An AI chatbot for the Ontario Centre of Forensic Sciences (CFS), developed using Rasa and OpenAI. Answers forensics investigators' questions by performing semantic search across a large library of CFS technical documents to find the most relevant search results.

• Linux OS
  • Ubuntu 18.04 / 20.04
  • Debian 9 / 10
  • Red Hat 7 / 8
  • CentOS 7 / 8
• Python 3.8
• pip 21.1.3
• docker
• docker-compose

python3 -m venv ~/rasa-cfs-env
cd ~/rasa-cfs-env
source ./bin/activate

git clone https://github.com/aoruize/rasa-cfs.git

cd rasa-cfs
pip install -r requirements.txt

Ensure you have access to the OpenAI Beta API. Login to OpenAI and copy your API key from https://beta.openai.com/docs/developer-quickstart/your-api-keys.

To temporarily add the environment variable, run the following command in the terminal, using your own API key:

export OPENAI_API_KEY='sk-...'

To add the environment variable to the operating system permanently, add it to your shell configuration files. Add the above command to the end of the following files:

~/.bashrc
~/.profile

cd ~/rasa-cfs-env/rasa-cfs
rasa run actions

1. Download

    curl -sSL -o install.sh https://storage.googleapis.com/rasa-x-releases/0.41.1/install.sh
   

2. Install all files into default folder, /etc/rasa:

    sudo bash ./install.sh 
   

3. Start up Rasa X and wait until all containers are running (Using -d will run Rasa X in the background):

    cd /etc/rasa
    sudo docker-compose up -d
   

4. Access and set admin password:

    cd /etc/rasa
    sudo python3 rasa_x_commands.py create --update admin me <PASSWORD>
   

Navigate to the hostname or IP where your server is reachable and log in using your newly created password.

Learn more: https://rasa.com/docs/rasa-x/installation-and-setup/install/docker-compose

1. This repository is set up with a GitHub Actions workflow, so that whenever a change is pushed to the actions directory, a new Docker image for the Custom Action Server will be built and pushed to a container registry.

To specify which container registry and image name, open .github/workflows/action_server.yml and change docker_image_name: to your desired Docker registry and image name.

Also, follow the instructions in Building an Action Server Image to specify your docker_registry_login: and docker_registry_password: .

2. Add OpenAI API key environment variable to .env file:

    cd /etc/rasa
    vim .env
   

   Add your API key to the bottom of the file:

    OPENAI_API_KEY=sk-...
   

3. Replace the default app image with the image of your custom action server:

   To avoid your changes in the docker-compose file being overwritten when you update versions, you should not apply your changes to docker-compose.yml. Instead, create a new file called docker-compose.override.yml inside your /etc/rasa directory and apply your changes there. Docker will automatically take that file into account and override any attributes in docker-compose.yml with changes from the override file.

   The contents of docker-compose.override.yml should look like this:

    version: '3.4'
    services:
        app:
            restart: always
            image: <image:tag>
            expose:
                - "5055"
            environment:
                OPENAI_API_KEY: "${OPENAI_API_KEY}"
   

4. If your Docker containers are already running, take them down and then start Rasa X again:

    cd /etc/rasa
    sudo docker-compose down
    sudo docker-compose up -d
   

Learn more: https://rasa.com/docs/rasa-x/installation-and-setup/customize#connecting-a-custom-action-server

Rasa Open Source: https://rasa.com/docs/rasa/

Rasa X: https://rasa.com/docs/rasa-x/

Rasa Action Server: https://rasa.com/docs/action-server

Implement in-house semantic search instead of using OpenAI by following this guide:

• Part 1: https://medium.com/mlearning-ai/semantic-search-with-s-bert-is-all-you-need-951bc710e160
• Part 2: https://medium.com/mlearning-ai/search-rank-and-recommendations-35cc717772cb
• Code: https://github.com/99sbr/semantic-search-with-sbert