Update app.py

app.py

@@ -3,7 +3,7 @@ import torch
 from transformers import CLIPProcessor, CLIPModel, BlipProcessor, BlipForConditionalGeneration
 from PIL import Image
 import numpy as np
-import openai  # GPT API 调用
+from openai import OpenAI
 
 # 初始化模型
 clip_model = CLIPModel.from_pretrained("openai/clip-vit-base-patch32")
@@ -11,11 +11,8 @@ clip_processor = CLIPProcessor.from_pretrained("openai/clip-vit-base-patch32")
 blip_processor = BlipProcessor.from_pretrained("Salesforce/blip-image-captioning-base")
 blip_model = BlipForConditionalGeneration.from_pretrained("Salesforce/blip-image-captioning-base")
 
-# GPT API 配置
-openai.api_key = "your_openai_api_key"
-
 # 定义功能函数
-def analyze_images(image_a, image_b):
+def analyze_images(image_a, image_b, api_key):
     # BLIP生成描述
     def generate_caption(image):
         inputs = blip_processor(image, return_tensors="pt")
@@ -44,17 +41,17 @@ def analyze_images(image_a, image_b):
     cosine_similarity = np.dot(features_a, features_b.T) / (np.linalg.norm(features_a) * np.linalg.norm(features_b))
     latent_diff = np.abs(features_a - features_b).tolist()
 
-    # GPT API 调用生成文字描述
-    gpt_prompt = (
-        f"图片A的描述为：{caption_a}。图片B的描述为：{caption_b}。\n"
-        "请对两张图片的内容和潜在特征区别进行详细分析，并输出一个简洁但富有条理的总结。"
-    )
-    gpt_response = openai.Completion.create(
-        engine="text-davinci-003",
-        prompt=gpt_prompt,
-        max_tokens=150
+    # 调用 DeepSeek API 生成详细分析
+    client = OpenAI(api_key=api_key, base_url="https://api.deepseek.com")
+    gpt_response = client.chat.completions.create(
+        model="deepseek-chat",
+        messages=[
+            {"role": "system", "content": "You are a helpful assistant."},
+            {"role": "user", "content": f"图片A的描述为：{caption_a}。图片B的描述为：{caption_b}。\n请对两张图片的内容和潜在特征区别进行详细分析，并输出一个简洁但富有条理的总结。"}
+        ],
+        stream=False
     )
-    textual_analysis = gpt_response['choices'][0]['text'].strip()
+    textual_analysis = gpt_response.choices[0].message.content.strip()
     
     # 返回结果
     return {
@@ -75,6 +72,8 @@ with gr.Blocks() as demo:
         with gr.Column():
             image_b = gr.Image(label="图片B", type="pil")  # 使用 PIL 类型
     
+    api_key_input = gr.Textbox(label="API Key", placeholder="输入您的 DeepSeek API Key", type="password")
+    
     analyze_button = gr.Button("分析图片")
     result_caption_a = gr.Textbox(label="图片A描述", interactive=False)
     result_caption_b = gr.Textbox(label="图片B描述", interactive=False)
@@ -83,13 +82,13 @@ with gr.Blocks() as demo:
     result_text_analysis = gr.Textbox(label="详细分析", interactive=False, lines=5)
     
     # 分析逻辑
-    def process_analysis(img_a, img_b):
-        results = analyze_images(img_a, img_b)
+    def process_analysis(img_a, img_b, api_key):
+        results = analyze_images(img_a, img_b, api_key)
         return results["caption_a"], results["caption_b"], results["similarity"], results["latent_diff"], results["text_analysis"]
     
     analyze_button.click(
         fn=process_analysis,
-        inputs=[image_a, image_b],
+        inputs=[image_a, image_b, api_key_input],
         outputs=[result_caption_a, result_caption_b, result_similarity, result_latent_diff, result_text_analysis]
     )