import asyncio import json import tempfile import os import edge_tts import gradio as gr from src.inference import MediVisionPipeline from src.model_loader import check_connection # --------------------------------------------------------------------------- # Pipeline singleton # --------------------------------------------------------------------------- _pipeline: MediVisionPipeline | None = None def get_pipeline() -> MediVisionPipeline: global _pipeline if _pipeline is None: _pipeline = MediVisionPipeline() return _pipeline # --------------------------------------------------------------------------- # Language config # --------------------------------------------------------------------------- _LANG_MAP = { "English": "en", "Tiếng Việt": "vn", "中文": "zh", "Español": "es", "Français": "fr", "日本語": "ja", } _LANG_CHOICES = list(_LANG_MAP.keys()) _LANG_BCP47 = { "en": "en-US", "vn": "vi-VN", "zh": "zh-CN", "es": "es-ES", "fr": "fr-FR", "ja": "ja-JP", } # edge-tts voice for each language (natural, neural voices) _TTS_VOICE = { "en": "en-US-JennyNeural", "vn": "vi-VN-HoaiMyNeural", "zh": "zh-CN-XiaoxiaoNeural", "es": "es-ES-ElviraNeural", "fr": "fr-FR-DeniseNeural", "ja": "ja-JP-NanamiNeural", } def _tts_generate(text: str, lang: str) -> str | None: """Generate TTS audio with edge-tts; returns temp file path or None on error.""" if not text or not text.strip(): return None voice = _TTS_VOICE.get(lang, _TTS_VOICE["en"]) try: tmp = tempfile.NamedTemporaryFile(delete=False, suffix=".mp3") tmp.close() async def _run(): communicate = edge_tts.Communicate(text.strip(), voice) await communicate.save(tmp.name) asyncio.run(_run()) return tmp.name except Exception: return None _I18N = { "en": { "img_label": "Upload Medical Image", "symptoms_label": "Symptoms Description", "symptoms_placeholder": "Describe what you feel — e.g. itchy red patch for 3 days, slight burning sensation...", "analyze_btn": "🔬 Analyze", "output_label": "Analysis Result", "empty_output": "Upload an image and/or describe symptoms, then click Analyze.", "diag_label": "Diagnosis Suggestion", "severity_label": "Severity", "actions_label": "Assessment & Guidance", "confidence_label": "Confidence Score", "disclaimer": "This is an AI assistant, not a licensed physician. Always consult a healthcare professional for serious conditions.", "placeholder": "Please upload an image or enter symptoms.", "backend_live": "AMD Cloud · Live", "backend_offline": "AMD Cloud · Offline", "error_title": "Backend Unavailable", "error_body": "AMD Cloud backend is unreachable. Please try again later.", "examples_label": "Quick Examples", "metrics_latency": "Latency", "metrics_throughput": "Throughput", "metrics_tokens": "tokens", "region_label": "Affected Body Region", "region_none": "Not specified", "region_optional_label": "Affected Body Region (Optional)", "input_hint": "Provide an image, describe symptoms, or both — at least one is required. Body region is optional.", "map_label": "Anatomical Map", "map_select": "click to select", "map_selected": "{n} region(s) selected", "img_mode_label": "Upload Mode", "img_mode_standard": "Standard (1 image)", "img_mode_compare": "Compare (2 images — before & after)", "img_label_day1": "Medical Image (Day 1)", "img_label_dayx": "Comparison Image (Day X)", "tab_patient": "Patient View", "tab_doctor": "Export for Doctor (SOAP)", "critical_warning": "⚠️ CRITICAL: Severe symptoms detected. Please visit a medical facility within 24 hours.", "conditions_label": "Possible Conditions", "soap_copy_btn": "Copy SOAP Note", "soap_empty": "Run an analysis to generate the SOAP note.", "tts_btn": "🔊 Listen", "chat_label": "Follow-up Questions", "chat_placeholder": "Ask a follow-up question about your condition...", "chat_send": "Send", "red_flags_label": "Red Flags", "watch_for_label": "Watch For", "analyzing_label": "Analyzing...", }, "vn": { "img_label": "Tải lên hình ảnh y tế", "symptoms_label": "Mô tả triệu chứng", "symptoms_placeholder": "Mô tả những gì bạn cảm thấy — ví dụ: vết đỏ ngứa 3 ngày, hơi rát...", "analyze_btn": "🔬 Phân tích", "output_label": "Kết quả phân tích", "empty_output": "Tải ảnh và/hoặc mô tả triệu chứng, rồi nhấn Phân tích.", "diag_label": "Gợi ý chẩn đoán", "severity_label": "Mức độ nghiêm trọng", "actions_label": "Đánh giá & Hướng dẫn", "confidence_label": "Độ tin cậy", "disclaimer": "Đây là trợ lý AI, không phải bác sĩ. Hãy tham khảo chuyên gia y tế cho các tình trạng nghiêm trọng.", "placeholder": "Vui lòng tải lên hình ảnh hoặc nhập triệu chứng.", "backend_live": "AMD Cloud · Trực tuyến", "backend_offline": "AMD Cloud · Ngoại tuyến", "error_title": "Hệ thống không khả dụng", "error_body": "Không thể kết nối AMD Cloud. Vui lòng thử lại sau.", "examples_label": "Ví dụ nhanh", "metrics_latency": "Độ trễ", "metrics_throughput": "Thông lượng", "metrics_tokens": "token", "region_label": "Vùng cơ thể bị ảnh hưởng", "region_none": "Không xác định", "region_optional_label": "Vùng cơ thể bị ảnh hưởng (Không bắt buộc)", "input_hint": "Cung cấp ảnh, mô tả triệu chứng hoặc cả hai — cần ít nhất một trong hai. Vùng cơ thể là tùy chọn.", "map_label": "Bản đồ giải phẫu", "map_select": "nhấn để chọn", "map_selected": "{n} vùng đã chọn", "img_mode_label": "Chế độ tải ảnh", "img_mode_standard": "Tiêu chuẩn (1 ảnh)", "img_mode_compare": "So sánh (2 ảnh — trước & sau)", "img_label_day1": "Ảnh y tế (Ngày 1)", "img_label_dayx": "Ảnh so sánh (Ngày X)", "tab_patient": "Dành cho bệnh nhân", "tab_doctor": "Xuất cho bác sĩ (SOAP)", "critical_warning": "⚠️ CẢNH BÁO: Triệu chứng nghiêm trọng được phát hiện. Vui lòng đến cơ sở y tế trong vòng 24 giờ.", "conditions_label": "Tình trạng có thể", "soap_copy_btn": "Sao chép SOAP", "soap_empty": "Thực hiện phân tích để tạo ghi chú SOAP.", "tts_btn": "🔊 Nghe", "chat_label": "Câu hỏi tiếp theo", "chat_placeholder": "Đặt câu hỏi tiếp theo về tình trạng của bạn...", "chat_send": "Gửi", "red_flags_label": "Cảnh báo", "watch_for_label": "Theo dõi", "analyzing_label": "Đang phân tích...", }, "zh": { "img_label": "上传医学图像", "symptoms_label": "症状描述", "symptoms_placeholder": "描述您的感受 — 例如：前臂红色瘙痒皮疹持续3天，略有灼热感...", "analyze_btn": "🔬 分析", "output_label": "分析结果", "empty_output": "请上传图片和/或描述症状，然后点击分析。", "diag_label": "诊断建议", "severity_label": "严重程度", "actions_label": "评估与指导", "confidence_label": "置信度", "disclaimer": "本工具为AI助手，不能替代执业医师。