Update app.py

app.py

@@ -1,65 +1,349 @@
 import gradio as gr
-import torch
-from transformers import pipeline
+import json
+import os
+from datetime import datetime
+from transformers import pipeline, AutoTokenizer, AutoModelForCausalLM
 
-# 简单的代码修复演示函数
-def code_fix_function(input_code):
-    """
-    简单的代码修复演示
-    在实际应用中，这里应该加载你训练好的模型
-    """
-    # 这里先使用一个简单的规则作为演示
-    # 实际应该替换为你的模型预测代码
-    
-    # 示例修复规则
-    fixes = {
-        "print()": "print()",
-        "if =": "if condition:",
-        "for i in range(:": "for i in range():",
-    }
-    
-    # 简单的模式匹配修复
-    fixed_code = input_code
-    for error, correction in fixes.items():
-        if error in input_code:
-            fixed_code = fixed_code.replace(error, correction)
-    
-    # 如果没匹配到特定错误，添加通用建议
-    if fixed_code == input_code:
-        fixed_code = input_code + "\n\n# 建议：检查语法错误和缩进"
-    
-    return fixed_code
+class SmartCodeFixer:
+    def __init__(self):
+        self.feedback_file = "user_feedback.json"
+        self.model = None
+        self.tokenizer = None
+        self.load_model()
+        self.load_feedback_data()
+    
+    def load_model(self):
+        """加载预训练模型"""
+        try:
+            # 使用一个较好的代码生成模型
+            model_name = "microsoft/DialoGPT-medium"  # 或者 "codellama/CodeLlama-7b-hf"
+            self.tokenizer = AutoTokenizer.from_pretrained(model_name)
+            self.model = AutoModelForCausalLM.from_pretrained(model_name)
+            print("模型加载成功!")
+        except Exception as e:
+            print(f"模型加载失败: {e}")
+            self.model = None
+    
+    def load_feedback_data(self):
+        """加载用户反馈数据用于学习"""
+        if os.path.exists(self.feedback_file):
+            with open(self.feedback_file, 'r', encoding='utf-8') as f:
+                self.feedback_data = json.load(f)
+        else:
+            self.feedback_data = []
+    
+    def save_feedback(self, original_code, fixed_code, user_feedback, is_correct):
+        """保存用户反馈用于模型进化"""
+        feedback_entry = {
+            "timestamp": datetime.now().isoformat(),
+            "original": original_code,
+            "fixed": fixed_code,
+            "feedback": user_feedback,
+            "correct": is_correct,
+            "language": self.detect_language(original_code)
+        }
+        
+        self.feedback_data.append(feedback_entry)
+        
+        # 保存到文件
+        with open(self.feedback_file, 'w', encoding='utf-8') as f:
+            json.dump(self.feedback_data, f, ensure_ascii=False, indent=2)
+        
+        # 定期重新训练模型（简化版）
+        if len(self.feedback_data) % 10 == 0:  # 每10个反馈重新学习
+            self.retrain_from_feedback()
+    
+    def detect_language(self, code):
+        """智能检测编程语言"""
+        code_lower = code.lower()
+        
+        language_indicators = {
+            'html': ['<!doctype', '<html', '<div', '<span', 'class="', 'id="'],
+            'python': ['def ', 'import ', 'print(', 'if __name__', 'lambda '],
+            'javascript': ['function ', 'console.log', 'document.', 'addEventListener'],
+            'java': ['public class', 'public static', 'System.out.println'],
+            'cpp': ['#include', 'using namespace', 'cout <<', 'std::'],
+            'css': ['{', '}', ':', ';', 'font-size', 'color:']
+        }
+        
+        scores = {lang: 0 for lang in language_indicators}
+        
+        for lang, indicators in language_indicators.items():
+            for indicator in indicators:
+                if indicator in code_lower:
+                    scores[lang] += 1
+        
+        return max(scores.items(), key=lambda x: x[1])[0]
+    
+    def ai_fix_code(self, code, language):
+        """使用AI模型修复代码"""
+        if self.model is None:
+            return self.rule_based_fix(code, language)
+        
+        try:
+            # 构建修复提示
+            prompt = f"""修复以下{language}代码的错误：
 
