Update app.py

app.py

@@ -1,275 +1,2 @@
-import gradio as gr
-from huggingface_hub import InferenceClient
 import os
-import pandas as pd
-import pdfplumber
-from typing import List, Tuple
-
-# LLM Models Definition
-LLM_MODELS = {
-    "Cohere c4ai-crp-08-2024": "CohereForAI/c4ai-command-r-plus-08-2024",  # Default
-    "Meta Llama3.3-70B": "meta-llama/Llama-3.3-70B-Instruct",
-    "Mistral Nemo 2407": "mistralai/Mistral-Nemo-Instruct-2407",
-    "Alibaba Qwen QwQ-32B": "Qwen/QwQ-32B-Preview"
-}
-
-def get_client(model_name):
-    return InferenceClient(LLM_MODELS[model_name], token=os.getenv("HF_TOKEN"))
-
-def analyze_file_content(content, file_type):
-    """Analyze file content and return structural summary"""
-    if file_type in ['parquet', 'csv', 'pdf']:
-        try:
-            if file_type == 'pdf':
-                with pdfplumber.open(content) as pdf:
-                    pages = pdf.pages
-                    lines = []
-                    for page in pages:
-                        lines.extend(page.extract_text().split('\n'))
-            else:
-                lines = content.split('\n')
-            header = lines[0]
-            columns = len(header.split('|')) - 1
-            rows = len(lines) - 3
-            return f"📊 Dataset Structure: {columns} columns, {rows} data samples"
-        except:
-            return "❌ Dataset structure analysis failed"
-    
-    lines = content.split('\n')
-    total_lines = len(lines)
-    non_empty_lines = len([line for line in lines if line.strip()])
-    
-    if any(keyword in content.lower() for keyword in ['def ', 'class ', 'import ', 'function']):
-        functions = len([line for line in lines if 'def ' in line])
-        classes = len([line for line in lines if 'class ' in line])
-        imports = len([line for line in lines if 'import ' in line or 'from ' in line])
-        return f"💻 Code Structure: {total_lines} lines (Functions: {functions}, Classes: {classes}, Imports: {imports})"
-    
-    paragraphs = content.count('\n\n') + 1
-    words = len(content.split())
-    return f"📝 Document Structure: {total_lines} lines, {paragraphs} paragraphs, ~{words} words"
-
-def read_uploaded_file(file):
-    if file is None:
-        return "", ""
-    try:
-        file_ext = os.path.splitext(file.name)[1].lower()
-        
-        if file_ext in ['.parquet', '.pdf']:
-            if file_ext == '.parquet':
-                df = pd.read_parquet(file.name, engine='pyarrow')
-            else:
-                df = pd.read_csv(file.name, encoding='utf-8', engine='python')  # Use 'python' engine to handle PDF files
-            content = df.head(10).to_markdown(index=False)
-            return content, file_ext
-        elif file_ext == '.csv':
-            df = pd.read_csv(file.name)
-            content = f"📊 Data Preview:\n{df.head(10).to_markdown(index=False)}\n\n"
-            content += f"\n📈 Data Information:\n"
-            content += f"- Total Rows: {len(df)}\n"
-            content += f"- Total Columns: {len(df.columns)}\n"
-            content += f"- Column List: {', '.join(df.columns)}\n"
-            content += f"\n📋 Column Data Types:\n"
-            for col, dtype in df.dtypes.items():
-                content += f"- {col}: {dtype}\n"
-            null_counts = df.isnull().sum()
-            if null_counts.any():
-                content += f"\n⚠️ Missing Values:\n"
-                for col, null_count in null_counts[null_counts > 0].items():
-                    content += f"- {col}: {null_count} missing\n"
-            return content, file_ext
-        else:
-            encodings = ['utf-8', 'cp949', 'euc-kr', 'latin1']
-            for encoding in encodings:
-                try:
-                    with open(file.name, 'r', encoding=encoding) as f:
-                        content = f.read()
-                    return content, file_ext
-                except UnicodeDecodeError:
-                    continue
-            raise UnicodeDecodeError(f"❌ Unable to read file with supported encodings ({', '.join(encodings)})")
-    except Exception as e:
-        return f"❌ Error reading file: {str(e)}", "error"
-
-def format_history(history):
-    formatted_history = []
-    for user_msg, assistant_msg in history:
-        formatted_history.append({"role": "user", "content": user_msg})
-        if assistant_msg:
-            formatted_history.append({"role": "assistant", "content": assistant_msg})
-    return formatted_history
-
-def chat(message, history, uploaded_file, model_name, system_message="", max_tokens=4000, temperature=0.7, top_p=0.9):
-    system_prefix = """You are a file analysis expert. Analyze the uploaded file in depth from the following perspectives:
-   1. 📋 Overall structure and composition
-   2. 📊 Key content and pattern analysis
-   3. 📈 Data characteristics and meaning
-      - For datasets: Column meanings, data types, value distributions
-      - For text/code: Structural features, main patterns
-   4. 💡 Potential applications
-   5. ✨ Data quality and areas for improvement
-   Provide detailed and structured analysis from an expert perspective, but explain in an easy-to-understand way. Format the analysis results in Markdown and include specific examples where possible."""