如有严重病情，请务必咨询专业医疗人员。", "placeholder": "请上传图片或输入症状描述。", "backend_live": "AMD Cloud · 在线", "backend_offline": "AMD Cloud · 离线", "error_title": "后端不可用", "error_body": "AMD Cloud 后端无法访问，请稍后重试。", "examples_label": "快速示例", "metrics_latency": "延迟", "metrics_throughput": "吞吐量", "metrics_tokens": "tokens", "region_label": "受影响的身体部位", "region_none": "未指定", "region_optional_label": "受影响的身体部位（可选）", "input_hint": "请上传图片或描述症状（至少提供其中一项）。身体部位为可选项。", "map_label": "解剖图", "map_select": "点击选择", "map_selected": "已选 {n} 个部位", "img_mode_label": "上传模式", "img_mode_standard": "标准模式（1张图片）", "img_mode_compare": "对比模式（2张图片 — 前后对比）", "img_label_day1": "医疗图像（第1天）", "img_label_dayx": "对比图像（第X天）", "tab_patient": "患者视图", "tab_doctor": "导出给医生（SOAP）", "critical_warning": "⚠️ 严重警告：检测到严重症状。请在24小时内前往医疗机构就诊。", "conditions_label": "可能的病症", "soap_copy_btn": "复制SOAP记录", "soap_empty": "运行分析以生成SOAP记录。", "tts_btn": "🔊 朗读", "chat_label": "后续问题", "chat_placeholder": "就您的病情提出后续问题...", "chat_send": "发送", "red_flags_label": "红旗症状", "watch_for_label": "需关注症状", "analyzing_label": "分析中...", }, "es": { "img_label": "Subir imagen médica", "symptoms_label": "Descripción de síntomas", "symptoms_placeholder": "Describa lo que siente — ej. erupción roja con picazón en el antebrazo desde hace 3 días...", "analyze_btn": "🔬 Analizar", "output_label": "Resultado del análisis", "empty_output": "Suba una imagen y/o describa sus síntomas, luego haga clic en Analizar.", "diag_label": "Sugerencia de diagnóstico", "severity_label": "Severidad", "actions_label": "Evaluación y orientación", "confidence_label": "Puntuación de confianza", "disclaimer": "Este es un asistente de IA, no un médico autorizado. Consulte siempre a un profesional de la salud para condiciones graves.", "placeholder": "Por favor, suba una imagen o describa sus síntomas.", "backend_live": "AMD Cloud · En línea", "backend_offline": "AMD Cloud · Sin conexión", "error_title": "Backend no disponible", "error_body": "El backend de AMD Cloud no está disponible. Por favor, inténtelo más tarde.", "examples_label": "Ejemplos rápidos", "metrics_latency": "Latencia", "metrics_throughput": "Rendimiento", "metrics_tokens": "tokens", "region_label": "Región corporal afectada", "region_none": "No especificado", "region_optional_label": "Región corporal afectada (Opcional)", "input_hint": "Suba una imagen, describa síntomas, o ambos — se requiere al menos uno. La región corporal es opcional.", "map_label": "Mapa anatómico", "map_select": "haga clic para seleccionar", "map_selected": "{n} región(es) seleccionada(s)", "img_mode_label": "Modo de carga", "img_mode_standard": "Estándar (1 imagen)", "img_mode_compare": "Comparar (2 imágenes — antes y después)", "img_label_day1": "Imagen médica (Día 1)", "img_label_dayx": "Imagen de comparación (Día X)", "tab_patient": "Vista del paciente", "tab_doctor": "Exportar para médico (SOAP)", "critical_warning": "⚠️ CRÍTICO: Síntomas graves detectados. Por favor, acuda a un centro médico en las próximas 24 horas.", "conditions_label": "Posibles condiciones", "soap_copy_btn": "Copiar nota SOAP", "soap_empty": "Ejecute un análisis para generar la nota SOAP.", "tts_btn": "🔊 Escuchar", "chat_label": "Preguntas de seguimiento", "chat_placeholder": "Haga una pregunta de seguimiento sobre su condición...", "chat_send": "Enviar", "red_flags_label": "Señales de alerta", "watch_for_label": "Vigilar", "analyzing_label": "Analizando...", }, "fr": { "img_label": "Télécharger une image médicale", "symptoms_label": "Description des symptômes", "symptoms_placeholder": "Décrivez ce que vous ressentez — ex. éruption rouge et prurigineuse sur l'avant-bras depuis 3 jours...", "analyze_btn": "🔬 Analyser", "output_label": "Résultat de l'analyse", "empty_output": "Téléchargez une image et/ou décrivez vos symptômes, puis cliquez sur Analyser.", "diag_label": "Suggestion de diagnostic", "severity_label": "Sévérité", "actions_label": "Évaluation et conseils", "confidence_label": "Score de confiance", "disclaimer": "Ceci est un assistant IA, pas un médecin agréé. Consultez toujours un professionnel de santé pour les situations graves.", "placeholder": "Veuillez télécharger une image ou décrire vos symptômes.", "backend_live": "AMD Cloud · En ligne", "backend_offline": "AMD Cloud · Hors ligne", "error_title": "Backend indisponible", "error_body": "Le backend AMD Cloud est inaccessible. Veuillez réessayer plus tard.", "examples_label": "Exemples rapides", "metrics_latency": "Latence", "metrics_throughput": "Débit", "metrics_tokens": "tokens", "region_label": "Région corporelle affectée", "region_none": "Non spécifié", "region_optional_label": "Région corporelle affectée (Facultatif)", "input_hint": "Fournissez une image, décrivez vos symptômes, ou les deux — au moins un est requis. La région corporelle est facultative.", "map_label": "Carte anatomique", "map_select": "cliquer pour sélectionner", "map_selected": "{n} région(s) sélectionnée(s)", "img_mode_label": "Mode de téléchargement", "img_mode_standard": "Standard (1 image)", "img_mode_compare": "Comparer (2 images — avant et après)", "img_label_day1": "Image médicale (Jour 1)", "img_label_dayx": "Image de comparaison (Jour X)", "tab_patient": "Vue patient", "tab_doctor": "Exporter pour le médecin (SOAP)", "critical_warning": "⚠️ CRITIQUE : Symptômes graves détectés. Veuillez vous rendre dans un établissement médical dans les 24 heures.", "conditions_label": "Conditions possibles", "soap_copy_btn": "Copier la note SOAP", "soap_empty": "Lancez une analyse pour générer la note SOAP.", "tts_btn": "🔊 Écouter", "chat_label": "Questions de suivi", "chat_placeholder": "Posez une question de suivi sur votre état...", "chat_send": "Envoyer", "red_flags_label": "Signaux d'alarme", "watch_for_label": "À surveiller", "analyzing_label": "Analyse en cours...", }, "ja": { "img_label": "医療画像をアップロード", "symptoms_label": "症状の説明", "symptoms_placeholder": "感じていることを説明してください — 例：3日前から前腕に赤くかゆい発疹、少し灼熱感...", "analyze_btn": "🔬 分析する", "output_label": "分析結果", "empty_output": "画像をアップロードし、症状を説明してから分析ボタンをクリックしてください。", "diag_label": "診断提案", "severity_label": "重症度", "actions_label": "評価とガイダンス", "confidence_label": "信頼スコア", "disclaimer": "これはAIアシスタントであり、有資格の医師ではありません。