FIT4002_Ad_Transparency

Repository for a final year project at Monash University around ad transparency

Prerequisites

1. Create .env file under bot directory with proper environment variables set. See bot/.env.default for an example.
2. Make sure the correct chromedriver is being used, see bot/driver/driver.py and look for CHROMEDRIVER_PATH. Similar steps if using Firefox driver

Instructions for running the project in local

DB Project

1. Install Node and npm.
2. Move to the /db directory and run npm install to install all dependencies.
3. Run project with npm start.

Bot Project

1. Make sure you have the correct browser installed, depending on the driver you want to use. Currently Chrome and Firefox supported
2. Move to the /bot directory and install all python requirements by running pip install --upgrade -r requirements.txt.
3. Ensure the DB Project is running.
4. run set "AD_USERNAME=<username>" && set "USE_PROXIES=0" && set "CHANGE_LOCATION=1" && set "NUM_TERMS=2" && set "DB_URL=http://localhost:8080" && set "UPLOAD_LOGS=0" && python app.py. Where is the username of the google profile to run.

Instructions for running the project in a docker container

Installing Docker

1. Download and run the latest docker toolbox executable for your OS from here. Include VirtualBox and Kitematic if not already installed.
2. Run the Docker Quickstart Terminal. The first time this is run it will add configuration variables to the computer.
3. Confirm docker is working by typing the command: docker run hello-world

Running the project

1. Ensure node dependencies are up to date by running npm install in /db directory.
2. Ensure the correct bot is being used through the AD_USERNAME environment variable in docker-compose.yml
3. Run docker-compose up --build. If node import error occurs, update the module locally and retry.

Running the DB container individually.

1. With docker toolbox cmd still open. Open terminal in /db directory`.
2. Build the custom db docker image by running docker build -t <image_name> .. Where <image_name> is an arbitrary identifier you can set.
3. Run docker image in a container by running docker run --rm <image_name>. This container will be deleted once exited.

Running the Bot container individually.

1. With docker toolbox cmd still open. Open terminal in /bot directory`.
2. Build the custom bot docker image by running docker build -t <image_name> .. Where <image_name> is an arbitrary identifier you can set.
3. Ensure the DB Container is running.
4. Run docker image in a container by running docker run --rm --env AD_USERNAME=<username> <image_name>. Where is the username of the google profile to run in this container. This container will be deleted once exited.

Helpfull commands in docker.

• Once a container is built, we can navigate as root with docker run -it -rm <image_name> /bin/sh.
• List all docker images with docker images
• Clear all old containers with docker container prune
• Confirm that the proxy addresses are working by running python bot/ip_check.py.

Instructions for running project in AWS.

• See README.md in /aws directory.

TODO: Automate the process of generating task-definition.json file. Automate the running of tasks. Automate the stopping of tasks.