video-create/tools/gemini3.py

import requests
import base64
from PIL import Image
import json
import io
import os
from dotenv import load_dotenv
from interfaces.image_output import ImageOutput
from utils.logger_config import setup_logger

logger = setup_logger(__name__)

load_dotenv()
API_KEY = os.getenv("GEMINI_API_KEY")

def image_to_base64(image_path):
    """将图片转换为base64字符串"""
    with Image.open(image_path) as img:
        # 转换为RGB模式（如果需要）
        if img.mode != 'RGB':
            img = img.convert('RGB')
        
        # 将图片转换为base64
        buffer = io.BytesIO()
        img.save(buffer, format='PNG')
        img_base64 = base64.b64encode(buffer.getvalue()).decode('utf-8')
        return img_base64


def generate_image_from_prompt_and_images(
    prompt="An office group photo of these people, they are making funny faces.",
    image_paths=['output_front.png'],
    aspect_ratio="5:4",
    resolution="1K",
    api_key=API_KEY,
    url="https://api.openaius.com/v1beta/models/gemini-3-pro-image-preview:generateContent"
):
    """
    使用Gemini API根据提示词和参考图片生成新图片
    
    Args:
        prompt (str): 提示词
        image_paths (list): 参考图片路径列表
        aspect_ratio (str): 图片宽高比
        resolution (str): 图片分辨率
        api_key (str): API密钥
        url (str): API端点URL
        
    Returns:
        dict: API响应结果
    """
    # 读取并转换所有图片
    images_base64 = []
    for img_path in image_paths:
        try:
            img_b64 = image_to_base64(img_path)
            images_base64.append(img_b64)
        except Exception as e:
            logger.error(f"Error loading image {img_path}: {e}")
            return None

    # 构建请求体
    payload = {
        "contents": [
            {
                "role": "user",
                "parts": [
                    {"text": prompt},
                    *[{"inline_data": {"mime_type": "image/png", "data": img_b64}} for img_b64 in images_base64]
                ]
            }
        ],
        "generation_config": {
            "response_modalities": ["TEXT", "IMAGE"],
            "image_config": {
                "aspect_ratio": aspect_ratio,
                "image_size": resolution
            }
        }
    }

    # 设置请求头
    headers = {
        "Authorization": f"Bearer {api_key}",
        "Content-Type": "application/json"
    }

    # 发送请求
    try:
        response = requests.post(url, headers=headers, json=payload, timeout=60)
        response.raise_for_status()  # 检查HTTP错误
        
        result = response.json()

        # 处理响应
        if "candidates" in result and len(result["candidates"]) > 0:
            candidate = result["candidates"][0]
            if "content" in candidate and "parts" in candidate["content"]:
                for part in candidate["content"]["parts"]:
                    if "text" in part:
                        logger.info(part["text"])
                    elif "inlineData" in part and part["inlineData"]["mimeType"] == "image/jpeg":
                        img_data = base64.b64decode(part["inlineData"]["data"])
        
        return ImageOutput(fmt="b64", ext="png", data=img_data)
        
    except requests.exceptions.RequestException as e:
        logger.error(f"Request failed: {e}")
        if hasattr(e, 'response') and e.response is not None:
            logger.error(f"Response status: {e.response.status_code}")
            logger.error(f"Response body: {e.response.text}")
    except json.JSONDecodeError as e:
        logger.error(f"Failed to parse JSON response: {e}")
        logger.error(f"Raw response: {response.text}")
    except Exception as e:
        logger.error(f"Unexpected error: {e}")
    
    return None


# 如果直接运行此脚本，则执行函数
if __name__ == "__main__":
    prompt = "生成一张美女图片，表情可爱"
    result = generate_image_from_prompt_and_images(prompt, image_paths=[])

    result.save_img("output_banana_pro.png")