Create app.py

app.py

@@ -0,0 +1,802 @@
+"""
+기업마당 AI 분석기 - 메인 애플리케이션
+"""
+import gradio as gr
+import pandas as pd
+import json
+import os
+import tempfile
+from pathlib import Path
+from typing import Optional, Tuple, List, Dict, Generator
+from datetime import datetime
+
+from utils import (
+    CATEGORY_CODES, REGION_LIST, SIDO_LIST, ORG_TYPE_OPTIONS, SORT_OPTIONS, STATUS_OPTIONS,
+    COMPANY_TYPE_OPTIONS, CORP_TYPE_OPTIONS, COMPANY_SIZE_OPTIONS, INDUSTRY_MAJOR_OPTIONS,
+    CORE_INDUSTRY_OPTIONS, NATIONAL_STRATEGIC_TECH, CREDIT_GRADE_OPTIONS, TCB_GRADE_OPTIONS,
+    ISO_CERT_OPTIONS, extract_region_from_text, extract_region_from_hashtags,
+    classify_org_type, parse_deadline, is_ongoing, calculate_age, calculate_company_age
+)
+
+from file_api import (
+    fetch_all_from_api, download_file, extract_text_from_file, extract_zip_files,
+    call_groq_api_stream, fetch_announcement_detail
+)
+
+
+# ============================================================
+# 공고 조회 함수
+# ============================================================
+def fetch_announcements(keyword="", category="전체", region="전체(지역)", org_type="전체",
+                        sort_by="등록일순", status_filter="진행중", page=1, rows=20) -> Tuple[pd.DataFrame, str]:
+    """기업마당 API로 공고 목록 조회"""
+    try:
+        items, error = fetch_all_from_api(category, region, keyword)
+        if error and not items:
+            return pd.DataFrame(), error
+        if not items:
+            return pd.DataFrame(), "⚠️ 검색 결과가 없습니다."
+        
+        data = []
+        for item in items:
+            if not isinstance(item, dict):
+                continue
+            author = item.get("author", "") or item.get("jrsdInsttNm", "")
+            title = item.get("title", "") or item.get("pblancNm", "")
+            exec_org = item.get("excInsttNm", "") or ""
+            hash_tags = item.get("hashTags", "")
+            
+            item_region = extract_region_from_hashtags(hash_tags)
+            if not item_region:
+                item_region = extract_region_from_text(title)
+            if not item_region:
+                item_region = extract_region_from_text(author)
+            
+            item_org_type = classify_org_type(author)
+            req_dt = item.get("reqstDt", "") or item.get("reqstBeginEndDe", "")
+            item_ongoing = is_ongoing(req_dt)
+            pub_date = item.get("pubDate", "") or item.get("creatPnttm", "") or ""
+            if pub_date and len(str(pub_date)) >= 10:
+                pub_date = str(pub_date)[:10]
+            
+            link = item.get("link", "") or item.get("pblancUrl", "")
+            pblanc_id = item.get("seq", "") or item.get("pblancId", "")
+            
+            attachments = []
+            file_url = item.get("flpthNm", "")
+            file_name = item.get("fileNm", "")
+            if file_url and file_name:
+                attachments.append({"url": file_url, "filename": file_name, "type": Path(file_name).suffix.lower()[1:] if Path(file_name).suffix else "unknown"})
+            
+            # 추가 첨부파일 (print 파일)
+            print_url = item.get("printFlpthNm", "")
+            print_name = item.get("printFileNm", "")
+            if print_url and print_name:
+                if not any(att['url'] == print_url for att in attachments):
+                    attachments.append({"url": print_url, "filename": print_name, "type": Path(print_name).suffix.lower()[1:] if Path(print_name).suffix else "unknown"})
+            
+            description = item.get("description", "") or item.get("bsnsSumryCn", "")
+            if description:
+                import re
+                description = re.sub(r'<[^>]+>', '', description).strip()
+            
+            deadline = parse_deadline(req_dt)
+            row = {
+                "지원분야": item.get("lcategory", "") or item.get("pldirSportRealmLclasCodeNm", ""),
+                "지원사업명": title, "신청기간": req_dt, "소관부처": author,
+                "수행기관": exec_org, "등록일": pub_date, "조회수": item.get("inqireCo", "") or "",
+                "상세링크": link, "공고ID": pblanc_id,
+                "사업개요": description[:200] + "..." if len(description) > 200 else description,
+                "첨부파일": attachments, "지원대상": item.get("trgetNm", ""),
+                "문의처": item.get("refrncNm", ""), "신청URL": item.get("rceptEngnHmpgUrl", ""),
+                "_org_type": item_org_type, "_ongoing": item_ongoing, "_deadline": deadline,
+                "_pub_date": pub_date, "_region": item_region,
+            }
+            data.append(row)
+        
+        if not data:
+            return pd.DataFrame(), "⚠️ 검색 결과가 없습니다."
