Connecting local computers to compute a problem. This program allows you to send functions to other computers on a local network and receive the computed output.
- Specify
PORT
inmain.py
to tell the server what port to run the program on. - Write
functions.py
which will include all the functions that could possibly be executed on the client. - Write
generate_function(user_id)
inmain.py
which will return 3 values:- the function itself
- a dictionary or tuple of the inputs to the function
- the name of the function
- Write
handle_received(received, user_id)
to specify what to do with the incoming data the client computed - Make sure that the client has all the dependencies installed, or make a function in
functions.py
to install them. - Run the server with
python3 server.py -m main.py -f functions.py
- Run the client with
python3 client.py SERVER_IP PORT
For help with running server.py
or client.py
run python3 server.py -h
or python3 client.py -h
respectively.
For a basic example, see example_functions.py
and example_main.py
. More complicated examples using Tensorflow and Keras will be added soon.
This program is in alpha and still greatly under development, so expect bugs. This is not to be used on a public network nor is any developer of this program responsible for damage or information leaked because of this program. You must be extremely careful when using this program with Tensorflow or Keras with a Tensorflow backend, as graphs become very tricky to manage.
Add examples using Keras and/or Tensorflow. Add on_connect and on_disconnect function.