Active-Art is a computational creativity system, written as part of a coursework assignement at the University of Kent.
Active-Art is a python3 project and therefore requires a few libraries to be installed.
Install the following packages before attempting to run:
- vaderSentiment
- syllables
- pandas
- datapoint
- openai
Navigate to the src/ dir and run the following command
python3 main.py 1
main.py takes one argument and it is the number of runs to retrieve from Strava, this will also then determine the number of poems generated.
main.py : takes care of running the entire system
promptgen.py : generates a prompt using strava and weather data
eval.py : evaluates a response from gpt
gpt.py : handles requests to gpt
strava.py : handles requests to strava
weather.py : handles requests to datapoint (metoffice)
results.txt : artifacts (poems) are pasted here at the end of run-time, includes primer, prompt and poem
artifacts-X.txt : artifacts generated throughout the project's lifetime.
The project uses my personal API keys and secrets for Datapoint, Strava and GPT. Therefore these are not shared. Please replace the dummy values with the ones that are need in strava.py and export your own API keys for datapoint and GPT under the names: DATAPOINT_API_KEY and OPENAI_API_KEY
To get access to the Strava API, a one time authentication has to be done in the browser once you have you API key.