it's set to true then it will use https irrespective of what the application is running.

Thank you!

I've added a small change to ensure that the js files are requested over https.

To implement this change, I've updated:

• flask_moment.py
• test_flask_moment.py
• README.md

I've run the tests and they're all passing. This is in relation to issue #64

Do you get the token in a header? All headers are sent to you in the environ dictionary, using the standard format used by WSGI applications.

Thanks. I have it in my list to review the signal handling changes so I'll have an update on this soon.

https://github.com/miguelgrinberg/Flask-SocketIO/blob/master/example/sessions.py

The backend works, sure. Are you running it? Look in the network tab of your browser to see more details.

@Ahzam-Lathiya there is an example in this repository that shows how to do this. I don't really see anything wrong in your code, you may want to test the example in this repo and adapt it to your needs.

I think it would be okay if we always use https:// for the js dependencies, without a setting. If you make this change and send a PR I'll merge it. Thanks!

@zero-shubham yes, the idea is that if an exception is raised inside a greenlet, that exception should be raised in the await of the function wrapped with async_(). I have tested this with requests and that appears to be working. If you have a small example that I can test here I can dig deeper and see why this isn't working in your case.

This is the repository for Flask-Migrate, which is a wrapper for Alembic. Your issue is with Alembic, but note that this package does not check column types by default, that needs to be enabled. See https://stackoverflow.com/questions/33944436/flask-migrate-and-changing-column-type/52181159#52181159 for the solution that I think will apply to your case.

Yeah, it will probably work. See the async_requests.py example in this repo, which integrates requests into an async application. PyMongo would use a similar solution. Would love to know if you succeed!

Then the problem is elsewhere, your connection requests aren't reaching the server.

I meant the server, not the client, sorry for not being clear. Add engineio_logger=True and logger=True to your SocketIO() class.

Yes, you would just use functions such as those shown in the API chapter.

Doesn't look like a CORS issue. Enable Socket.IO and Engine.IO logs to have more details about this error.

Use request.form.get('form-field-name-here').

The error is from Flask. You need to return a response.

Try this:

@app.route('/send', methods=['POST'])
def sendTestMessage():
    emit('testmessage', {'data': request.json}, namespace='/', broadcast=True)

Okay, thanks. But flask-socketio only supports socketio, right?

Correct.

is it possible to send data to clients when a url gets a post request?

Yes. Your example will broadcast the message to all your Socket.IO clients. From an HTTP handler it is a bit tricky to know how to only address that same client that sent the HTTP request through a parallel Socket.IO connection.

Your face_rec.py should be this:

def main_face2():
    from __main__ import socketio

    # ...

The main file of the application is called __main__, regardless of the filename.

This is something that needs to be decided for each application. Some applications have public users, so this does not present any privacy issue. Some others, obviously you wouldn't want the list of users to be known, so you would not have a get_user endpoint in your API.