-def main():
-    # 创建Gradio界面
-    iface = gr.Interface(
-        fn=code_fix_function,
-        inputs=gr.Textbox(
-            lines=10, 
-            placeholder="输入需要修复的代码...",
-            label="输入代码"
-        ),
-        outputs=gr.Textbox(
-            lines=10, 
-            label="修复后的代码",
-            show_copy_button=True
-        ),
-        title="🐑 Capricode 代码修复助手",
-        description="输入有问题的代码，获取AI修复建议",
-        examples=[
-            ["print('Hello World'"],  # 缺少右括号
-            ["if = 10"],  # 错误的条件语句
-            ["for i in range(10"],  # 缺少右括号
-        ]
+错误代码:
+```{language}
+{code}
+"""
+            
+            inputs = self.tokenizer.encode(prompt, return_tensors="pt")
+            outputs = self.model.generate(
+                inputs, 
+                max_length=len(inputs[0]) + 100,
+                num_return_sequences=1,
+                temperature=0.7,
+                do_sample=True
+            )
+            
+            response = self.tokenizer.decode(outputs[0], skip_special_tokens=True)
+            # 提取修复后的代码
+            if "```" in response:
+                fixed_code = response.split("```")[2].strip()
+            else:
+                fixed_code = response.replace(prompt, "").strip()
+            
+            return fixed_code
+            
+        except Exception as e:
+            print(f"AI修复失败: {e}")
+            return self.rule_based_fix(code, language)
+    
+    def rule_based_fix(self, code, language):
+        """基于规则的代码修复"""
+        fixes = {
+            'html': self.fix_html,
+            'python': self.fix_python,
+            'javascript': self.fix_javascript,
+            'java': self.fix_java,
+            'cpp': self.fix_cpp,
+            'css': self.fix_css
+        }
+        
+        fix_function = fixes.get(language, self.fix_generic)
+        return fix_function(code)
+    
+    def fix_html(self, code):
+        """修复HTML代码"""
+        fixes = []
+        
+        # 检查标签闭合
+        if '<div>' in code and '</div>' not in code:
+            code += '\n</div>'
+            fixes.append("添加了缺失的 </div> 标签")
+        
+        # 检查属性引号
+        if 'class=' in code and 'class="' not in code:
+            code = code.replace('class=', 'class="')
+            if '"' not in code[code.find('class="')+7:code.find('class="')+20]:
+                code = code.replace('class="', 'class=""')
+            fixes.append("修复了属性引号")
+        
+        # 添加基本的HTML结构
+        if '<!DOCTYPE html>' not in code and '<html>' not in code:
+            code = f"""<!DOCTYPE html>
+<html>
+<head>
+    <meta charset="UTF-8">
+    <title>Document</title>
+</head>
+<body>
+{code}
+</body>
+</html>"""
+            fixes.append("添加了基本的HTML结构")
+        
+        return code, fixes
+    
+    def fix_python(self, code):
+        """修复Python代码"""
+        fixes = []
+        
+        # 修复括号
+        if code.count('(') > code.count(')'):
+            code += ')' * (code.count('(') - code.count(')'))
+            fixes.append("修复了不匹配的括号")
+        
+        # 修复引号
+        if code.count('"') % 2 != 0:
+            code += '"'
+            fixes.append("修复了不匹配的双引号")
+        
+        if code.count("'") % 2 != 0:
+            code += "'"
+            fixes.append("修复了不匹配的单引号")
+        
+        # 修复冒号
+        lines = code.split('\n')
+        for i, line in enumerate(lines):
+            if any(keyword in line for keyword in ['if ', 'for ', 'def ', 'class ', 'while ']) and not line.rstrip().endswith(':'):
+                lines[i] = line.rstrip() + ':'
+                fixes.append("在条件/函数声明后添加了冒号")
+        
+        return '\n'.join(lines), fixes
+    
+    def fix_javascript(self, code):
+        """修复JavaScript代码"""
+        fixes = []
+        
+        # 修复括号
+        if code.count('(') > code.count(')'):
+            code += ')' * (code.count('(') - code.count(')'))
+            fixes.append("修复了不匹配的括号")
+        
+        # 修复花括号
+        if code.count('{') > code.count('}'):
+            code += '}' * (code.count('{') - code.count('}'))
+            fixes.append("修复了不匹配的花括号")
+        
+        return code, fixes
+    
+    def fix_css(self, code):
+        """修复CSS代码"""
+        fixes = []
+        
+        # 修复选择器
+        if ':' in code and ';' not in code:
+            code += ';'
+            fixes.append("添加了缺失的分号")
