Codebase for the submission from the Hybrid Intelligence Team 1 in the Google Cloud GenAI Hackathon 2023.
This repository is meant to be run on a machine with the following packages installed:
- <=Python-3.10
- <=pip-22.0
- <=git-2.34
The examples in this README are written for GNU/Linux, but should be valid for any machine with standard GNU Coreutils and BusyBox installed (e.g. BSD, MacOS, WSL, etc.).
The API requires proper credentials to access outside of GCP. The recommended method is through gcloud authentication. The instructions below are for installing gcloud on Ubuntu and most Debian-based GNU/Linux distributions.
# Install prerequisites
sudo apt install -y apt-transport-https ca-certificates gnupg
# Add new source to package list
echo "deb [signed-by=/usr/share/keyrings/cloud.google.gpg] https://packages.cloud.google.com/apt cloud-sdk main" | sudo tee -a /etc/apt/sources.list.d/google-cloud-sdk.list
# Download the public key for Google Cloud and add to keyring
curl https://packages.cloud.google.com/apt/doc/apt-key.gpg | sudo apt-key --keyring /usr/share/keyrings/cloud.google.gpg add -
# Update package manager and install gcloud
sudo apt update && sudo apt install -y google-cloud-cli
# Check gloud successfully installed
gcloud version
# Authenticate with Google Credentials (copy link into browser)
gcloud auth application-default login
# Set project id
gcloud config set project gen-hybrid-intelligence-team-1
gcloud auth application-default set-quota-project gen-hybrid-intelligence-team-1
The repository code should run on any Python3 environment as long as it
contains all the dependencies from the requirements.txt file. It is
recommended to manage the environment with venv, pipenv, conda, poetry,
etc. for portibility.
For simplicity, the example instructions here uses the python3-venv module. More robust solutions can be used, but venv has the advantage of being easily accessible on most operating systems.
cd /path/to/repo
# Create a new venv directory
python3 -m venv ./venv
# Activate the newly created venv profile
. venv/bin/activate
Assuming the build-in venv module was used, install all the required dependencies into the environment. The pip3 tool was used here, but it can vary depending on the chosen environment.
cd /path/to/repo
# Activate the venv profile
. venv/bin/activate
# Install the requirements into the newly created venv profile
pip3 install -r ./requirements.txt
The main application is accessible with the prototype Streamlit application. As such, the app can be run with the following commands:
cd /path/to/repo
# Activate the venv profile
. venv/bin/activate
# Run the Streamlit app
streamlit bin/streamlit_shopper.py
# Open a browser and navigate to the URL
# (e.g. https://localhost:8501)
The ./doc directory in this repository includes API references as well
as a quickstart and step-by-step guide. These files are automatically
generated from Google-style Python docstrings with the pdoc package as
shown below.
cd /path/to/repo
# OPTIONAL: Activate venv profile if using venv
. venv/bin/activate
# OPTIONAL: Install pdoc if not already installed
pip3 install pdoc
# Run pdoc on localhost
pdoc -d "google" ./src
# Save HTML documentation
pdoc -d "google" -o ./doc ./src