PhoneMapServer

An similar idea to Google's MapReduce, but for phones.

Setup

To get dependencies run sudo ./build.sh

Also make sure the configuration in config.py is correct for your machine, or (once we've implemented it) make sure you've set the appropriate environment variables.

To get postgresql, run:

sudo apt install postgresql

Then in a terminal write:

sudo -su postgres

And then:

createuser -P -s -e YOUR_USER

(set some password of your choosing, say, YOUR_PW)

followed by:

createdb -O YOUR_USER YOUR_DB_NAME

and finally, to test:

psql YOUR_USER -h YOUR_HOST OUR_DB_NAME

(you'll get a port number at this point, say, YOUR_PORT)

That's it! You can quit with:

\q

followed by ctrl-D to stop being the postgres user.

From there you'll want to edit your config.py so it has a line such as:

SQLALCHEMY_DATABASE_URI = postgresql://YOUR_USER:YOUR_PW@YOUR_HOST:YOUR_PORT/YOUR_DB_NAME

Mine for example is:

SQLALCHEMY_DATABASE_URI = postgresql://rob:[topsecretpw]@127.0.0.1:5432/phonemap

Running Tests

Go into the PhoneMapServer directory and in terminal run python3 -m "nose"

Current API endpoints:

connect: The usual socketio call; required to initiate . Emits a my_response with data saying CON_OK.

disconnect: What is a connection without a disconnection?

my_ping: Used by the JS front end to calculate the response time. Responds with my_pong.

my_event: A standard call. Not used for anything, echos whatever you send it under my_response. Useful for the JS frontend.

my_broadcast_event: Similar to my_event, but echos the message over all devices. Still uses my_response.

get_code: Responds a set_code with some code and data as args, which consists of the JS code to be executed and the data to be crunched. This will return the next data to be crunched for the oldest task. May also respond with a no_tasks in the event that no more tasks are available.

get_code_by_id: Same response as get_code. Must be sent with a task_id arg so as to specify which task is desired, instead of first task available obtained fromget_code.

start_code: Marks the code as started in the database. required for submission, and should follow a successful get_code call. May respond with a stop_executing in unexpected scenarios; eg the code you have just started has already been completed, or if you haven't requested code to execute yet (which would be very confusing indeed).

execution_failed: Informs the server the currently running code has failed to execute.

return: Marks the completed crunching of a piece of data, updating the database appropriately. Also broadcasts the result under my_response, as usual.

get_task_list: Returns a task_list with arg list containing a list of JSON data for each task.

Note: In the event of an error, error may be emitted, which means the server has attempted to gracefully recover and tried to inform you of the error under the error argument.