dota2stats.io

The stack behind dota2stats.io.

Wraps OpenDota's replay parser to provide a hosted API and web interface for parsing Dota 2 replay files into JSON objects.

Using the service

Anyone can upload a replay file via the form on the homepage. This kicks off the replay parser and adds data from the match to the database. Alternatively, call the HTTP API from any other program.

See the homepage for documentation on the API.

Parsing the results

The parsed replay is a text file, where each line is an event in the match stored as JSON. A helper class ReplaySummariser is provided in scripts/helpers/parser.py which turns this output into something useful.

Fantasy points

Fantasy points have been broken in the Dota 2 client for a while, but they can be calculated by the ReplaySummariser. See scripts/fantasy_points.py for an example.

To install dependencies and run the script,

1. Make sure you have Python >= 3.8 installed and in your path
2. Clone this repository
3. From the repository root run python -m venv venv to create a new virtual environment
4. Run source venv/bin/activate (Mac) or venv\Scripts\activate (Windows) to activate the virtual environment
5. Run pip install -r requirements.txt to install dependencies
6. Run the script with python scripts/fantasy_points.py /path/to/replay/file.dem

Architecture

dota2stats.io is made up of 4 components:

• Web front-end (Python/Flask)
• Replay parser (OpenDota's Java parser)
• Task queue & file database (Redis)
• Task worker (Python/Celery)

The front-end is the only component accessible outside the network. Docker volumes are used to provide persistent storage for the uploaded replay files.

When the front-end recieves a replay file it saves the file to disk then kicks off a Celery task. It's this task that passes the file to the replay parser, stores the results and updates the status in Redis.

Deployment

The whole app is launched via docker-compose, and the front-end is accessible on port 5000.

docker-compose build
docker-compose up

CD is managed through two workflows using Github Actions. When a PR is opened, the first workflow (build_images.yml) tries to build the Docker images in app and celery-worker. The second workflow (deploy.yml) runs when a commit is made to master and deploys the new master branch to the hosting server.