+        
+        df = pd.DataFrame(data)
+        total_before_filter = len(df)
+        
+        if org_type == "중앙부처":
+            df = df[df["_org_type"] == "중앙부처"]
+        elif org_type == "지자체":
+            df = df[df["_org_type"] == "지자체"]
+        if region and region != "전체(지역)":
+            df = df[df["_region"] == region]
+        if status_filter == "진행중":
+            df = df[df["_ongoing"] == True]
+        
+        if sort_by == "등록일순":
+            df = df.sort_values(by="_pub_date", ascending=False)
+        elif sort_by == "마감일순":
+            df = df.sort_values(by="_deadline", ascending=True, na_position='last')
+        
+        if len(df) == 0:
+            return pd.DataFrame(), f"⚠️ 필터 조건에 맞는 결과가 없습니다. (전체 {total_before_filter}건 중)"
+        
+        total_filtered = len(df)
+        start_idx = (page - 1) * rows
+        end_idx = start_idx + rows
+        df_page = df.iloc[start_idx:end_idx].copy()
+        df_page.insert(0, "번호", range(total_filtered - start_idx, total_filtered - start_idx - len(df_page), -1))
+        internal_cols = [c for c in df_page.columns if c.startswith("_")]
+        df_page = df_page.drop(columns=internal_cols)
+        
+        status = f"✅ {len(df_page)}건 표시 (페이지 {page}) | 필터 결과: {total_filtered}건 | API 수집: {total_before_filter}건"
+        return df_page, status
+    except Exception as e:
+        import traceback
+        return pd.DataFrame(), f"❌ 오류: {str(e)[:80]}"
+
+
+# ============================================================
+# AI 분석 함수
+# ============================================================
+def analyze_announcement(detail_url, project_name, api_attachments=None, api_description="", progress=gr.Progress()):
+    """공고 첨부파일을 다운로드하고 AI로 분석"""
+    if not detail_url:
+        yield "❌ 분석할 공고를 선택해주세요."
+        return
+    
+    output = f"# 📄 {project_name}\n\n---\n\n"
+    attachments = api_attachments if api_attachments else []
+    all_text = f"## 공고명: {project_name}\n\n"
+    
+    if api_description:
+        all_text += f"### 사업개요:\n{api_description}\n\n"
+        output += f"📋 **사업개요**\n{api_description}\n\n"
+        yield output
+    
+    output += f"📎 **첨부파일 {len(attachments)}개 발견**\n\n"
+    yield output
+    
+    extracted_texts = []
+    if attachments:
+        with tempfile.TemporaryDirectory() as tmp_dir:
+            for i, att in enumerate(attachments):
+                progress(0.2 + (0.4 * i / len(attachments)), desc=f"파일 다운로드 중... ({i+1}/{len(attachments)})")
+                output += f"📥 다운로드 중: `{att['filename']}`\n"
+                yield output
+                
+                file_path, error = download_file(att['url'], tmp_dir, att['filename'])
+                if error:
+                    output += f"  - ⚠️ {error}\n"
+                    yield output
+                    continue
+                
+                if file_path and file_path.lower().endswith('.zip'):
+                    extracted = extract_zip_files(file_path, tmp_dir)
+                    for ext_file in extracted:
+                        text, err = extract_text_from_file(ext_file)
+                        if text:
+                            extracted_texts.append({"filename": os.path.basename(ext_file), "text": text})
+                            output += f"  - ✅ 텍스트 추출 성공 ({len(text):,} 글자)\n"
+                        yield output
+                elif file_path:
+                    text, err = extract_text_from_file(file_path)
+                    if text:
+                        extracted_texts.append({"filename": os.path.basename(file_path), "text": text})
+                        output += f"  - ✅ 텍스트 추출 성공 ({len(text):,} 글자)\n"
+                    yield output
+            
+            if extracted_texts:
+                all_text += "### 첨부파일 내용:\n\n"
+                for ext in extracted_texts:
+                    text_preview = ext['text'][:5000] if len(ext['text']) > 5000 else ext['text']
+                    all_text += f"#### 📄 {ext['filename']}\n{text_preview}\n\n"
+    
+    if len(all_text) < 100 and not extracted_texts:
+        output += "\n❌ **분석할 내용이 충분하지 않습니다.**\n"
+        yield output
+        return
+    
+    output += f"\n📊 **분석 준비 완료** (총 {len(all_text):,}자)\n\n---\n\n## 🤖 AI 분석 결과\n\n"
+    yield output
+    
+    progress(0.7, desc="AI 분석 중...")
+    system_prompt = """당신은 정부 지원사업 공고 분석 전문가입니다.
+주어진 공고 내용을 분석하여 다음 항목을 명확하게 정리해주세요:
+- 사업명, 주관기관, 지원 목적
+- 신청 자격 요건, 제외 대상
+- 지원 금액/규모, 지원 항목/내용
+- 신청 기간, 신청 방법, 제출 서류
+- 중요 유의사항, 제한 사항
+- 이 사업의 핵심 포인트를 3줄로 요약"""
+    
+    messages = [
+        {"role": "system", "content": system_prompt},
+        {"role": "user", "content": f"다음 지원사업 공고를 분��해주세요:\n\n{all_text[:15000]}"}
+    ]
+    
+    for chunk in call_groq_api_stream(messages):
+        output += chunk
+        yield output
+    
+    output += "\n\n---\n✅ **분석 완료**"
+    yield output
+
+
+# ============================================================
+# 맞춤 과제 매칭 함수
+# ============================================================
+def analyze_uploaded_documents(files, progress=gr.Progress()):
+    """업로드된 문서들을 분석하여 기업 정보 추출"""
+    if not files:
+        yield "❌ 분석할 파일을 업로드해주세요."