+        
+        return code, fixes
+    
+    def fix_java(self, code):
+        """修复Java代码"""
+        fixes = []
+        
+        # 添加基本的类结构
+        if 'public class' in code and '{' not in code:
+            code = code.replace('public class', 'public class Main {') + '\n    public static void main(String[] args) {\n        \n    }\n}'
+            fixes.append("添加了基本的类结构")
+        
+        return code, fixes
+    
+    def fix_cpp(self, code):
+        """修复C++代码"""
+        fixes = []
+        
+        if '#include' in code and 'int main' not in code:
+            code += '\n\nint main() {\n    return 0;\n}'
+            fixes.append("添加了main函数")
+        
+        return code, fixes
+    
+    def fix_generic(self, code):
+        """通用修复"""
+        fixes = ["进行了通用语法检查"]
+        return code, fixes
+    
+    def retrain_from_feedback(self):
+        """根据用户反馈重新训练模型（简化版）"""
+        print("正在从用户反馈中学习...")
+        # 这里可以添加增量学习的逻辑
+        # 目前先记录反馈，后续可以真正重新训练模型
+
+# 创建修复器实例
+fixer = SmartCodeFixer()
+
+def process_code(input_code, use_ai=True):
+    """处理代码修复"""
+    language = fixer.detect_language(input_code)
+    
+    if use_ai and fixer.model is not None:
+        fixed_code = fixer.ai_fix_code(input_code, language)
+        fixes = ["使用AI模型修复"]
+    else:
+        fixed_code, fixes = fixer.rule_based_fix(input_code, language)
+    
+    # 生成修复报告
+    report = f"""🔧 修复报告
+📝 检测语言: {language}
+✅ 修复内容: {', '.join(fixes) if fixes else '代码看起来没问题'}
+
+修复后的代码:"""
+    
+    return fixed_code, report
+
+def handle_feedback(original_code, fixed_code, user_feedback, is_correct):
+    """处理用户反馈"""
+    fixer.save_feedback(original_code, fixed_code, user_feedback, is_correct)
+    return "感谢您的反馈！系统正在学习改进... 💡"
+
+# 创建Gradio界面
+with gr.Blocks(theme=gr.themes.Soft()) as demo:
+    gr.Markdown("# 🚀 Capricode 智能代码修复助手")
+    gr.Markdown("支持 HTML, Python, JavaScript, Java, C++, CSS 等多种语言！")
+    
+    with gr.Tab("代码修复"):
+        with gr.Row():
+            with gr.Column():
+                input_code = gr.Textbox(
+                    label="📥 输入需要修复的代码",
+                    placeholder="粘贴你的代码到这里...",
+                    lines=10
+                )
+                use_ai = gr.Checkbox(label="使用AI智能修复", value=True)
+                fix_btn = gr.Button("🔧 修复代码", variant="primary")
+            
+            with gr.Column():
+                output_code = gr.Textbox(
+                    label="📤 修复后的代码",
+                    lines=10,
+                    show_copy_button=True
+                )
+                report = gr.Textbox(
+                    label="📊 修复报告",
+                    lines=3
+                )
+    
+    with gr.Tab("反馈学习"):
+        gr.Markdown("## 💡 帮助系统变得更好")
+        with gr.Row():
+            with gr.Column():
+                feedback_original = gr.Textbox(label="原始代码", lines=3)
+                feedback_fixed = gr.Textbox(label="修复后的代码", lines=3)
+                user_feedback = gr.Textbox(
+                    label="您的反馈建议",
+                    placeholder="这里可以如何改进？",
+                    lines=3
+                )
+                is_correct = gr.Radio(
+                    choices=[("正确修复", True), ("需要改进", False)],
+                    label="修复是否正确？"
+                )
+                feedback_btn = gr.Button("提交反馈", variant="secondary")
+                feedback_result = gr.Textbox(label="反馈结果", interactive=False)
+    
+    # 事件处理
+    fix_btn.click(
+        fn=process_code,
+        inputs=[input_code, use_ai],
+        outputs=[output_code, report]
     )
     
-    # 启动应用
-    iface.launch(
-        server_name="0.0.0.0", 
-        server_port=7860,
-        share=False
+    feedback_btn.click(
+        fn=handle_feedback,
+        inputs=[feedback_original, feedback_fixed, user_feedback, is_correct],
+        outputs=[feedback_result]
+    )
+    
+    # 示例
+    gr.Markdown("## 🎯 试试这些例子:")
+    gr.Examples(
+        examples=[
+            ["<div>Hello World", True],  # HTML
+            ["print('Hello World'", True],  # Python
+            ["function test() {", True],  # JavaScript
+            ["public class MyClass", True],  # Java
+        ],
+        inputs=[input_code, use_ai]
     )
 
 if __name__ == "__main__":
-    main()
+    demo.launch(
+        server_name="0.0.0.0",
+        server_port=7860,
+        share=True
+    )