This moves the listeners to an async service so that there is no longer a need for the time.sleep hack after indexing a file to get it to work.

To fix this, I used an async service and added a different type of event - a 'file indexed' event that the message_listener is listening to.

In order to make all this work, I did change the rabbitmq dependency from a blockingChannel to an AbstractChannel. I tested to make sure that files trigger extractors when added/indexed and also, that manual submission of datasets and files still works.

To test this, do the following.

First, comment out in docker-compose.dev.yml the part related to the message_listener.

Second, run the dependencies using docker-dev.sh up as usual.

Third, run the message_listener using pycharm, in debug mode.

Fourth, run the backend in debug mode. Run the front end as usual.

Test the following:

Create a dataset and add a file - it should create an event that is picked up by the message_listener when the file is indexed, and this should trigger the check_feed_listeners and any extractors that should run on the file.

Also, to verify that everything extractor related works as before, manually submit files and datasets to extractors.