+        return
+    
+    output = "# 📄 업로드 문서 분석 결과\n\n"
+    all_extracted_text = []
+    
+    for i, file in enumerate(files):
+        progress((i + 1) / len(files), desc=f"파일 분석 중... ({i+1}/{len(files)})")
+        filename = os.path.basename(file.name) if hasattr(file, 'name') else f"파일_{i+1}"
+        output += f"## 📎 {filename}\n\n"
+        yield output
+        
+        try:
+            text, error = extract_text_from_file(file.name if hasattr(file, 'name') else file)
+            if text:
+                all_extracted_text.append({"filename": filename, "text": text})
+                preview = text[:500] + "..." if len(text) > 500 else text
+                output += f"✅ 텍스트 추출 성공 ({len(text):,}자)\n\n```\n{preview}\n```\n\n"
+            else:
+                output += f"⚠️ 텍스트 추출 실패: {error}\n\n"
+        except Exception as e:
+            output += f"❌ 오류: {str(e)}\n\n"
+        yield output
+    
+    if all_extracted_text:
+        output += "---\n\n## 🤖 AI 기업정보 추출\n\n"
+        yield output
+        
+        combined_text = "\n\n".join([f"[{item['filename']}]\n{item['text'][:3000]}" for item in all_extracted_text])
+        
+        system_prompt = """당신은 기업 서류 분석 전문가입니다.
+주어진 문서들에서 다음 정보를 추출해주세요:
+1. 사업자 정보 (사업자등록번호, 법인등록번호, 상호, 대표자, 설립일, 주소, 업종)
+2. 재무 정보 (자본금, 매출액, 영업이익, 당기순이익)
+3. 인력 정보 (상시근로자 수, 4대보험 가입자 수)
+4. 인증/등록 정보 (부설연구소, 벤처기업 인증 등)
+5. 기타 특이사항
+JSON 형식으로 정리해주세요."""
+        
+        messages = [
+            {"role": "system", "content": system_prompt},
+            {"role": "user", "content": f"다음 기업 서류들을 분석해주세요:\n\n{combined_text[:12000]}"}
+        ]
+        
+        for chunk in call_groq_api_stream(messages):
+            output += chunk
+            yield output
+    
+    output += "\n\n---\n✅ **문서 분석 완료**"
+    yield output
+
+
+def match_announcements_with_profile(profile_data, announcements_df, progress=gr.Progress()):
+    """기업 프로필과 공고를 매칭"""
+    if not profile_data:
+        yield "❌ 기업 프로필을 먼저 입력해주세요."
+        return
+    
+    if announcements_df is None or (isinstance(announcements_df, pd.DataFrame) and announcements_df.empty):
+        yield "❌ 매칭할 공고 데이터가 없습니다. 먼저 공고를 검색해주세요."
+        return
+    
+    output = "# 🎯 맞춤 과제 매칭 결과\n\n"
+    output += "## 📋 입력된 기업 프로필\n\n"
+    output += f"```json\n{json.dumps(profile_data, ensure_ascii=False, indent=2)[:2000]}\n```\n\n"
+    output += "---\n\n## 🔍 AI 매칭 분석 중...\n\n"
+    yield output
+    
+    announcements_text = ""
+    df_to_use = announcements_df if isinstance(announcements_df, pd.DataFrame) else pd.DataFrame()
+    for idx, row in df_to_use.head(20).iterrows():
+        announcements_text += f"""
+### {row.get('지원사업명', '')}
+- 지원분야: {row.get('지원분야', '')}
+- 소관부처: {row.get('소관부처', '')}
+- 신청기간: {row.get('신청기간', '')}
+- 지원대상: {row.get('지원대상', '')}
+---
+"""
+    
+    system_prompt = """당신은 정부 지원사업 매칭 전문가입니다.
+기업 프로필과 공고 목록을 분석하여 신청 가능한 과제를 추천해주세요.
+각 공고에 대해:
+- ✅ 적합: 신청 자격 충족
+- ⚠️ 확인필요: 일부 조건 확인 필요
+- ❌ 부적합: 자격 미달
+추천 순위와 이유를 설명해주세요."""
