Game Walkthroughs Weaviate demo

A Weaviate demo dataset with game walkthroughs.

Build Status

Branch	Status
Master

Running the demo

Requirements

• python 3 and venv
• docker and docker-compose are installed

Spin up the service

docker-compose up -d

It may take a while to download the image. After the service is up we still need some time until weaviate is reachable by url (around 30 secs - 1 minute). To check if the graph is reachable, simply open http://localhost:8080/ and see if there is any response from the graph.

While waiting for it, we can configure the data seed.

Configure how we populate the data

Game data

• We can populate the game data inside games file

• Game data is formatted as: [Game name];[Developer name];[Genre(s) - comma separated];[Platform(s) - comma separated]

• Example:

  GTA V;Rockstar Games;Action-adventure Game,First-person Shooter;PlayStation 3,PlayStation 4,PlayStation 5,Xbox One,Xbox 360,Xbox Series X,Microsoft Windows
  

  This will later be interpreted as:

  • create a Game instance named GTA V with developer of Rockstar Game
  • create 2 genres of Action-adventure Game and First-person Shooter, and later associate both of it with GTA V
  • create 7 platforms and associate all of it with GTA V

• All creation are get_or_create method, so same instance will not be duplicated.

Video and Subtitle data

• We can populate the video and its subtitle data inside video_links file
• Game data is formatted as: [Game name];[Video link from Youtube]
• example: GTA V;https://www.youtube.com/watch?v=Vncf_9LLagc
• The script will get the video's metadata and its Youtube generated auto caption using youtube-dl, create an instance inside the graph and associate with the Game

Running the script

• Create python 3 venv, get into it, and install all the requirement

  pip3 install -r requirements.txt
  

• Make sure that weaviate is reachable by opening http://localhost:8080/, if there is any 200 response then it should be good

Creating schema

• Create the schema by running create_schema.py:

python3 project/create_schema.py

By now, if you open the weaviate playground pointed on your localhost and check all the schema you will see something like this:

Populating schema

• Populate the schema by running populate schema.py

python3 project/create_schema.py

It will take quite a moment according how much the videos is listed and duration of each video.

After done, we will get soemthing like this:

We can now query or explore the data.