深刻な症状については必ず医療専門家に相談してください。", "placeholder": "画像をアップロードするか、症状を入力してください。", "backend_live": "AMD Cloud · オンライン", "backend_offline": "AMD Cloud · オフライン", "error_title": "バックエンド利用不可", "error_body": "AMD Cloudバックエンドに接続できません。後でもう一度お試しください。", "examples_label": "クイック例", "metrics_latency": "レイテンシ", "metrics_throughput": "スループット", "metrics_tokens": "トークン", "region_label": "患部の体の部位", "region_none": "指定なし", "region_optional_label": "患部の体の部位（任意）", "input_hint": "画像または症状（あるいは両方）を入力してください — どちらか一方が必要です。体の部位は任意です。", "map_label": "解剖マップ", "map_select": "クリックして選択", "map_selected": "{n} 部位選択中", "img_mode_label": "アップロードモード", "img_mode_standard": "標準（画像1枚）", "img_mode_compare": "比較（画像2枚 — 経過観察）", "img_label_day1": "医療画像（第1日）", "img_label_dayx": "比較画像（第X日）", "tab_patient": "患者向け", "tab_doctor": "医師向けエクスポート（SOAP）", "critical_warning": "⚠️ 重大：重篤な症状が検出されました。24時間以内に医療機関を受診してください。", "conditions_label": "考えられる疾患", "soap_copy_btn": "SOAPノートをコピー", "soap_empty": "分析を実行してSOAPノートを生成します。", "tts_btn": "🔊 読み上げ", "chat_label": "フォローアップの質問", "chat_placeholder": "症状についてフォローアップの質問をしてください...", "chat_send": "送信", "red_flags_label": "赤信号", "watch_for_label": "注意症状", "analyzing_label": "分析中...", }, } def _detect_lang_from_header(accept_language: str) -> str: """Map Accept-Language header to a display name from _LANG_CHOICES.""" if not accept_language: return "English" first = accept_language.split(",")[0].split(";")[0].strip().lower() code = first.split("-")[0] mapping = {"vi": "Tiếng Việt", "zh": "中文", "es": "Español", "fr": "Français", "ja": "日本語"} return mapping.get(code, "English") # --------------------------------------------------------------------------- # Backend status # --------------------------------------------------------------------------- def get_backend_status_html(lang: str = "en") -> str: t = _I18N.get(lang, _I18N["en"]) connected, _ = check_connection() if connected: dot, label, color = "#22c55e", t["backend_live"], "#86efac" else: dot, label, color = "#ef4444", t["backend_offline"], "#fca5a5" return ( f"
" f" {label}" f"
" ) # --------------------------------------------------------------------------- # Severity / confidence helpers # --------------------------------------------------------------------------- _SEVERITY_COLOR = { "Low": ("#22c55e", "#dcfce7"), "Medium": ("#eab308", "#fef9c3"), "High": ("#f97316", "#ffedd5"), "Urgent": ("#ef4444", "#fee2e2"), } # Translate English severity keys to display language _SEVERITY_TRANSLATE = { "en": {"Low": "Low", "Medium": "Medium", "High": "High", "Urgent": "Urgent"}, "vn": {"Low": "Thấp", "Medium": "Trung bình", "High": "Cao", "Urgent": "Khẩn cấp"}, "zh": {"Low": "低", "Medium": "中", "High": "高", "Urgent": "紧急"}, "es": {"Low": "Leve", "Medium": "Moderada", "High": "Alta", "Urgent": "Urgente"}, "fr": {"Low": "Faible", "Medium": "Modérée", "High": "Élevée", "Urgent": "Urgente"}, "ja": {"Low": "軽度", "Medium": "中等度", "High": "重度", "Urgent": "緊急"}, } # --------------------------------------------------------------------------- # Anatomical Region Selector # --------------------------------------------------------------------------- # Regions shown as a compact grid; value passed to prompt in English _BODY_REGIONS = [ "Head / Face", "Neck", "Chest", "Abdomen", "Upper Back", "Lower Back", "Left Arm", "Right Arm", "Left Hand", "Right Hand", "Left Leg", "Right Leg", "Left Foot", "Right Foot", "Groin / Genital", "Buttocks", ] # Translated display names — same order as _BODY_REGIONS _REGION_TRANSLATIONS = { "en": ["Head / Face", "Neck", "Chest", "Abdomen", "Upper Back", "Lower Back", "Left Arm", "Right Arm", "Left Hand", "Right Hand", "Left Leg", "Right Leg", "Left Foot", "Right Foot", "Groin / Genital", "Buttocks"], "vn": ["Đầu / Mặt", "Cổ", "Ngực", "Bụng", "Lưng trên", "Lưng dưới", "Tay trái", "Tay phải", "Bàn tay trái", "Bàn tay phải", "Chân trái", "Chân phải", "Bàn chân trái", "Bàn chân phải", "Bẹn / Sinh dục", "Mông"], "zh": ["头部 / 面部", "颈部", "胸部", "腹部", "上背部", "下背部", "左臂", "右臂", "左手", "右手", "左腿", "右腿", "左脚", "右脚", "腹股沟 / 生殖器", "臀部"], "es": ["Cabeza / Cara", "Cuello", "Pecho", "Abdomen", "Espalda alta", "Espalda baja", "Brazo izquierdo", "Brazo derecho", "Mano izquierda", "Mano derecha", "Pierna izquierda", "Pierna derecha", "Pie izquierdo", "Pie derecho", "Ingle / Genitales", "Nalgas"], "fr": ["Tête / Visage", "Cou", "Poitrine", "Abdomen", "Haut du dos", "Bas du dos", "Bras gauche", "Bras droit", "Main gauche", "Main droite", "Jambe gauche", "Jambe droite", "Pied gauche", "Pied droit", "Aine / Organes génitaux", "Fesses"], "ja": ["頭部 / 顔", "首", "胸部", "腹部", "上背部", "下背部", "左腕", "右腕", "左手", "右手", "左脚", "右脚", "左足", "右足", "鼠径部 / 性器", "臀部"], } def _localized_regions(lang: str) -> list: return _REGION_TRANSLATIONS.get(lang, _REGION_TRANSLATIONS["en"]) def _display_to_en(display: str) -> str: """Map any localized display name (any lang) back to the English region key.""" if display in _BODY_REGIONS: return display for translations in _REGION_TRANSLATIONS.values(): if display in translations: return _BODY_REGIONS[translations.index(display)] return "" # Each region maps to one or more SVG shape IDs that should highlight _REGION_SHAPE_MAP = { "Head / Face": ["svg-head"], "Neck": ["svg-neck"], "Chest": ["svg-chest"], "Abdomen": ["svg-abdomen"], "Upper Back": ["svg-upper-back"], "Lower Back": ["svg-lower-back"], "Left Arm": ["svg-left-arm"], "Right Arm": ["svg-right-arm"], "Left Hand": ["svg-left-hand"], "Right Hand": ["svg-right-hand"], "Left Leg": ["svg-left-leg"], "Right Leg": ["svg-right-leg"], "Left Foot": ["svg-left-foot"], "Right Foot": ["svg-right-foot"], "Groin / Genital": ["svg-groin"], "Buttocks": ["svg-buttocks"], } _DIM = "#374151" # default fill _HI = "#ED1C24" # highlighted fill def _body_map_svg(selected: list, lang: str = "en") -> str: t = _I18N.get(lang, _I18N["en"]) active = set() for r in (selected or []): for sid in _REGION_SHAPE_MAP.get(r, []): active.add(sid) def f(sid): return _HI if sid in active else _DIM def stroke(sid): return "#ED1C24" if sid in active else "#4b5563" def glow(sid): return "drop-shadow(0 0 3px #ED1C24)" if sid in active else "none" # Build svg_id → localized name for tooltips locs = _REGION_TRANSLATIONS.get(lang, _REGION_TRANSLATIONS["en"]) _svg_title: dict[str, str] = {} for en_name, svg_ids in _REGION_SHAPE_MAP.items(): loc_name = locs[_BODY_REGIONS.index(en_name)] if en_name in _BODY_REGIONS else en_name for sid in svg_ids: _svg_title[sid] = loc_name def title(sid): return f"<title>{_svg_title.get(sid, sid)}" counter_html = ( f'
' f'{t["map_selected"].format(n=len(active))}
' if active else f'
{t["map_select"]}
' ) return f"""
{t["map_label"]}
{title('svg-head')} {title('svg-neck')} {title('svg-chest')} {title('svg-abdomen')} {title('svg-upper-back')} {title('svg-lower-back')} {title('svg-left-arm')} {title('svg-right-arm')} {title('svg-left-hand')} {title('svg-right-hand')} {title('svg-groin')} {title('svg-buttocks')} {title('svg-left-leg')} {title('svg-right-leg')} {title('svg-left-foot')} {title('svg-right-foot')} {counter_html}
""" def _severity_badge(severity: str, lang: str = "en") -> str: color, bg = _SEVERITY_COLOR.get(severity, ("#6b7280", "#f3f4f6")) display = _SEVERITY_TRANSLATE.get(lang, _SEVERITY_TRANSLATE["en"]).get(severity, severity) return ( f"{display}" ) def _metrics_bar(metrics: dict, t: dict) -> str: latency_ms = metrics.get("latency_ms", 0) tok_per_sec = metrics.get("tokens_per_sec", 0) total_tokens = metrics.get("total_tokens", 0) latency_val = f"{latency_ms:,} ms" if latency_ms else "—" throughput_val = f"{tok_per_sec} {t['metrics_tokens']}/s" if tok_per_sec else "—" tokens_val = f"{total_tokens:,} {t['metrics_tokens']}" if total_tokens else "—" def chip(icon, label, val): return ( f"
" f"{icon}" f"{label}" f"{val}" f"
" ) return ( f"
" f"⚡ AMD MI300X" f"{chip('⏱', t['metrics_latency'], latency_val)}" f"{chip('🚀', t['metrics_throughput'], throughput_val)}" f"{chip('◊', t['metrics_tokens'], tokens_val)}" f"
" ) def _confidence_bar(score: int, label: str) -> str: if score == 0: return "" fill = "#ED1C24" if score >= 85 else "#f97316" if score >= 70 else "#eab308" return ( f"
" f"
" f" {label}: {score}%
" f"
" f"
" f"
" f"
" ) def _empty_output_html(lang: str) -> str: t = _I18N.get(lang, _I18N["en"]) return ( f"
" f"{t['empty_output']}" f"
" ) def _empty_soap_html(lang: str) -> str: return _build_soap_html("", lang) def _loading_output_html(lang: str) -> str: t = _I18N.get(lang, _I18N["en"]) dots = "".join( f"
" for i in range(3) ) return ( f"
" f"
⚕️
" f"
" f"{t.get('analyzing_label', 'Analyzing...')}
" f"
" f"AMD MI300X · ROCm · 3-Step Pipeline
" f"
{dots}
" f"
" ) def _condition_probability_bars(conditions: list, t: dict) -> str: """Render probability bars for each possible condition.""" if not conditions: return "—" bars = [] for c in conditions: if isinstance(c, dict): name = c.get("name", "Unknown") prob = int(c.get("probability", 50)) icd10 = c.get("icd10", "") else: name, prob, icd10 = str(c), 50, "" fill = "#ef4444" if prob >= 70 else "#f97316" if prob >= 45 else "#eab308" icd_badge = ( f"" f"{icd10}" ) if icd10 else "" bars.append( f"
" f"
" f" {name}{icd_badge}" f" {prob}%" f"
" f"
" f"
" f"
" f"
" ) return "".join(bars) def _red_flags_panel(red_flags: list, watch_symptoms: list, urgency_reason: str, t: dict = None) -> str: """Render red flags and watch symptoms warning panel. Returns empty string if nothing to show.""" if t is None: t = _I18N["en"] has_flags = bool(red_flags) has_watch = bool(watch_symptoms) has_urgency = bool(urgency_reason) if not has_flags and not has_watch and not has_urgency: return "" flags_html = "" if has_flags: items = "".join( f"
  • ⚠ {f}
    • " for f in red_flags ) flags_html = ( f"
    {t.get('red_flags_label', 'Red Flags')}
    " f"" ) watch_html = "" if has_watch: items = "".join( f"
  • 👁 {w}
    • " for w in watch_symptoms ) watch_html = ( f"
    {t.get('watch_for_label', 'Watch For')}
    " f"" ) urgency_html = "" if has_urgency: urgency_html = ( f"
    " f"ⓘ {urgency_reason}
    " ) border_color = "#ef4444" if has_flags else "#f59e0b" bg_color = "#1c0a0a" if has_flags else "#1c1000" return ( f"
    " f"{flags_html}{watch_html}{urgency_html}" f"
    " ) def _build_result_html(result: dict, lang: str) -> str: t = _I18N.get(lang, _I18N["en"]) triage = result.get("triage_level", "Low") patient_msg = result.get("patient_message", "") conditions = result.get("possible_conditions", []) red_flags = result.get("red_flags", []) watch_symptoms = result.get("watch_symptoms", []) urgency_reason = result.get("urgency_reason", "") metrics = result.get("_metrics", {}) backend_tag = ( "AMD Cloud" ) triage_colors = { "High": ("#ef4444", "#7f1d1d"), "Medium": ("#f97316", "#431407"), "Low": ("#22c55e", "#052e16"), } t_color, t_bg = triage_colors.get(triage, ("#22c55e", "#052e16")) critical_banner = "" if triage == "High": critical_banner = f"""
    {t['critical_warning']}