+    
+    messages = [
+        {"role": "system", "content": system_prompt},
+        {"role": "user", "content": f"기업 프로필:\n{json.dumps(profile_data, ensure_ascii=False)}\n\n공고 목록:\n{announcements_text[:8000]}"}
+    ]
+    
+    for chunk in call_groq_api_stream(messages):
+        output += chunk
+        yield output
+    
+    output += "\n\n---\n✅ **매칭 분석 완료**"
+    yield output
+
+
+# ============================================================
+# CSS 스타일
+# ============================================================
+CUSTOM_CSS = """
+@import url('https://fonts.googleapis.com/css2?family=Noto+Sans+KR:wght@400;500;700&display=swap');
+.gradio-container {
+    font-family: 'Noto Sans KR', sans-serif !important;
+    background: linear-gradient(135deg, #f5f7fa 0%, #e8eef5 100%) !important;
+    max-width: 1600px !important;
+    margin: 0 auto !important;
+}
+.header-banner {
+    background: linear-gradient(135deg, #1a5cb0 0%, #0d4a94 100%);
+    color: white;
+    padding: 24px 28px;
+    border-radius: 16px;
+    margin-bottom: 20px;
+    box-shadow: 0 4px 20px rgba(26, 92, 176, 0.3);
+}
+.section-header {
+    background: linear-gradient(135deg, #667eea 0%, #764ba2 100%);
+    color: white;
+    padding: 12px 16px;
+    border-radius: 8px;
+    margin: 16px 0 12px 0;
+    font-weight: 600;
+}
+.feature-badge {
+    display: inline-block;
+    background: rgba(255,255,255,0.2);
+    padding: 4px 12px;
+    border-radius: 20px;
+    font-size: 12px;
+    margin: 4px 4px 0 0;
+}
+.analyze-btn {
+    background: linear-gradient(135deg, #10B981 0%, #059669 100%) !important;
+    border: none !important;
+    color: white !important;
+    font-weight: 600 !important;
+}
+.analyze-btn:hover {
+    background: linear-gradient(135deg, #059669 0%, #047857 100%) !important;
+}
+.match-btn {
+    background: linear-gradient(135deg, #8B5CF6 0%, #7C3AED 100%) !important;
+    border: none !important;
+    color: white !important;
+    font-weight: 600 !important;
+}
+.match-btn:hover {
+    background: linear-gradient(135deg, #7C3AED 0%, #6D28D9 100%) !important;
+}
+.analysis-output {
+    background: white !important;
+    border: 2px solid #e5e7eb !important;
+    border-radius: 12px !important;
+    padding: 20px !important;
+    min-height: 500px !important;
+    max-height: 700px !important;
+    overflow-y: auto !important;
+    font-size: 14px !important;
+    line-height: 1.8 !important;
+}
+.filter-info {
+    background: #FEF3C7;
+    border: 1px solid #F59E0B;
+    border-radius: 8px;
+    padding: 12px;
+    margin-bottom: 16px;
+    font-size: 13px;
+    color: #92400E;
+}
+.footer-text {
+    text-align: center;
+    color: #868e96;
+    font-size: 13px;
+    margin-top: 16px;
+}
+.profile-section {
+    border: 1px solid #e5e7eb;
+    border-radius: 12px;
+    padding: 16px;
+    margin-bottom: 16px;
+    background: white;
+}
+"""
+
+
+# ============================================================
+# 메인 인터페이스
+# ============================================================
+def create_interface():
+    with gr.Blocks(title="기업마당 AI 분석기", css=CUSTOM_CSS) as demo:
+        gr.HTML("""
+        <div class="header-banner">
+            <div style="display: flex; align-items: center; gap: 16px;">
+                <div style="width: 56px; height: 56px; background: white; border-radius: 14px; display: flex; align-items: center; justify-content: center; font-size: 28px;">🏢</div>
+                <div>
+                    <h1 style="margin: 0; font-size: 24px;">기업마당 지원사업 AI 분석기</h1>
+                    <p style="margin: 4px 0 0 0; opacity: 0.9;">공고 검색 · 첨부파일 자동 분석 · AI 요약 · 맞춤 과제 추출</p>
+                    <div style="margin-top: 8px;">
+                        <span class="feature-badge">📄 HWP/HWPX 지원</span>
+                        <span class="feature-badge">🤖 AI 분석</span>
+                        <span class="feature-badge">🎯 맞춤 매칭</span>
+                    </div>
+                </div>
+            </div>
+        </div>
+        """)
+        
+        # 상태 변수
+        selected_url = gr.State("")
+        selected_name = gr.State("")
+        selected_attachments = gr.State([])
+        selected_description = gr.State("")
+        current_df = gr.State(value=pd.DataFrame())
+        company_profile = gr.State(value={})
+        
+        with gr.Tabs():
+            # 탭 1: 공고 검색
+            with gr.Tab("🔍 공고 검색"):
+                with gr.Row():
+                    keyword_input = gr.Textbox(label="🔍 검색어", placeholder="예: AI, 스타트업, R&D", scale=3)
+                    category_dropdown = gr.Dropdown(label="📂 지원분야", choices=list(CATEGORY_CODES.keys()), value="전체", scale=1)
+                    region_dropdown = gr.Dropdown(label="📍 지역", choices=REGION_LIST, value="전체(지역)", scale=1)
+                with gr.Row():
+                    org_type_dropdown = gr.Dropdown(label="🏛️ 기관유형", choices=ORG_TYPE_OPTIONS, value="전체", scale=1)
+                    sort_dropdown = gr.Dropdown(label="📊 정렬", choices=SORT_OPTIONS, value="등록일순", scale=1)
+                    status_dropdown = gr.Dropdown(label="📌 공고상태", choices=STATUS_OPTIONS, value="진행중", scale=1)
+                with gr.Row():
+                    page_input = gr.Number(label="📄 페이지", value=1, minimum=1, scale=1)
+                    rows_dropdown = gr.Dropdown(label="📊 표시개수", choices=[10, 15, 20, 30, 50], value=20, scale=1)
+                    search_btn = gr.Button("🔎 검색", variant="primary", scale=2)
+                
+                status_output = gr.Textbox(label="📊 조회 결과", interactive=False)
