How to run

To run A-CMA DQN-based agents, you should run two applications along side: the A-CMA refactoring tool and its DQN server.

Run DQN server

DQN server contains a deep neural network named Deep Q-Network which is developed by python language. So in order to run DQN server, you should have python 3 installed and running on your computer.

First of all, clone DQN server from GitHub:

$ git clone https://github.com/hrahmadi71/acma_dqn_server.git

Then you should go to the directory and create a virtual environment in it:

$ cd acma_dqn_server
$ python3 -m venv .venv

After you’ve create virtual environment, you should enable it. On Linux or Mac, use this command:

$ source .venv/bin/activate

On Windows, use this command:

$ .venv\Scrips\activate

After enabling virtual environment, you can install the requirements of DQN server:

(.venv)$ pip install -r requirements.txt

Then run the server (on the default address which is localhost:5000):

(.venv)$ flask run

There’s 4 pre-trained models which you can use for refactoring: model_1 to model_4. To use one of pretrained models, go to address localhost:5000 via a web-browser, you must see a swagger panel. Then in the swagger panel, click on /load_weights/ then click on Try it out button and change the string to the name of a model. for example:

{
  "model_name": "model_1"
}

Run A-CMA

In order to run A-CMA tool, you must have java installed and running on your computer. After you’ve installed and configured java on your computer, you can run A-CMA, so first of all, clone it from its repository:

$ git clone https://github.com/hrahmadi71/a-cma.git

Then go to the a-cma directory:

$ cd ./a-cma

Now you can compile a-cma or run it without compiling. If you are in Linux or Mac you can run it via bash:

$ bash A-CMA/run.sh

If you are using Windows you can run the bat file:

$ .\A-CMA\run.bat