Panterinha — FURIA Fan Chatbot

An unofficial Flask chatbot prototype for Counter-Strike fans. Panterinha combines fuzzy intent recognition with modular web-scraping adapters to answer questions about FURIA's teams, matches, events, rankings, and lineups.

Note

This is an independent educational fan project. It is not affiliated with or endorsed by FURIA or HLTV. Live answers depend on third-party page structure and may become unavailable when those pages change.

What it demonstrates

• A Flask JSON endpoint consumed by a browser chat interface.
• Accent-insensitive message normalization.
• Fuzzy keyword matching with RapidFuzz.
• Intent routing separated from response formatting.
• Team-context detection for the main and women's rosters.
• Modular parsers for matches, events, rankings, and lineups.
• Brazilian timezone conversion for match and event timestamps.
• Containerized execution with Docker and deployment configuration for Fly.io.
• Offline automated tests for HTTP routes, intent selection, team context, and HTML parsing.

Supported questions

Panterinha recognizes questions about:

• Recent and upcoming matches.
• Main and women's lineups.
• Upcoming events.
• Global and Brazilian rankings.
• Where to watch matches.
• FURIA facts and official social links.

Request flow

Browser chat
    -> POST /chat
    -> normalize + fuzzy intent match
    -> select team context
    -> local response or scraping adapter
    -> formatted JSON response

Only intents that need changing information call the external adapters. Greetings, help, social links, and trivia are produced locally.

Technology stack

• Python 3.13
• Flask
• RapidFuzz and Unidecode
• Beautiful Soup and Cloudscraper
• HTML, CSS, and JavaScript
• Gunicorn and Docker
• Pytest and GitHub Actions

Running locally

git clone https://github.com/ViniOcCode/furia-chatbot.git
cd furia-chatbot
python -m venv .venv

Activate the environment:

# Linux/macOS
source .venv/bin/activate

# Windows PowerShell
.venv\Scripts\Activate.ps1

Install and start the development server:

pip install -r requirements.txt
python main.py

Open http://localhost:5000.

Docker

docker build -t furia-chatbot:1.0 .
docker run --rm -p 8080:5000 furia-chatbot:1.0

Open http://localhost:8080. Additional container commands are documented in README.docker.md.

Tests

pip install -r requirements-dev.txt
pytest -q

The tests intentionally use controlled HTML fixtures rather than contacting HLTV, making the CI suite deterministic and respectful of the external service.

Project structure

app/
  controllers/chat.py       Flask routes
  models/chatresponses.py   Intent routing and response formatting
  models/matches.py         Match parser
  models/events.py          Event parser
  models/lineup.py          Roster parser
  models/ranking.py         Ranking parser
  models/utils.py           Keywords, teams, dates, and shared HTTP client
  static/                   Browser UI assets
  templates/                Chat page
tests/                      Offline behavior and parser tests
main.py                     Application entry point

Limitations

• The chatbot uses fuzzy keyword routing rather than a trained language model.
• Scraping adapters are coupled to third-party HTML and require maintenance when markup changes.
• Some static trivia and ranking routes reflect the project's original 2025 prototype period.
• Responses are informational and should be checked against official sources for current competitive data.

License

MIT

Português

Sobre o projeto

Panterinha é um protótipo não oficial de chatbot para fãs de Counter-Strike. A aplicação usa Flask, reconhecimento aproximado de intenções e módulos de scraping para responder perguntas sobre partidas, eventos, rankings e escalações da FURIA.

Este é um projeto educacional independente, sem afiliação ou endosso da FURIA ou da HLTV. Respostas dinâmicas dependem da estrutura de páginas externas e podem deixar de funcionar quando essas páginas mudam.

O que o projeto demonstra

• API Flask consumida pela interface de chat.
• Normalização de mensagens e reconhecimento de intenção com RapidFuzz.
• Detecção de contexto entre os times principal e feminino.
• Parsers separados para partidas, eventos, rankings e escalações.
• Conversão de datas para o fuso brasileiro.
• Docker, configuração de deploy e testes automatizados sem acesso à rede.

Execução

Crie um ambiente virtual, instale requirements.txt, execute python main.py e acesse http://localhost:5000.

Para executar os testes, instale requirements-dev.txt e rode pytest -q.