feat: AI套图分层方案 + Gemini集成 - 4种图案类型处理 + 正片叠底 + 宽高比 + 模型选择
Co-authored-by: Cursor <cursoragent@cursor.com>
This commit is contained in:
223
AI改图-双图.py
Normal file
223
AI改图-双图.py
Normal file
@@ -0,0 +1,223 @@
|
||||
#!/usr/bin/env python3
|
||||
# -*- coding: utf-8 -*-
|
||||
"""
|
||||
双图合成处理 - 基于 multi_image_merge 方法
|
||||
使用 Gemini 3 Pro Image Preview 模型进行双图创意合成
|
||||
|
||||
特点：
|
||||
- 支持本地图片和网络图片 URL 两种输入方式
|
||||
- 支持自定义提示词
|
||||
- 创意合成、场景融合
|
||||
|
||||
使用方法：
|
||||
1. 安装依赖：pip install openai requests
|
||||
2. 在下方配置区填入您的 API Key
|
||||
3. 修改文件底部的参数变量
|
||||
4. 运行：python 双图合成处理.py
|
||||
|
||||
获取 API Key：https://api.laozhang.ai/token
|
||||
"""
|
||||
|
||||
import base64
|
||||
import re
|
||||
import os
|
||||
from openai import OpenAI
|
||||
|
||||
# ========== 配置区（请填入您的 API Key）==========
|
||||
API_KEY = "sk-KZ9TGgZAWnzY1T3c90B277F72e184c26A53f22440e08E86e" # 替换为您的 API Key
|
||||
BASE_URL = "https://api.laozhang.ai/v1"
|
||||
|
||||
# 模型选择
|
||||
MODEL = "gemini-3-pro-image-preview" # 最新版，支持更高质量
|
||||
# MODEL = "gemini-2.5-flash-image" # 稳定版，价格更低（$0.025 vs $0.05）
|
||||
# ================================================
|
||||
|
||||
|
||||
def extract_and_save_image(content: str, filename: str) -> bool:
|
||||
"""从响应内容中提取 base64 图片并保存"""
|
||||
# 匹配 markdown 格式的 base64 图片
|
||||
match = re.search(r'!\[.*?\]\((data:image/\w+;base64,([^)]+))\)', content)
|
||||
if match:
|
||||
base64_data = match.group(2)
|
||||
# 确保 base64 填充正确
|
||||
padding = 4 - len(base64_data) % 4
|
||||
if padding != 4:
|
||||
base64_data += '=' * padding
|
||||
image_data = base64.b64decode(base64_data)
|
||||
with open(filename, 'wb') as f:
|
||||
f.write(image_data)
|
||||
return True
|
||||
return False
|
||||
|
||||
|
||||
def is_url(path: str) -> bool:
|
||||
"""判断路径是否为 URL"""
|
||||
return path.startswith('http://') or path.startswith('https://')
|
||||
|
||||
|
||||
def multi_image_merge(image1_path: str, image2_path: str, prompt: str, output_filename: str = None) -> bool:
|
||||
"""
|
||||
多图合成
|
||||
|
||||
参数:
|
||||
image1_path: 第一张图片路径（本地文件路径或 URL）
|
||||
image2_path: 第二张图片路径（本地文件路径或 URL）
|
||||
prompt: 合成提示词
|
||||
output_filename: 输出文件名（可选）
|
||||
|
||||
返回: 是否成功
|
||||
"""
|
||||
print("\n" + "="*60)
|
||||
print("🎨 双图合成处理")
|
||||
print("="*60)
|
||||
|
||||
client = OpenAI(api_key=API_KEY, base_url=BASE_URL)
|
||||
|
||||
# 构建 content 数组
|
||||
content = [{"type": "text", "text": prompt}]
|
||||
|
||||
# 处理第一张图片
|
||||
if is_url(image1_path):
|
||||
# 网络图片 URL
|
||||
content.append({"type": "image_url", "image_url": {"url": image1_path}})
|
||||
print(f"🖼️ 图片1 (URL): {image1_path[:50]}...")
|
||||
else:
|
||||
# 本地图片 Base64
|
||||
if not os.path.exists(image1_path):
|
||||
print(f"❌ 错误：图片1不存在: {image1_path}")
|
||||
return False
|
||||
with open(image1_path, "rb") as f:
|
||||
image1_b64 = base64.b64encode(f.read()).decode("utf-8")
|
||||
content.append({
|
||||
"type": "image_url",
|
||||
"image_url": {
|
||||
"url": f"data:image/jpeg;base64,{image1_b64}"
|
||||
}
|
||||
})
|
||||
print(f"🖼️ 图片1 (本地): {image1_path}")
|
||||
|
||||
# 处理第二张图片
|
||||
if is_url(image2_path):
|
||||
# 网络图片 URL
|
||||
content.append({"type": "image_url", "image_url": {"url": image2_path}})
|
||||
print(f"🖼️ 图片2 (URL): {image2_path[:50]}...")
