How to multithread call another endpoint method in python?

Question

I have an api with 2 endpoints, one is a simple post receiving a json and the other is an endpoint which calls the 1st multiple times depending on the length of a list of jsons and save the return to a list.

First method

@app.route('/getAudience', methods=['POST', 'OPTIONS'])
def get_audience(audience_=None):
    try:
        if audience_:
            audience = audience_
        else:
            audience = request.get_json()
    except (BadRequest, ValueError):
        return make_response(jsonify(exception_response), 500)

    return get_audience_response(audience, exception_response)

Second method

@app.route('/getMultipleAudience', methods=['POST', 'OPTIONS'])
def get_multiple_audience():
    try:
        audiences = request.json
    except (BadRequest, ValueError):
        return make_response(jsonify(exception_response), 500)

    response = []
    for audience in audiences:
        new_resp = json.loads(get_audience(audience).data)
        response.append(new_resp)

    return make_response(jsonify(response))

I wanted to call the first method starting a thread per object in the list of the second method so I tried this:

def get_multiple_audience():
    with app.app_context():
        try:
            audiences = request.get_json()
        except (BadRequest, ValueError):
            return make_response(jsonify(exception_response), 500)

        for audience in audiences:
            thread = Thread(target=get_audience, args=audience)
            thread.start()

        thread.join()
        return make_response(jsonify(response))

And got this error:

Exception in thread Thread-6:
Traceback (most recent call last):
  File "C:\Python27\lib\threading.py", line 801, in __bootstrap_inner
    self.run()
  File "C:\Python27\lib\threading.py", line 754, in run
    self.__target(*self.__args, **self.__kwargs)
  File "C:\Python27\lib\site-packages\flask_cors\decorator.py", line 123, in wrapped_function
    options = get_cors_options(current_app, _options)
  File "C:\Python27\lib\site-packages\flask_cors\core.py", line 286, in get_cors_options
    options.update(get_app_kwarg_dict(appInstance))
  File "C:\Python27\lib\site-packages\flask_cors\core.py", line 299, in get_app_kwarg_dict
    app_config = getattr(app, 'config', {})
  File "C:\Python27\lib\site-packages\werkzeug\local.py", line 347, in __getattr__
    return getattr(self._get_current_object(), name)
  File "C:\Python27\lib\site-packages\werkzeug\local.py", line 306, in _get_current_object
    return self.__local()
  File "C:\Python27\lib\site-packages\flask\globals.py", line 51, in _find_app
    raise RuntimeError(_app_ctx_err_msg)
RuntimeError: Working outside of application context.

So then I tried to modify the first method like this:

@app.route('/getAudience', methods=['POST', 'OPTIONS'])
def get_audience(audience_=None):
    with app.app_context():
        try:
        ...

And got the same error. Can anyone give me a hint, advice, best practice or solution?

If you want to test your web service, the best way is to create a separate/independent application using Requests. — Laurent LAPORTE
– Laurent LAPORTE, Commented Aug 10, 2017 at 16:52

darksky · Accepted Answer · 2017-08-10 20:27:20Z

There are multiple problems here. Firstly, here:

for audience in audiences:
    thread = Thread(target=get_audience, args=audience)
    thread.start()

thread.join()

You are only waiting for the last thread to complete. You should have a list of all the threads, and wait for all of them to complete.

threads = []
for audience in audiences:
    thread = Thread(target=get_audience, args=audience)
    threads.append(thread)
    thread.start()

for thread in threads:
    thread.join()

Second problem is that you are returning a single response which isn't even set anywhere. But that's not how multi-threading works. You will have multiple results from all the threads and you will have to keep track of them. So you can create a results array to hold the answers for each thread's return value. Here I will make a simple function sum as an example.

results = []
threads = []

def sum(a, b):
    results.append(a + b)

@app.route("/test")
def test():
    with app.app_context():
        for i in range(5):
            t = Thread(target=sum, args=(1, 2))
            threads.append(t)
            t.start()

        for t in threads:
            t.join()

        return jsonify(results)

This will happily work, and it will return the result of all the calls to sum() function.

Now if I change sum to:

@app.route("/mysum/a/b")
def sum(a, b):
    results.append(a + b)
    return jsonify(a + b)

I will get a similar error as the one you were getting earlier: namely, RuntimeError: Working outside of request context., even thought the return value would still be correct: [3, 3, 3, 3, 3]. What's happening here is that your sum function is now trying to return a flask response, but it is residing inside its own temporary thread and doesn't have access to any of flask's internal contexts. What you should do is to never return a value inside a temporary worker thread, but have a pool to store them for future reference.

But this doesn't mean you can't have a /mysum route. Indeed, you can, but the logic has to be separated. To put it all together:

results = []
threads = []

def sum(a, b):
    return a + b

def sum_worker(a, b):
    results.append(sum(a, b))

@app.route("/mysum/a/b")
def mysum(a, b):
    return jsonify(sum(a, b))

@app.route("/test")
def test():
    with app.app_context():
        for i in range(5):
            t = Thread(target=sum_worker, args=(1, 2))
            threads.append(t)
            t.start()

        for t in threads:
            t.join()

        return jsonify(results)

Note that this code is very crude and is only for demonstration purposes. I can't recommend making global variables throughout your app, so some cleanup is required.

Collectives™ on Stack Overflow

How to multithread call another endpoint method in python?

1 Answer 1

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Your Answer

Sign up or log in

Post as a guest

Related