    """ msg_html = "".join( f"

    {line}

    " for line in patient_msg.split("\n") if line.strip() ) if patient_msg else "

    —

    " cond_bars = _condition_probability_bars(conditions, t) alert_panel = _red_flags_panel(red_flags, watch_symptoms, urgency_reason, t) return f"""
    MediVision {backend_tag}
    AMD MI300X · ROCm · Qwen2.5-VL-7B · 3-Step Pipeline
    {_metrics_bar(metrics, t)} {critical_banner}
    {t['severity_label']}
    {_SEVERITY_TRANSLATE.get(lang, _SEVERITY_TRANSLATE["en"]).get(triage, triage)}
    {t['conditions_label']}
    {cond_bars}
    {alert_panel}
    {t['actions_label']}
    {msg_html}
    ⚠️ {t['disclaimer']}
    """ def _build_soap_html(soap_text: str, lang: str = "en") -> str: t = _I18N.get(lang, _I18N["en"]) if not soap_text: return ( f"
    " f"{t['soap_empty']}
    " ) lines_html = "".join( f"
    {line}
    " for line in soap_text.split("\n") if line.strip() ) return f"""
    SOAP Clinical Note
    {lines_html}
    """ # --------------------------------------------------------------------------- # UI update helpers # --------------------------------------------------------------------------- def _error_html(t: dict, exc: Exception) -> str: return ( "
    " "
    ⚠️
    " f"
    {t['error_title']}
    " f"
    {t['error_body']}
    " f"
    {exc}
    " "
    " ) def _ui_updates(lang_choice: str, current_regions=None): """Return gr.update() for the 5 translatable input-area components (no output_html).""" lang = _LANG_MAP.get(lang_choice, "en") t = _I18N[lang] new_choices = _localized_regions(lang) # Translate current selection into new language translated = [] for r in (current_regions or []): en = _display_to_en(r) if en in _BODY_REGIONS: translated.append(new_choices[_BODY_REGIONS.index(en)]) hint_html = ( f"

    {t['input_hint']}

    " ) chat_label_html = ( f"
    {t['chat_label']}
    " ) return ( gr.update(label=t["img_label"], choices=[t["img_mode_standard"], t["img_mode_compare"]]), gr.update(label=t["img_label_day1"]), gr.update(label=t["img_label_dayx"]), gr.update(label=t["symptoms_label"], placeholder=t["symptoms_placeholder"]), gr.update(value=t["analyze_btn"]), gr.update(label=t["region_optional_label"], choices=new_choices, value=translated), gr.update(value=hint_html), gr.update(placeholder=t["chat_placeholder"]), gr.update(value=t["chat_send"]), gr.update(value=t["tts_btn"]), gr.update(value=chat_label_html), ) def _regions_to_prompt(selected) -> str: """Convert list of display names (any lang) to English prompt string.""" if not selected: return "" if isinstance(selected, str): selected = [selected] en_regions = [_display_to_en(r) for r in selected] return ", ".join(r for r in en_regions if r) def on_region_change(selected, lang_choice: str = "English"): """Re-render the body map SVG when selection changes (map display→EN first).""" if isinstance(selected, str): selected = [selected] en_keys = [_display_to_en(r) for r in (selected or [])] lang = _LANG_MAP.get(lang_choice, "en") return _body_map_svg(en_keys, lang) def on_svg_click(svg_id: str, current_regions: list, lang_choice: str) -> tuple: """Toggle a body region from an SVG click. Returns (new_dropdown_value, new_svg_html).""" # Map svg-id → English region name svg_to_region = {v[0]: k for k, v in _REGION_SHAPE_MAP.items() if v} clicked_en = svg_to_region.get(svg_id, "") if not clicked_en: return current_regions, _body_map_svg([_display_to_en(r) for r in (current_regions or [])]) lang = _LANG_MAP.get(lang_choice, "en") choices = _localized_regions(lang) en_list = _BODY_REGIONS clicked_display = choices[en_list.index(clicked_en)] if clicked_en in en_list else clicked_en current = list(current_regions or []) # Toggle: if already selected remove, else add en_current = [_display_to_en(r) for r in current] if clicked_en in en_current: current = [r for r in current if _display_to_en(r) != clicked_en] else: current.append(clicked_display) new_en = [_display_to_en(r) for r in current] return current, _body_map_svg(new_en, lang) def on_lang_change(lang_choice: str, selected_regions): lang = _LANG_MAP.get(lang_choice, "en") t = _I18N.get(lang, _I18N["en"]) mode_upd, day1_upd, dayx_upd, sym_upd, btn_upd, region_upd, hint_upd, chat_ph_upd, chat_send_upd, _tts_upd, chat_lbl_upd = _ui_updates( lang_choice, current_regions=selected_regions ) return (mode_upd, day1_upd, dayx_upd, sym_upd, btn_upd, region_upd, hint_upd, _empty_output_html(lang), _empty_soap_html(lang), get_backend_status_html(lang), chat_ph_upd, chat_send_upd, gr.update(visible=False, value=t["tts_btn"]), chat_lbl_upd) def on_load(request: gr.Request): lang_display = _detect_lang_from_header( request.headers.get("accept-language", "") ) mode_upd, day1_upd, dayx_upd, sym_upd, btn_upd, region_upd, hint_upd, chat_ph_upd, chat_send_upd, _tts_upd, chat_lbl_upd = _ui_updates( lang_display, current_regions=[] ) lang = _LANG_MAP.get(lang_display, "en") t = _I18N.get(lang, _I18N["en"]) return ( lang_display, mode_upd, day1_upd, dayx_upd, sym_upd, btn_upd, region_upd, hint_upd, _body_map_svg([], lang), _empty_output_html(lang), _empty_soap_html(lang), get_backend_status_html(lang), chat_ph_upd, chat_send_upd, gr.update(visible=False, value=t["tts_btn"]), chat_lbl_upd, ) # --------------------------------------------------------------------------- # Predict # --------------------------------------------------------------------------- def predict(image_1, image_2, symptoms: str, lang_choice: str, selected_regions): lang = _LANG_MAP.get(lang_choice, "en") t = _I18N[lang] _empty_ctx = {} if not image_1 and not image_2 and not (symptoms or "").strip(): return ( _empty_output_html(lang), _empty_soap_html(lang), get_backend_status_html(lang), _empty_ctx, [], gr.update(visible=False), gr.update(visible=False), "", ) region = _regions_to_prompt(selected_regions) try: result = get_pipeline().process( image_1, image_2, (symptoms or "").strip(), lang=lang, region=region ) patient_msg = result.get("patient_message", "") ctx = { "visual_description": result.get("visual_description", ""), "possible_conditions": result.get("possible_conditions", []), "triage_level": result.get("triage_level", "Low"), "urgency_reason": result.get("urgency_reason", ""), "red_flags": result.get("red_flags", []), "watch_symptoms": result.get("watch_symptoms", []), "patient_message": patient_msg, } return ( _build_result_html(result, lang), _build_soap_html(result.get("soap_note", ""), lang), get_backend_status_html(lang), ctx, [], gr.update(visible=True), gr.update(visible=bool(patient_msg)), patient_msg, ) except Exception as exc: return ( _error_html(t, exc), _empty_soap_html(lang), get_backend_status_html(lang), _empty_ctx, [], gr.update(visible=False), gr.update(visible=False), "", ) def on_tts_click(patient_msg: str, lang_choice: str): """Generate TTS audio for the patient message.""" lang = _LANG_MAP.get(lang_choice, "en") audio_path = _tts_generate(patient_msg, lang) return audio_path # --------------------------------------------------------------------------- # Chat # --------------------------------------------------------------------------- def on_chat_send(question: str, history: list, context: dict, lang_choice: str): from src.agents import chat_agent lang = _LANG_MAP.get(lang_choice, "en") if not question or not question.strip(): return history, "" # Convert messages-format [{role,content},...] to [[user,bot],...] tuples for chat_agent msgs = list(history or []) tuples = [] for i in range(0, len(msgs) - 1, 2): if msgs[i].get("role") == "user" and msgs[i+1].get("role") == "assistant": tuples.append([msgs[i]["content"], msgs[i+1]["content"]]) try: answer, _ = chat_agent(question.strip(), context, tuples, lang) except Exception as exc: answer = f"⚠️ {exc}" history = list(history or []) + [ {"role": "user", "content": question.strip()}, {"role": "assistant", "content": answer}, ] return history, "" # --------------------------------------------------------------------------- # CSS # --------------------------------------------------------------------------- CSS = """ @import url('https://fonts.googleapis.com/css2?family=Inter:wght@400;500;600;700;900&display=swap'); /* ── Base ── */ body, .gradio-container { background-color: #030712 !important; color: #f9fafb !important; font-family: 'Inter', 'Segoe UI', Arial, sans-serif; } /* ── Accessibility: focus rings ── */ *:focus-visible { outline: 2px solid #ED1C24 !important; outline-offset: 2px !important; border-radius: 4px; } input:focus, textarea:focus { border-color: #ED1C24 !important; box-shadow: 0 0 0 2px rgba(237,28,36,0.25) !important; } /* ── iOS: prevent auto-zoom on small inputs (must be ≥16px) ── */ @media (max-width: 768px) { input[type="text"], input[type="search"], select, textarea.scroll-hide { font-size: 16px !important; } } /* ── Analyze button ── */ button.primary, .gr-button-primary { background: linear-gradient(135deg, #ED1C24 0%, #b01318 100%) !important; color: #fff !important; border: none !important; font-weight: 700 !important; font-size: 1rem !important; letter-spacing: 0.05em !important; text-transform: uppercase !important; width: 100% !important; padding: 14px 0 !important; border-radius: 8px !important; box-shadow: 0 0 16px rgba(237,28,36,0.45), 0 4px 12px rgba(0,0,0,0.4) !important; transition: box-shadow 0.25s ease, opacity 0.2s !important; } button.primary:hover { opacity: 0.92 !important; box-shadow: 0 0 28px rgba(237,28,36,0.65), 0 6px 16px rgba(0,0,0,0.5) !important; } button.primary:active { opacity: 0.8 !important; transform: scale(0.99) !important; } /* Chat "Send" — override full-width / glow from above */ #chat-section button.primary, #chat-section .gr-button-primary { width: auto !important; min-width: 72px !important; padding: 10px 18px !important; font-size: 0.875rem !important; letter-spacing: 0.02em !important; text-transform: none !important; box-shadow: 0 2px 8px rgba(237,28,36,0.3) !important; } /* ── Panels / boxes ── */ .gr-box, .gr-panel { background: #111827 !important; border: 1px solid #1f2937 !important; border-radius: 10px !important; } /* ── Form labels ── */ label span, .gr-form > label { color: #cbd5e1 !important; font-size: 0.82rem !important; font-weight: 600 !important; text-transform: uppercase; letter-spacing: 0.04em; } footer { display: none !important; } /* ── TTS button ── */ #tts-btn { background: #1e3a5f !important; color: #93c5fd !important; border: 1px solid #2563eb !important; font-weight: 600 !important; font-size: 0.875rem !important; letter-spacing: 0.03em !important; text-transform: none !important; padding: 8px 18px !important; min-height: 44px !important; border-radius: 6px !important; box-shadow: none !important; width: auto !important; transition: background 0.2s, border-color 0.2s !important; } #tts-btn:hover { background: #1d4ed8 !important; border-color: #3b82f6 !important; opacity: 1 !important; box-shadow: none !important; } /* ── Tab bar — min 44px touch target, clear selected state ── */ [role="tab"] { min-height: 44px !important; padding: 8px 16px !important; color: #9ca3af !important; font-size: 0.875rem !important; font-weight: 500 !important; transition: color 0.2s !important; } [role="tab"][aria-selected="true"] { color: #f9fafb !important; } /* ── Animations ── */ @keyframes redflash { 0%, 100% { opacity: 1; box-shadow: 0 0 12px rgba(239,68,68,0.6); } 50% { opacity: 0.7; box-shadow: 0 0 24px rgba(239,68,68,0.9); } } @keyframes fadeSlideIn { from { opacity: 0; transform: translateY(8px); } to { opacity: 1; transform: translateY(0); } } @keyframes medispin { from { transform: rotate(0deg); } to { transform: rotate(360deg); } } @keyframes medipulse { 0%, 80%, 100% { transform: scale(0.6); opacity: 0.4; } 40% { transform: scale(1.1); opacity: 1; } } /* ── Respect prefers-reduced-motion ── */ @media (prefers-reduced-motion: reduce) { *, *::before, *::after { animation-duration: 0.01ms !important; animation-iteration-count: 1 !important; transition-duration: 0.01ms !important; } } /* ── Scrollbar ── */ ::-webkit-scrollbar { width: 6px; height: 6px; } ::-webkit-scrollbar-track { background: #111827; } ::-webkit-scrollbar-thumb { background: #374151; border-radius: 3px; } ::-webkit-scrollbar-thumb:hover { background: #4b5563; } /* ── Topbar ── */ #topbar { display: flex !important; align-items: center !important; justify-content: space-between !important; flex-wrap: wrap !important; gap: 8px !important; padding: 8px 0 !important; border-bottom: 1px solid #1f2937; margin-bottom: 16px; } #topbar > .gr-row, #topbar > div { width: 100% !important; } #lang-col { min-width: 160px !important; max-width: 200px !important; } #lang-col label span { text-transform: none !important; font-size: 0.82rem !important; } /* ── Chat section: smooth reveal ── */ #chat-section { animation: fadeSlideIn 0.3s ease; } /* ── Main 2-column → stack on mobile ── */ #main-row { flex-wrap: wrap !important; } @media (max-width: 768px) { #main-row > div { min-width: 100% !important; width: 100% !important; flex: 0 0 100% !important; } /* Chatbot height on mobile */ #chat-box > div { height: 200px !important; } } /* ── SVG body map: tap without double-tap zoom ── */ .bpart { touch-action: manipulation; } /* ── Image upload: mobile ── */ @media (pointer: coarse), (max-width: 768px) { .upload-container [data-testid="drop-zone"] .upload-text:first-child, .svelte-upload .file-preview-title, span.drag-text, .drag-drop-label { display: none !important; } .upload-container { min-height: 120px !important; } .upload-container button { font-size: 0.9rem !important; padding: 10px 20px !important; } } /* ── Quick Examples — horizontal scroll on mobile ── */ @media (max-width: 768px) { .gr-samples table, table.gr-samples-table { display: flex !important; overflow-x: auto !important; -webkit-overflow-scrolling: touch !important; gap: 8px !important; padding-bottom: 6px !important; } .gr-samples tr, .gr-samples-table tr { display: inline-flex !important; flex-direction: column !important; min-width: 220px !important; background: #111827 !important; border: 1px solid #1f2937 !important; border-radius: 8px !important; padding: 8px 10px !important; white-space: normal !important; } .gr-samples thead, .gr-samples-table thead { display: none !important; } } /* ── Safe area insets (notch / home-bar) ── */ @supports (padding-bottom: env(safe-area-inset-bottom)) { .gradio-container { padding-left: max(12px, env(safe-area-inset-left)) !important; padding-right: max(12px, env(safe-area-inset-right)) !important; padding-bottom: max(16px, env(safe-area-inset-bottom)) !important; } } """ # --------------------------------------------------------------------------- # Layout # --------------------------------------------------------------------------- HEADER_HTML = """
    MediVision
    Multilingual Dermatology & Wound Care AI Assistant
    AMD Instinct™ MI300X ROCm · Qwen2.5-VL-7B EN · VI · ZH · ES · FR · JA AMD Developer Hackathon 2026
    """ FOOTER_HTML = """
    Powered by AMD MI300X + ROCm  ·  Track 3: Vision & Multimodal AI  ·  MIT License
    """ BLOCKS_JS = """ () => { /* ── SVG body-map click → hidden bridge input ── */ document.addEventListener('click', function(e) { var el = e.target.closest('.bpart'); if (!el) return; var svgId = el.id; if (!svgId) return; var bridge = document.getElementById('svg-click-bridge'); if (!bridge) return; var input = bridge.querySelector('input, textarea'); if (!input) return; var setter = Object.getOwnPropertyDescriptor(window.HTMLInputElement.prototype, 'value').set || Object.getOwnPropertyDescriptor(window.HTMLTextAreaElement.prototype, 'value').set; setter.call(input, svgId); input.dispatchEvent(new InputEvent('input', { bubbles: true, data: svgId })); input.dispatchEvent(new Event('change', { bubbles: true })); }); } """ _TAB_LABELS_JS = """{ 'English': ['Patient View', 'Export for Doctor (SOAP)'], 'Tiếng Việt': ['Dành cho bệnh nhân', 'Xuất cho bác sĩ (SOAP)'], '中文': ['患者视图', '导出给医生（SOAP）'], 'Español': ['Vista del paciente', 'Exportar para médico (SOAP)'], 'Français': ['Vue patient', 'Exporter pour le médecin (SOAP)'], '日本語': ['患者向け', '医師向けエクスポート（SOAP）'], }""" UPDATE_TABS_JS = f"""(langDisplay) => {{ var TAB_LABELS = {_TAB_LABELS_JS}; var labels = TAB_LABELS[langDisplay] || TAB_LABELS['English']; var tabContainer = document.getElementById('output-tabs'); if (!tabContainer) return langDisplay; /* Gradio renders tab buttons inside a div[role=tablist] */ var tablist = tabContainer.querySelector('[role=tablist]') || tabContainer; var tabButtons = Array.from(tablist.querySelectorAll('button[role=tab]')); if (tabButtons.length < 2) tabButtons = Array.from(tablist.querySelectorAll('button')); tabButtons.slice(0, 2).forEach(function(btn, i) {{ if (labels[i] === undefined) return; /* Find the text span inside the button (Gradio wraps label in a span) */ var span = btn.querySelector('span') || btn; span.textContent = labels[i]; }}); return langDisplay; }}""" with gr.Blocks(css=CSS, js=BLOCKS_JS, theme=gr.themes.Base(), title="MediVision — Dermatology & Wound Care AI") as demo: gr.HTML(HEADER_HTML) # ── Topbar ─────────────────────────────────────────────────────────────── with gr.Row(elem_id="topbar"): with gr.Column(scale=5): status_bar = gr.HTML(value="
