gemini_pictionary_fullcode.py

Pictionary.py

import pygame
import requests
import json
import base64

# Initialize Pygame
pygame.init()

# Screen dimensions
width, height = 640, 480
screen = pygame.display.set_mode((width, height))
pygame.display.set_caption("Pictionary with Gemini")

# Colors
white = (255, 255, 255)
black = (0, 0, 0)
purple = (128, 0, 128)

# Drawing variables
drawing = False
last_pos = (0, 0)
brush_size = 5

# Button variables
button_width = 200
button_height = 50
button_x = (width - button_width) // 2
validate_button_y = height - button_height - 20
clear_button_y = validate_button_y - button_height - 20

# Gemini API credentials
api_key = "YOUR_API_KEY_HERE"
model = "gemini-1.5-pro"
api_url = f"https://api.google.com/generative/v1beta2/models/{model}:generateContent?key={api_key}"

# Function to draw the drawing surface
def draw_surface():
    pygame.draw.rect(screen, white, (0, 0, width, height - 100))

# Function to draw a button
def draw_button(text, x, y, width, height, color):
    font = pygame.font.Font(None, 24)
    text_surface = font.render(text, True, white)
    text_rect = text_surface.get_rect(center=(x + width // 2, y + height // 2))
    pygame.draw.rect(screen, color, (x, y, width, height))
    screen.blit(text_surface, text_rect)

# Function to send image to Gemini and get guess
def get_gemini_guess(image_data):
    headers = {"Content-Type": "application/json"}
    data = {
        "prompt": {
            "text": """
            You are a pictionary player. Given the provided image, identify its contents 
            and reply in the form of a JSON payload response containing the following properties:
            {
              "guessedImage": the name of the image drawn 
            }
            """
        },
        "image": {"data": image_data.decode("utf-8")},
    }
    response = requests.post(api_url, headers=headers, json=data)
    response.raise_for_status()
    response_json = response.json()
    return response_json["candidates"][0]["output"]["text"]


# Function to display the Gemini guess
def display_guess(guess):
    font = pygame.font.Font(None, 36)
    text_surface = font.render("Gemini guesses: " + guess, True, purple)
    text_rect = text_surface.get_rect(center=(width // 2, height // 2))
    screen.blit(text_surface, text_rect)

# Main game loop
running = True
guess = ""
while running:
    for event in pygame.event.get():
        if event.type == pygame.QUIT:
            running = False
        elif event.type == pygame.MOUSEBUTTONDOWN:
            if event.button == 1:
                drawing = True
                last_pos = event.pos
            elif event.button == 3:
                # Right click to clear canvas
                draw_surface()
        elif event.type == pygame.MOUSEBUTTONUP:
            if event.button == 1:
                drawing = False
        elif event.type == pygame.MOUSEMOTION:
            if drawing:
                pygame.draw.line(screen, black, last_pos, event.pos, brush_size)
                last_pos = event.pos
        elif event.type == pygame.KEYDOWN:
            if event.key == pygame.K_RETURN:
                # Convert the drawing surface to a base64 encoded string
                pygame.image.save(screen, "drawing.png")
                with open("drawing.png", "rb") as image_file:
                    image_data = base64.b64encode(image_file.read())
                try:
                    guess = get_gemini_guess(image_data)
                    guess = json.loads(guess)["guessedImage"]
                except Exception as e:
                    guess = f"Error: {e}"

    # Draw buttons
    draw_button("Validate Picture", button_x, validate_button_y, button_width, button_height, purple)
    draw_button("Clear", button_x, clear_button_y, button_width, button_height, purple)

    # Display Gemini guess
    if guess:
        display_guess(guess)

    # Update display
    pygame.display.flip()

pygame.quit()

MartialTerran/Pictionary.py
Forked from romanejaquez/gist:d8dd3b36d39efce7f1c53a6db4969844