+                results_output = gr.Dataframe(label="📋 공고 목록 (행 클릭으로 선택)", wrap=True, interactive=False)
+                with gr.Row():
+                    prev_btn = gr.Button("◀️ 이전", size="sm")
+                    next_btn = gr.Button("다음 ▶️", size="sm")
+                    export_btn = gr.Button("📥 CSV 내보내기", size="sm", variant="secondary")
+                csv_output = gr.File(label="📁 다운로드", visible=True)
+            
+            # 탭 2: AI 분석
+            with gr.Tab("🤖 AI 분석"):
+                with gr.Row():
+                    with gr.Column(scale=1):
+                        selected_info = gr.Textbox(label="📌 선택된 공고", placeholder="공고 검색 탭에서 선택", lines=3, interactive=False)
+                        analyze_btn = gr.Button("🚀 AI 분석 시작", variant="primary", size="lg", elem_classes=["analyze-btn"])
+                    with gr.Column(scale=2):
+                        analysis_output = gr.Markdown(value="### 📊 분석 결과\n\n*공고를 선택하고 분석 버튼을 클릭하세요*", height=500)
+            
+            # 탭 3: 맞춤 과제 추출
+            with gr.Tab("🎯 맞춤 과제 추출"):
+                gr.HTML("""
+                <div style="background: linear-gradient(135deg, #667eea 0%, #764ba2 100%); color: white; padding: 20px; border-radius: 12px; margin-bottom: 20px;">
+                    <h2 style="margin: 0 0 8px 0;">🎯 나만의 맞춤 과제 추출</h2>
+                    <p style="margin: 0; opacity: 0.9;">기업 정보를 입력하고 문서를 업로드하면 AI가 신청 가능한 과제를 자동으로 매칭해드립니다.</p>
+                </div>
+                """)
+                
+                with gr.Tabs():
+                    # 서브탭 1: 기업 기본정보
+                    with gr.Tab("1️⃣ 기업 기본정보"):
+                        with gr.Row():
+                            with gr.Column():
+                                gr.HTML('<div class="section-header">📋 사업자 정보</div>')
+                                biz_number = gr.Textbox(label="사업자등록번호", placeholder="000-00-00000")
+                                corp_number = gr.Textbox(label="법인등록번호", placeholder="000000-0000000")
+                                company_name = gr.Textbox(label="상호/법인명", placeholder="(주)회사명")
+                                establish_date = gr.Textbox(label="설립일자", placeholder="YYYY-MM-DD")
+                                company_type = gr.Dropdown(label="기업형태", choices=COMPANY_TYPE_OPTIONS, value="법인사업자")
+                                corp_type = gr.Dropdown(label="법인 종류", choices=CORP_TYPE_OPTIONS, value="주식회사")
+                                company_size = gr.Dropdown(label="기업규모", choices=COMPANY_SIZE_OPTIONS, value="소기업")
+                            
+                            with gr.Column():
+                                gr.HTML('<div class="section-header">🏆 인증 현황</div>')
+                                venture_cert = gr.Checkbox(label="벤처기업 인증")
+                                innobiz_cert = gr.Checkbox(label="이노비즈 인증")
+                                mainbiz_cert = gr.Checkbox(label="메인비즈 인증")
+                                sme_cert = gr.Checkbox(label="중소기업확인서 보유")
+                                small_biz_cert = gr.Checkbox(label="소상공인확인서 보유")
+                                social_venture = gr.Checkbox(label="소셜벤처")
+                                startup_flag = gr.Checkbox(label="스타트업")
+                            
+                            with gr.Column():
+                                gr.HTML('<div class="section-header">📍 소재지 정보</div>')
+                                hq_sido = gr.Dropdown(label="본사 소재지 (시/도)", choices=SIDO_LIST, value="서울특별시")
+                                hq_sigungu = gr.Textbox(label="본사 소재지 (시/군/구)", placeholder="예: 강남구")
+                                innovation_city = gr.Checkbox(label="혁신도시 입주")
+                                industrial_complex = gr.Checkbox(label="산업단지 입주")
+                                free_zone = gr.Checkbox(label="규제자유특구 소재")
+                                non_capital = gr.Checkbox(label="비수도권 (지방기업)")
+                                
+                                gr.HTML('<div class="section-header">📊 업종 정보</div>')
+                                industry_major = gr.Dropdown(label="주업종 (대분류)", choices=INDUSTRY_MAJOR_OPTIONS, value="C. 제조업")
+                                is_manufacturing = gr.Checkbox(label="제조업 여부")
+                                is_knowledge_service = gr.Checkbox(label="지식서비스업 여부")
+                    
+                    # 서브탭 2: 대표자/인력 정보
+                    with gr.Tab("2️⃣ 대표자/인력 정보"):
+                        with gr.Row():
+                            with gr.Column():
+                                gr.HTML('<div class="section-header">👤 대표자 정보</div>')
+                                ceo_name = gr.Textbox(label="대표자명", placeholder="홍길동")
+                                ceo_gender = gr.Radio(label="대표자 성별", choices=["남성", "여성"], value="남성")
+                                ceo_birthdate = gr.Textbox(label="대표자 생년월일", placeholder="YYYY-MM-DD")
+                                youth_ceo = gr.Checkbox(label="청년창업자 (만39세 미만)")
+                                senior_ceo = gr.Checkbox(label="시니어창업자 (만40세 이상)")
+                                women_company = gr.Checkbox(label="여성기업확인서 보유")
+                                disabled_company = gr.Checkbox(label="장애인기업확인서 보유")
+                            
+                            with gr.Column():
+                                gr.HTML('<div class="section-header">👥 고용 현황</div>')
+                                insurance_employees = gr.Number(label="4대보험 가입자 수", value=0, minimum=0)