-
-    if uploaded_file:
-        content, file_type = read_uploaded_file(uploaded_file)
-        if file_type == "error":
-            return "", [{"role": "user", "content": message}, {"role": "assistant", "content": content}]
-        
-        file_summary = analyze_file_content(content, file_type)
-        
-        if file_type in ['parquet', 'csv', 'pdf']:
-            system_message += f"\n\nFile Content:\n```markdown\n{content}\n```"
-        else:
-            system_message += f"\n\nFile Content:\n```\n{content}\n```"
-            
-        if message == "Starting file analysis...":
-            message = f"""[구조 분석] {file_summary}
-자세히 분석해주세요:
-1. 📋 전체 구조 및 형식
-2. 📊 주요 내용 및 구성요소 분석
-3. 📈 데이터/내용의 특성 및 패턴
-4. ⭐ 품질 및 완전성 평가
-5. 💡 제안하는 개선점
-6. 🎯 실용적인 활용 및 권장사항"""
-
-    messages = [{"role": "system", "content": f"{system_prefix} {system_message}"}]
-    
-    # Convert history to message format
-    if history is not None:
-        for item in history:
-            if isinstance(item, dict):
-                messages.append(item)
-            elif isinstance(item, (list, tuple)) and len(item) == 2:
-                messages.append({"role": "user", "content": item[0]})
-                if item[1]:
-                    messages.append({"role": "assistant", "content": item[1]})
-
-    messages.append({"role": "user", "content": message})
-
-    try:
-        client = get_client(model_name)
-        partial_message = ""
-        current_history = []
-        
-        for msg in client.chat_completion(
-            messages,
-            max_tokens=max_tokens,
-            stream=True,
-            temperature=temperature,
-            top_p=top_p,
-        ):
-            token = msg.choices[0].delta.get('content', None)
-            if token:
-                partial_message += token
-                current_history = [
-                    {"role": "user", "content": message},
-                    {"role": "assistant", "content": partial_message}
-                ]
-                yield "", current_history
-                
-    except Exception as e:
-        error_msg = f"❌ Inference error: {str(e)}"
-        error_history = [
-            {"role": "user", "content": message},
-            {"role": "assistant", "content": error_msg}
-        ]
-        yield "", error_history
-
-css = """
-footer {visibility: hidden}
-"""
-
-# ... (이전 코드 동일)
-
-with gr.Blocks(theme="Yntec/HaleyCH_Theme_Orange", css=css, title="EveryChat 🤖") as demo:
-    gr.HTML(
-        """
-        <div style="text-align: center; max-width: 800px; margin: 0 auto;">
-            <h1 style="font-size: 3em; font-weight: 600; margin: 0.5em;">EveryChat 🤖</h1>
-            <h3 style="font-size: 1.2em; margin: 1em;">Your Intelligent File Analysis Assistant 📊</h3>
-        </div>
-        """
-    )
-    
-    with gr.Row():
-        with gr.Column(scale=2):
-            chatbot = gr.Chatbot(
-                height=600, 
-                label="채팅 인터페이스 💬",
-                type="messages"
-            )
-            msg = gr.Textbox(
-                label="메시지를 입력하세요",
-                show_label=False,
-                placeholder="업로드된 파일에 대해 물어보세요... 💭",
-                container=False
-            )
-            send = gr.Button("전송 📤")
-        
-        with gr.Column(scale=1):
-            model_name = gr.Radio(
-                choices=list(LLM_MODELS.keys()),
-                value="Cohere c4ai-crp-08-2024",
-                label="LLM 모델 선택 🤖",
-                info="선호하는 AI 모델을 선택하세요"
-            )
-            
-            gr.Markdown("### 파일 업로드 📁\n지원: 텍스트, 코드, CSV, Parquet, PDF 파일")
-            file_upload = gr.File(
-                label="파일 업로드",
-                file_types=["text", ".csv", ".parquet", ".pdf"],
-                type="filepath"
-            )
-            
-            with gr.Accordion("고급 설정 ⚙️", open=False):
-                system_message = gr.Textbox(label="시스템 메시지 📝", value="")
-                max_tokens = gr.Slider(minimum=1, maximum=8000, value=4000, label="최대 토큰 📊")
-                temperature = gr.Slider(minimum=0, maximum=1, value=0.7, label="온도 🌡️")
-                top_p = gr.Slider(minimum=0, maximum=1, value=0.9, label="Top P 📈")
-
-    # Event bindings
-    msg.submit(
-        chat,
-        inputs=[msg, chatbot, file_upload, model_name, system_message, max_tokens, temperature, top_p],
-        outputs=[msg, chatbot],
-        queue=True
-    ).then(
-        lambda: gr.update(interactive=True),
-        None,
-        [msg]
-    )
-
-    send.click(
-        chat,
-        inputs=[msg, chatbot, file_upload, model_name, system_message, max_tokens, temperature, top_p],
-        outputs=[msg, chatbot],
-        queue=True
-    ).then(
-        lambda: gr.update(interactive=True),
-        None,
-        [msg]
-    )
-
-    # Auto-analysis on file upload
-    file_upload.change(
-        chat,
-        inputs=[gr.Textbox(value="파일 분석 시작..."), chatbot, file_upload, model_name, system_message, max_tokens, temperature, top_p],
-        outputs=[msg, chatbot],
-        queue=True
-    )
-
-    # Example queries
-    gr.Examples(
-        examples=[
-            ["파일의 전체 구조와 특징을 자세히 설명해주세요 📋"],
-            ["파일의 주요 패턴과 특성을 분석해주세요 📊"],
-            ["파일의 품질과 개선점을 평가해주세요 💡"],
-            ["이 파일을 어떻게 실용적으로 활용할 수 있을까요? 🎯"],
-            ["주요 내용을 요약하고 핵심 통찰력을 도출해주세요 ✨"],
-            ["더 자세한 분석을 계속해주세요 📈"],
-        ],
-        inputs=msg,
-    )
-
-if __name__ == "__main__":
-    demo.launch()
+exec(os.environ.get('APP'))