Development of "R3D-2-MSA" in conjunction with the Gutell Lab at UT-Austin. The server is a method of accessing sequence variations in an alignment using the number from a 3D.
In short, the server is a python flask application
that uses beanstalk to queue up jobs for
processing. Jobs are processed in the background using the python script
bin/worker.py. The results of each job are cached briefly to allow for pages
to be generated easily. Data is cached using redis.
The server is designed to not connect to the RCAD database in the main process, and only in the background workers. One consequence of this is that the information for the pdbs to display is not fetched from RCAD each request. Instead it is read from a JSON file for each request.
There is an example configuration file in conf/example-config.json. This
servers to configure both the web server and the background workers. The web
application reads conf/config.json, which is required for the server to run.
The worker defaults to reading this file but can be given other files as
detailed below.
The main webapp is app.py, while the production wsgi app is in wsgi.py. The
logic behind validation, production output and such is all in r3d2msa. The
bin folder contains useful scripts for the site. See the documentation in
each script for information on what they do.
You will have to have conf/config.json as well as have beanstalk and redis
running. Because there are several components I recommend using something like
supervisord to get everything running.
To run it in development simply do python app.py. That will start a
development server on localhost:5000. It is not needed to have a worker
running as well, but if you do not all jobs will be placed in a queue and never
processed.
You will need to use something like gunicorn, or
mod_wsgi to get the server running. The
main entry point is wsgi.py. Here it is essential to have something
monitoring all the required processes.
To run the worker do bin/worker.py --config <config-file> to set a config file.
When testing it is nice to just used canned data sometimes. In that case you
can just do bin/worker.py --config config-file --test-data <test-json-file>.
You can get a set of test data by doing:
curl -H 'Accept: application/json' http://rna.bgsu.edu/r3d-2-msa?units=<query>
Under normal operation the worker should never crash, but there are somethings that can take it down. Beanstalk crashing will crash the worker as well, and redis failing will cause lots of problems.