+                                regular_employees = gr.Number(label="상시근로자 수", value=0, minimum=0)
+                                youth_employees = gr.Number(label="청년고용 인원", value=0, minimum=0)
+                                female_ratio = gr.Slider(label="여성고용 비율 (%)", minimum=0, maximum=100, value=0)
+                                new_hire_plan = gr.Number(label="신규채용 계획 (명)", value=0, minimum=0)
+                            
+                            with gr.Column():
+                                gr.HTML('<div class="section-header">🔬 연구인력/역량</div>')
+                                rd_personnel = gr.Number(label="연구인력 수", value=0, minimum=0)
+                                phd_researchers = gr.Number(label="박사급 연구원", value=0, minimum=0)
+                                research_center = gr.Checkbox(label="기업부설연구소 등록")
+                                rd_dept = gr.Checkbox(label="연구개발전담부서 등록")
+                                patent_count = gr.Number(label="보유 특허 수", value=0, minimum=0)
+                    
+                    # 서브탭 3: 재무 정보
+                    with gr.Tab("3️⃣ 재무 정보"):
+                        with gr.Row():
+                            with gr.Column():
+                                gr.HTML('<div class="section-header">💰 매출 및 수익</div>')
+                                revenue_current = gr.Number(label="최근년도 매출액 (백만원)", value=0, minimum=0)
+                                revenue_prev = gr.Number(label="전년도 매출액 (백만원)", value=0, minimum=0)
+                                operating_profit = gr.Number(label="영업이익 (백만원)", value=0)
+                                net_income = gr.Number(label="당기순이익 (백만원)", value=0)
+                                export_amount = gr.Number(label="수출액 (천달러)", value=0, minimum=0)
+                            
+                            with gr.Column():
+                                gr.HTML('<div class="section-header">📊 재무건전성</div>')
+                                capital = gr.Number(label="자본금 (백만원)", value=0, minimum=0)
+                                total_assets = gr.Number(label="자산총계 (백만원)", value=0, minimum=0)
+                                debt_ratio = gr.Slider(label="부채비율 (%)", minimum=0, maximum=500, value=0)
+                                credit_grade = gr.Dropdown(label="신용등급", choices=CREDIT_GRADE_OPTIONS, value="미평가")
+                                tcb_grade = gr.Dropdown(label="TCB 등급", choices=TCB_GRADE_OPTIONS, value="미평가")
+                                capital_impairment = gr.Checkbox(label="자본잠식 여부")
+                            
+                            with gr.Column():
+                                gr.HTML('<div class="section-header">🔬 R&D 투자</div>')
+                                rd_investment = gr.Number(label="연간 R&D 투자액 (백만원)", value=0, minimum=0)
+                                gov_project_exp = gr.Checkbox(label="정부과제 수행 경험")
+                                gov_support_3yr = gr.Number(label="최근 3년 정부지원금 (백만원)", value=0, minimum=0)
+                    
+                    # 서브탭 4: 기술분야/제한사항
+                    with gr.Tab("4️⃣ 기술분야/제한사항"):
+                        with gr.Row():
+                            with gr.Column():
+                                gr.HTML('<div class="section-header">🔬 기술 분야</div>')
+                                core_industry = gr.CheckboxGroup(label="10대 핵심산업", choices=CORE_INDUSTRY_OPTIONS)
+                                strategic_tech = gr.CheckboxGroup(label="12대 국가전략기술", choices=NATIONAL_STRATEGIC_TECH)
+                                green_tech = gr.Checkbox(label="녹색기술 분야")
+                                digital_transform = gr.Checkbox(label="디지털전환 분야")
+                                defense_industry = gr.Checkbox(label="국방/방산 분야")
+                            
+                            with gr.Column():
+                                gr.HTML('<div class="section-header">📜 인증/ISO</div>')
+                                iso_certs = gr.CheckboxGroup(label="ISO 인증", choices=ISO_CERT_OPTIONS)
+                                gmp_cert = gr.Checkbox(label="GMP 인증")
+                            
+                            with gr.Column():
+                                gr.HTML('<div class="section-header">⚠️ 결격사유 확인</div>')
+                                tax_delinquent = gr.Checkbox(label="국세 체납")
+                                local_tax_delinquent = gr.Checkbox(label="지방세 체납")
+                                gov_project_fail = gr.Checkbox(label="정부과제 불성실")
+                                bankruptcy = gr.Checkbox(label="휴/폐업 이력")
+                                financial_default = gr.Checkbox(label="금융기관 연체")
+                    
+                    # 서브탭 5: 문서 업로드 및 매칭
+                    with gr.Tab("5️⃣ 문서 업로드 & 매칭"):
+                        gr.HTML("""
+                        <div style="background: #EBF5FF; border: 1px solid #3B82F6; border-radius: 8px; padding: 16px; margin-bottom: 16px;">
+                            <h3 style="margin: 0 0 8px 0; color: #1E40AF;">📁 문서 업로드</h3>
+                            <p style="margin: 0; color: #1E3A8A;">사업자등록증, 등기부등본, 재무제표, 중소기업확인서 등을 업로드하면 AI가 자동으로 정보를 추출합니다.</p>
+                        </div>
+                        """)
+                        
+                        with gr.Row():
+                            with gr.Column(scale=1):
+                                file_upload = gr.File(
+                                    label="📎 문서 업로드 (HWP, PDF, TXT, XLSX)",
+                                    file_count="multiple",
