vri-scores

A tool to quickly extract Virtual Regatta Inshore race results from screenshots

Features

• Screenshot OCR: Automatically extracts race results from screenshots using Tesseract OCR.
• Text Ranking Parsing: Supports direct text input with "Ranking:" followed by a list of names.
• Race Combination: Combines multiple screenshots in the same message into one race.
• Emoji Reactions: Uses number emojis (e.g., 1️⃣, 2️⃣) to label races.
• Total Score Calculation: Aggregates scores across races, handling DSQ, DNF, and DNS. Low point scoring system (see the Rule A4).
• Series Tie-Breaks (Rule A8): Applies the RRS A8 to break ties in series scores, using best-to-worst race results and, if needed, last-race performance.
• Fuzzy Name Matching: Attempts to correct minor OCR errors in participant names by matching against previously seen names (using Levenshtein distance).
• Reset Command: Type !reset to clear the bot's internal race data for the channel, allowing a new regatta to start. The last generated table message remains in the chat.
• Single Table Display: While a regatta is active, only the latest generated race table is kept in the channel; previous tables for that regatta are automatically deleted upon update.
• Reaction-Based Updates: Updates to the race table occur when reacting to a ranking message with a number emoji. Editing a ranking message does not automatically update the table; you must re-react.
• Fast and Responsive: Fast extraction of race results with a final aggregated table display.

Example of ranking from a screenshot:

⬆️ As you may notice, the recognition was not ideal in the example above. This is the reality of fast and cheap OCR.

Example of a race table after three races:

Not supported (yet)

• medal races
• worst race[s] exclusion
• non-integer scores

Installation

pip install -r requirements.txt
sudo apt-get install tesseract-ocr  # apt-based Linux
brew install tesseract              # macOS

You also need the fonts-noto-cjk font.

Deployment

Bare Metal

To run on bare metal, simply run:

python main.py

Ensure that Tesseract and required system dependencies are installed as per the Installation instructions.

Docker

To run using Docker, use the provided Dockerfile and docker-compose.yml:

docker-compose up --build

This command builds the Docker image and starts the container.

Support

Now you can buy me a coffee to encourage further development!