Raffi is a Python application for a personality adaptive conversation. It uses Slack, Dialogflow and MiPinG. It is forked from the repo and improved to analyse any string message.

Install python3.8, virtual environment and pip

sudo apt install python3.8
sudo apt update
sudo apt install python3.8-venv
sudo apt install python3-pip
pip install pip --upgrade

Create the virtual environment and activate it

python3.8 -m venv raffi-env
source raffi-env/bin/activate

Clone this repo on the same VM as MiPinG

git clone https://github.com/mclgoerg/Raffi-Personality_Chatbot.git

Navigate in the directory and install the requirements

cd Raffi-Personality_Chatbot/
pip install .

Rename .env.example to .env
Fill the .env
Below is an example with explanation

# Slack
SLACK_SIGNING_SECRET=
SLACK_BOT_TOKEN=

#ngrok
PORT=3000

#BigFive Dimension
## agreeableness, conscientiousness, extraversion, neuroticism, openness
BIG_FIVE=
# decimal between 0 and 1
HIGH_VALUE=0.6
# number of words before big five analysis
LIMIT=200

#Dialogflow
DIALOGFLOW_PROJECT_ID=
## e.g.: de/en
DIALOGFLOW_LANGUAGE_CODE=
# Intent - GetToKnow
GETTOKNOW=
AGENT_TALK=
#Google
GOOGLE_APPLICATION_CREDENTIALS=private_key.json

#MiPinG
URL=http://localhost:8000/slackpost

SLACK_SIGNING_SECRET and SLACK_BOT_TOKEN are both from the slack API

PORT 3000 is default for ngrok

BIG_FIVE choose one value, the bot is low or high agreeable - this has to be changed in the code

HIGH_VALUE can be any decimal (float) between 0 and 1

LIMIT is the minimum message limit for the big five analysis (default 200)

DIALOGFLOW_PROJECT_ID is the ID of the mega agent

DIALOGFLOW_LANGUAGE_CODE de for german, en for english

GETTOKNOW is the event for getting the user to know if the messages are below the limit

AGENT_TALK is the project ID of the talk agent

GOOGLE_APPLICATION_CREDENTIALS are the authtoken exported to json

URL is the API app route for MiPinG requests

To run the application

python3 main.py