+                                    file_types=[".hwp", ".hwpx", ".pdf", ".txt", ".xlsx", ".xls"]
+                                )
+                                analyze_docs_btn = gr.Button("📄 문서 분석", variant="secondary", size="lg")
+                            
+                            with gr.Column(scale=2):
+                                doc_analysis_output = gr.Markdown(value="### 📄 문서 분석 결과\n\n*문서를 업로드하고 분석 버튼을 클릭하세요*", height=400)
+                        
+                        gr.HTML('<hr style="margin: 24px 0;">')
+                        
+                        with gr.Row():
+                            save_profile_btn = gr.Button("💾 프로필 저장", variant="secondary", size="lg")
+                            match_btn = gr.Button("🎯 맞춤 과제 매칭 시작", variant="primary", size="lg", elem_classes=["match-btn"])
+                        
+                        profile_status = gr.Textbox(label="프로필 저장 상태", interactive=False)
+                        match_output = gr.Markdown(value="### 🎯 매칭 결과\n\n*프로필을 저장하고 매칭 버튼을 클릭하세요*", height=500)
+        
+        # ============================================================
+        # 이벤트 핸들러
+        # ============================================================
+        def search_fn(keyword, category, region, org_type, sort_by, status_filter, page, rows):
+            df, status = fetch_announcements(keyword or "", category or "전체", region or "전체(지역)",
+                                             org_type or "전체", sort_by or "등록일순",
+                                             status_filter or "진행중", int(page) if page else 1, int(rows) if rows else 20)
+            display_cols = ["번호", "지원분야", "지원사업명", "신청기간", "소관부처", "등록일", "지원대상"]
+            display_df = df[[c for c in display_cols if c in df.columns]] if not df.empty else df
+            return display_df, status, df
+        
+        def prev_fn(page, keyword, category, region, org_type, sort_by, status_filter, rows):
+            new_page = max(1, int(page) - 1) if page else 1
+            df, status = fetch_announcements(keyword or "", category or "전체", region or "전체(지역)",
+                                             org_type or "전체", sort_by or "등록일순",
+                                             status_filter or "진행중", new_page, int(rows) if rows else 20)
+            display_cols = ["번호", "지원분야", "지원사업명", "신청기간", "소관부처", "등록일", "지원대상"]
+            display_df = df[[c for c in display_cols if c in df.columns]] if not df.empty else df
+            return display_df, status, df, new_page
+        
+        def next_fn(page, keyword, category, region, org_type, sort_by, status_filter, rows):
+            new_page = int(page) + 1 if page else 2
+            df, status = fetch_announcements(keyword or "", category or "전체", region or "전체(지역)",
+                                             org_type or "전체", sort_by or "등록일순",
+                                             status_filter or "진행중", new_page, int(rows) if rows else 20)
+            display_cols = ["번호", "지원분야", "지원사업명", "신청기간", "소관부처", "등록일", "지원대상"]
+            display_df = df[[c for c in display_cols if c in df.columns]] if not df.empty else df
+            return display_df, status, df, new_page
+        
+        def export_to_csv(df):
+            if df is None or (isinstance(df, pd.DataFrame) and df.empty):
+                return None
+            filepath = "/tmp/bizinfo_announcements.csv"
+            export_cols = [c for c in df.columns if not c.startswith("_")]
+            df[export_cols].to_csv(filepath, index=False, encoding='utf-8-sig')
+            return filepath
+        
+        def on_row_select(evt: gr.SelectData, df):
+            if evt.index[0] < len(df):
+                row = df.iloc[evt.index[0]]
+                url = row.get("상세링크", "")
+                name = row.get("지원사업명", "")
+                attachments = row.get("첨부파일", [])
+                description = row.get("사업개요", "")
+                att_info = ""
+                if attachments and len(attachments) > 0:
+                    att_info = f"\n\n📎 첨부파일 {len(attachments)}개:"
+                    for att in attachments:
+                        att_info += f"\n  - {att.get('filename', '파일')}"
+                info = f"📌 {name}\n\n🔗 {url}{att_info}"
+                return url, name, attachments, description, info
+            return "", "", [], "", ""
+        
+        def save_profile_fn(biz_num, corp_num, comp_name, est_date, comp_type, corp_tp, comp_size,
+                           venture, innobiz, mainbiz, sme, small_biz, social, startup,
+                           sido, sigungu, innov_city, ind_complex, free_z, non_cap,
+                           ind_major, is_manu, is_know,
+                           ceo_nm, ceo_gen, ceo_birth, youth, senior, women, disabled,
+                           ins_emp, reg_emp, youth_emp, fem_ratio, new_hire,
+                           rd_per, phd, res_ctr, rd_dep, patent,
+                           rev_cur, rev_prev, op_profit, net_inc, export,
+                           cap, assets, debt, credit, tcb, impair,
+                           rd_inv, gov_exp, gov_sup,
+                           core_ind, strat_tech, green, digital, defense,
+                           iso, gmp,
+                           tax_del, local_tax, gov_fail, bankrupt, fin_def):
+            
+            profile = {
+                "사업자정보": {
+                    "사업자등록번호": biz_num, "법인등록번호": corp_num, "상호": comp_name,
+                    "설립일자": est_date, "기업형태": comp_type, "법인종류": corp_tp, "기업규모": comp_size