|
||||
else:
|
||||
# 本地图片 Base64
|
||||
if not os.path.exists(image2_path):
|
||||
print(f"❌ 错误：图片2不存在: {image2_path}")
|
||||
return False
|
||||
with open(image2_path, "rb") as f:
|
||||
image2_b64 = base64.b64encode(f.read()).decode("utf-8")
|
||||
content.append({
|
||||
"type": "image_url",
|
||||
"image_url": {
|
||||
"url": f"data:image/jpeg;base64,{image2_b64}"
|
||||
}
|
||||
})
|
||||
print(f"🖼️ 图片2 (本地): {image2_path}")
|
||||
|
||||
print(f"📝 提示词: {prompt}")
|
||||
print("📡 发送请求...")
|
||||
|
||||
try:
|
||||
response = client.chat.completions.create(
|
||||
model=MODEL,
|
||||
messages=[
|
||||
{
|
||||
"role": "user",
|
||||
"content": content
|
||||
}
|
||||
]
|
||||
)
|
||||
|
||||
result_content = response.choices[0].message.content
|
||||
|
||||
# 确定输出文件名
|
||||
if output_filename:
|
||||
output_file = output_filename
|
||||
else:
|
||||
output_file = "result_multi_merge.png"
|
||||
|
||||
if extract_and_save_image(result_content, output_file):
|
||||
print(f"✅ 成功！图片已保存: {output_file}")
|
||||
return True
|
||||
else:
|
||||
print(f"⚠️ 未找到图片数据")
|
||||
return False
|
||||
|
||||
except Exception as e:
|
||||
print(f"❌ 错误: {str(e)}")
|
||||
return False
|
||||
|
||||
|
||||
def main():
|
||||
"""主函数"""
|
||||
print("="*60)
|
||||
print("🎨 双图合成处理")
|
||||
print(f"模型: {MODEL}")
|
||||
print("="*60)
|
||||
|
||||
# 检查 API Key
|
||||
if API_KEY == "sk-YOUR_API_KEY" or not API_KEY:
|
||||
print("\n❌ 请先配置您的 API Key！")
|
||||
print(" 获取地址: https://api.laozhang.ai/token")
|
||||
print(" 然后修改脚本顶部的 API_KEY 变量")
|
||||
return
|
||||
|
||||
# 检查图片是否存在（仅本地文件）
|
||||
if not is_url(IMAGE1_PATH) and not os.path.exists(IMAGE1_PATH):
|
||||
print(f"\n❌ 错误：图片1文件不存在: {IMAGE1_PATH}")
|
||||
print(" 请检查文件路径是否正确")
|
||||
return
|
||||
|
||||
if not is_url(IMAGE2_PATH) and not os.path.exists(IMAGE2_PATH):
|
||||
print(f"\n❌ 错误：图片2文件不存在: {IMAGE2_PATH}")
|
||||
print(" 请检查文件路径是否正确")
|
||||
return
|
||||
|
||||
print(f"\n📋 配置信息：")
|
||||
print(f" 图片1: {IMAGE1_PATH}")
|
||||
print(f" 图片2: {IMAGE2_PATH}")
|
||||
print(f" 提示词: {PROMPT}")
|
||||
print()
|
||||
|
||||
# 执行合成
|
||||
result = multi_image_merge(
|
||||
image1_path=IMAGE1_PATH,
|
||||
image2_path=IMAGE2_PATH,
|
||||
prompt=PROMPT,
|
||||
output_filename=OUTPUT_FILENAME
|
||||
)
|
||||
|
||||
if result:
|
||||
print("\n" + "="*60)
|
||||
print("🎉 合成完成！")
|
||||
print("="*60)
|
||||
else:
|
||||
print("\n" + "="*60)
|
||||
print("❌ 合成失败，请检查错误信息")
|
||||
print("="*60)
|
||||
|
||||
|
||||
# ========== 参数配置区（请修改以下参数）==========
|
||||
|
||||
# 1. 第一张图片路径（支持本地文件路径或网络 URL）
|
||||
IMAGE1_PATH = "5.png" # 本地文件示例: "1.png" 或 "C:/images/photo1.jpg"
|
||||
# IMAGE1_PATH = "https://images.unsplash.com/photo-1514888286974-6c03e2ca1dba?w=800&q=80" # URL 示例
|
||||
|
||||
# 2. 第二张图片路径（支持本地文件路径或网络 URL）
|
||||
IMAGE2_PATH = "6.png" # 本地文件示例: "2.png" 或 "C:/images/photo2.jpg"
|
||||
# IMAGE2_PATH = "https://images.unsplash.com/photo-1560806887-1e4cd0b6cbd6?w=800&q=80" # URL 示例
|
||||
|
||||
# 3. 合成提示词（描述如何合成这两张图片）
|
||||
PROMPT = "把第2张图的衣股花型图案替换成图1的花型图案" # 修改为您想要的合成效果描述
|
||||
|
||||
# 4. 输出文件名（可选，留空则使用默认名称）
|
||||
OUTPUT_FILENAME = None # 例如: "合成结果.png" 或 None（使用默认名称）
|
||||
|
||||
# ================================================
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
main()
|
||||
|
||||
Reference in New Issue
Block a user