Chess Bot (Dialogflow + Stockfish)

Flask webhook that lets a Dialogflow agent play chess against Stockfish. Game state is stored in MySQL (Cloud SQL) keyed by the Dialogflow session URL.

Features

• Dialogflow webhook endpoint for chess actions
• Stockfish-backed move generation
• Persistent games in MySQL
• Cloud Run deployment script

Project Layout

• main.py: Flask app + Dialogflow webhook handlers
• chess_engine.py: Chess and Stockfish helpers
• entities/: Dialogflow entity definitions
• sql_commands/db_script.sql: database/table notes
• Dockerfile, deploy.sh: Cloud Run container + deploy commands
• stockfish_20011801_x64: bundled Stockfish binary

Requirements

• Python 3.8
• MySQL (Cloud SQL)

Install deps:

pip install -r requirements.txt

Configuration

The app expects Cloud SQL via a Unix socket. Set these environment variables:

• DB_USER
• DB_PASS
• DB_NAME
• CLOUD_SQL_CONN (e.g. PROJECT:REGION:INSTANCE)
• DB_SOCKET_DIR (optional, default /cloudsql)
• PORT (for gunicorn, e.g. 8080)

API

POST /api/v1/assistant

Dialogflow webhook handler. The request must include:

• queryResult.action
• queryResult.parameters
• session

Supported actions:

• CreateGame: starts a new game
• MakeMove: applies a user move, then plays Stockfish response
• RevertMove: reverts last two plies
• ResignGame: deletes the session row
• ShowValidMoves: lists legal moves

Parameters used:

• Color: white | black | random
• Level: 1-10 (Stockfish strength)
• ChessPiece: N, B, R, Q, K or empty for pawn
• Moves: move tokens (single square or [from,to])
• Specialmoves: short_castle | long_castle

Responses are Dialogflow fulfillmentText strings.

Database

The game table stores:

• session_url
• board (pickled chess.Board)
• player_color
• engine_level

See sql_commands/db_script.sql for a starter schema.

Deployment (Cloud Run)

Build and deploy using the included script:

./deploy.sh

This script expects Google Cloud credentials, a Cloud Run service, and Cloud SQL instance/secrets wired as shown in deploy.sh.