+                },
+                "인증현황": {
+                    "벤처기업": venture, "이노비즈": innobiz, "메인비즈": mainbiz,
+                    "중소기업확인서": sme, "소상공인확인서": small_biz, "소셜벤처": social, "스타트업": startup
+                },
+                "소재지": {
+                    "시도": sido, "시군구": sigungu, "혁신도시": innov_city,
+                    "산업단지": ind_complex, "규제자유특구": free_z, "비수도권": non_cap
+                },
+                "업종": {"대분류": ind_major, "제조업": is_manu, "지식서비스업": is_know},
+                "대표자": {
+                    "이름": ceo_nm, "성별": ceo_gen, "생년월일": ceo_birth,
+                    "청년창업자": youth, "시니어창업자": senior, "여성기업": women, "장애인기업": disabled
+                },
+                "고용현황": {
+                    "4대보험가입자": ins_emp, "상시근로자": reg_emp, "청년고용": youth_emp,
+                    "여성비율": fem_ratio, "신규채용계획": new_hire
+                },
+                "연구역량": {
+                    "연구인력": rd_per, "박사급": phd, "부설연구소": res_ctr,
+                    "전담부서": rd_dep, "특허수": patent
+                },
+                "재무정보": {
+                    "매출액_당해": rev_cur, "매출���_전년": rev_prev, "영업이익": op_profit,
+                    "당기순이익": net_inc, "수출액": export, "자본금": cap, "자산총계": assets,
+                    "부채비율": debt, "신용등급": credit, "TCB등급": tcb, "자본잠식": impair
+                },
+                "R&D투자": {"연간투자액": rd_inv, "정부과제경험": gov_exp, "최근3년지원금": gov_sup},
+                "기술분야": {
+                    "핵심산업": core_ind, "국가전략기술": strat_tech,
+                    "녹색기술": green, "디지털전환": digital, "국방방산": defense
+                },
+                "인증": {"ISO": iso, "GMP": gmp},
+                "결격사유": {
+                    "국세체납": tax_del, "지방세체납": local_tax, "불성실이력": gov_fail,
+                    "휴폐업": bankrupt, "금융연체": fin_def
+                }
+            }
+            return profile, "✅ 프로필이 저장되었습니다."
+        
+        # 이벤트 연결
+        search_btn.click(fn=search_fn, inputs=[keyword_input, category_dropdown, region_dropdown, org_type_dropdown,
+                                                sort_dropdown, status_dropdown, page_input, rows_dropdown],
+                        outputs=[results_output, status_output, current_df])
+        
+        keyword_input.submit(fn=search_fn, inputs=[keyword_input, category_dropdown, region_dropdown, org_type_dropdown,
+                                                    sort_dropdown, status_dropdown, page_input, rows_dropdown],
+                            outputs=[results_output, status_output, current_df])
+        
+        # 페이지네이션 이벤트
+        prev_btn.click(fn=prev_fn, inputs=[page_input, keyword_input, category_dropdown, region_dropdown, 
+                                           org_type_dropdown, sort_dropdown, status_dropdown, rows_dropdown],
+                      outputs=[results_output, status_output, current_df, page_input])
+        
+        next_btn.click(fn=next_fn, inputs=[page_input, keyword_input, category_dropdown, region_dropdown,
+                                           org_type_dropdown, sort_dropdown, status_dropdown, rows_dropdown],
+                      outputs=[results_output, status_output, current_df, page_input])
+        
+        # CSV 내보내기
+        export_btn.click(fn=export_to_csv, inputs=[current_df], outputs=[csv_output])
+        
+        results_output.select(fn=on_row_select, inputs=[current_df],
+                             outputs=[selected_url, selected_name, selected_attachments, selected_description, selected_info])
+        
+        analyze_btn.click(fn=analyze_announcement, inputs=[selected_url, selected_name, selected_attachments, selected_description],
+                         outputs=[analysis_output])
+        
+        analyze_docs_btn.click(fn=analyze_uploaded_documents, inputs=[file_upload], outputs=[doc_analysis_output])
+        
+        save_profile_btn.click(
+            fn=save_profile_fn,
+            inputs=[biz_number, corp_number, company_name, establish_date, company_type, corp_type, company_size,
+                   venture_cert, innobiz_cert, mainbiz_cert, sme_cert, small_biz_cert, social_venture, startup_flag,
+                   hq_sido, hq_sigungu, innovation_city, industrial_complex, free_zone, non_capital,
+                   industry_major, is_manufacturing, is_knowledge_service,
+                   ceo_name, ceo_gender, ceo_birthdate, youth_ceo, senior_ceo, women_company, disabled_company,
+                   insurance_employees, regular_employees, youth_employees, female_ratio, new_hire_plan,
+                   rd_personnel, phd_researchers, research_center, rd_dept, patent_count,
+                   revenue_current, revenue_prev, operating_profit, net_income, export_amount,
+                   capital, total_assets, debt_ratio, credit_grade, tcb_grade, capital_impairment,
+                   rd_investment, gov_project_exp, gov_support_3yr,
+                   core_industry, strategic_tech, green_tech, digital_transform, defense_industry,
+                   iso_certs, gmp_cert,
+                   tax_delinquent, local_tax_delinquent, gov_project_fail, bankruptcy, financial_default],
+            outputs=[company_profile, profile_status]
+        )
+        
+        match_btn.click(fn=match_announcements_with_profile, inputs=[company_profile, current_df], outputs=[match_output])
+    
+    return demo
+
+
+if __name__ == "__main__":
+    demo = create_interface()
+    demo.launch(server_name="0.0.0.0", server_port=7860, ssr_mode=False)