    ") with gr.Column(scale=0, elem_id="lang-col"): lang_radio = gr.Dropdown( choices=_LANG_CHOICES, value="English", label="🌐 Language", container=False, show_label=False, ) # Hidden bridge: SVG clicks write svg-element-id here → triggers on_svg_click svg_click_bridge = gr.Textbox( value="", visible=False, elem_id="svg-click-bridge", ) # ── Main content ────────────────────────────────────────────────────────── with gr.Row(equal_height=False, elem_id="main-row"): with gr.Column(scale=1, min_width=300): img_mode = gr.Radio( choices=[_I18N["en"]["img_mode_standard"], _I18N["en"]["img_mode_compare"]], value=_I18N["en"]["img_mode_standard"], label=_I18N["en"]["img_mode_label"], elem_id="img-mode-radio", ) with gr.Row(equal_height=True): input_img = gr.Image( type="filepath", label=_I18N["en"]["img_label_day1"], height=200, ) input_img_2 = gr.Image( type="filepath", label=_I18N["en"]["img_label_dayx"], height=200, visible=False, ) symptoms_txt = gr.Textbox( label="Symptoms Description", placeholder="Describe what you feel — e.g. itchy red patch for 3 days...", lines=4, ) input_hint_html = gr.HTML( value=f"

    " f"{_I18N['en']['input_hint']}

    ", elem_id="input-hint", ) with gr.Row(equal_height=True): with gr.Column(scale=0, min_width=80): body_map_html = gr.HTML(value=_body_map_svg([])) with gr.Column(scale=1): region_selector = gr.Dropdown( choices=_BODY_REGIONS, value=[], multiselect=True, label="Affected Body Region", container=True, ) submit_btn = gr.Button("🔬 Analyze", variant="primary", size="lg", elem_id="analyze-btn") gr.Examples( examples=[ [None, "I have a red, itchy rash on my forearm for 3 days. It burns slightly.", "English"], [None, "Small wound on my hand after a cut, slightly swollen with some redness.", "English"], [None, "Vết thương nhỏ ở bàn tay, hơi sưng và có dấu hiệu đỏ xung quanh.", "Tiếng Việt"], [None, "手臂上出现红色瘙痒皮疹，已持续3天，略有灼热感。", "中文"], [None, "Tengo una erupción roja y con picazón en el antebrazo desde hace 3 días.", "Español"], [None, "J'ai une éruption rouge et prurigineuse sur l'avant-bras depuis 3 jours.", "Français"], [None, "3日前から前腕に赤くてかゆい発疹があり、少し灼熱感があります。", "日本語"], ], inputs=[input_img, symptoms_txt, lang_radio], label="Quick Examples", ) with gr.Column(scale=1, min_width=340): with gr.Tabs(elem_id="output-tabs"): with gr.TabItem(_I18N["en"]["tab_patient"], elem_id="tab-patient"): output_html = gr.HTML(value=_empty_output_html("en")) with gr.TabItem(_I18N["en"]["tab_doctor"], elem_id="tab-doctor"): soap_html = gr.HTML(value=_empty_soap_html("en")) tts_btn = gr.Button( _I18N["en"]["tts_btn"], variant="secondary", size="sm", elem_id="tts-btn", visible=False, min_width=100, ) tts_audio = gr.Audio( value=None, label=None, autoplay=True, visible=False, show_label=False, show_download_button=False, ) # ── Follow-up Q&A chat ──────────────────────────────────────── with gr.Group(visible=False, elem_id="chat-section") as chat_section: chat_label_html = gr.HTML( "
    " f"{_I18N['en']['chat_label']}
    " ) chat_box = gr.Chatbot( value=[], elem_id="chat-box", height=280, show_label=False, type="messages", ) with gr.Row(equal_height=True): chat_input = gr.Textbox( placeholder=_I18N["en"]["chat_placeholder"], show_label=False, lines=1, scale=5, container=False, ) chat_send_btn = gr.Button( _I18N["en"]["chat_send"], variant="primary", scale=1, min_width=72, ) # ── States ──────────────────────────────────────────────────────────────── analysis_context_state = gr.State({}) chat_history_state = gr.State([]) patient_msg_state = gr.State("") # ── Events ─────────────────────────────────────────────────────────────── # Image mode toggle: show/hide second image upload img_mode.change( fn=lambda m, lc: gr.update(visible=_I18N[_LANG_MAP.get(lc, "en")]["img_mode_compare"] in m), inputs=[img_mode, lang_radio], outputs=[input_img_2], ) # SVG click → toggle region in dropdown + re-render SVG for _svg_evt in [svg_click_bridge.input, svg_click_bridge.change]: _svg_evt( fn=on_svg_click, inputs=[svg_click_bridge, region_selector, lang_radio], outputs=[region_selector, body_map_html], ) # Dropdown change → re-render SVG (keeps sync when user edits dropdown directly) region_selector.change( fn=on_region_change, inputs=[region_selector, lang_radio], outputs=[body_map_html], ) lang_radio.change( fn=on_lang_change, inputs=[lang_radio, region_selector], outputs=[img_mode, input_img, input_img_2, symptoms_txt, submit_btn, region_selector, input_hint_html, output_html, soap_html, status_bar, chat_input, chat_send_btn, tts_btn, chat_label_html], ).then( fn=None, inputs=[lang_radio], js=UPDATE_TABS_JS, ) submit_btn.click( fn=lambda lc: _loading_output_html(_LANG_MAP.get(lc, "en")), inputs=[lang_radio], outputs=[output_html], ).then( fn=predict, inputs=[input_img, input_img_2, symptoms_txt, lang_radio, region_selector], outputs=[output_html, soap_html, status_bar, analysis_context_state, chat_history_state, chat_section, tts_btn, patient_msg_state], api_name="analyze", ).then( fn=lambda h: h, inputs=[chat_history_state], outputs=[chat_box], ) tts_btn.click( fn=on_tts_click, inputs=[patient_msg_state, lang_radio], outputs=[tts_audio], ).then( fn=lambda a: gr.update(visible=bool(a)), inputs=[tts_audio], outputs=[tts_audio], ) # Chat send (button click or Enter) def _do_send(q, hist, ctx, lc): return on_chat_send(q, hist, ctx, lc) chat_send_btn.click( fn=_do_send, inputs=[chat_input, chat_history_state, analysis_context_state, lang_radio], outputs=[chat_history_state, chat_input], ).then(fn=lambda h: h, inputs=[chat_history_state], outputs=[chat_box]) chat_input.submit( fn=_do_send, inputs=[chat_input, chat_history_state, analysis_context_state, lang_radio], outputs=[chat_history_state, chat_input], ).then(fn=lambda h: h, inputs=[chat_history_state], outputs=[chat_box]) demo.load( fn=on_load, inputs=[], outputs=[ lang_radio, img_mode, input_img, input_img_2, symptoms_txt, submit_btn, region_selector, input_hint_html, body_map_html, output_html, soap_html, status_bar, chat_input, chat_send_btn, tts_btn, chat_label_html, ], ).then( fn=None, inputs=[lang_radio], js=UPDATE_TABS_JS, ) gr.HTML(FOOTER_HTML) if __name__ == "__main__": demo.launch(server_name="0.0.0.0", server_port=7860)