rag/data_sources.json

{
  "version": "1.0",
  "last_updated": "2025-10-18",
  "description": "Healthcare documents for RAG system - Official sources only",
  "sources": [
    {
      "id": "who_healthy_diet_2020",
      "category": "nutrition",
      "name": "WHO Healthy Diet Fact Sheet",
      "url": "https://www.who.int/news-room/fact-sheets/detail/healthy-diet",
      "type": "html",
      "language": "en",
      "required": true,
      "cache_days": 90,
      "notes": "WHO official guidelines on healthy diet"
    },
    {
      "id": "cdc_physical_activity",
      "category": "fitness",
      "name": "CDC Physical Activity Guidelines",
      "url": "https://www.cdc.gov/physicalactivity/basics/adults/index.htm",
      "type": "html",
      "language": "en",
      "required": true,
      "cache_days": 90,
      "notes": "CDC guidelines for adult physical activity"
    },
    {
      "id": "nimh_mental_health",
      "category": "mental_health",
      "name": "NIMH Mental Health Information",
      "url": "https://www.nimh.nih.gov/health/topics",
      "type": "html",
      "language": "en",
      "required": true,
      "cache_days": 90,
      "notes": "National Institute of Mental Health resources"
    },
    {
      "id": "who_nutrition_landscape_2020",
      "category": "nutrition",
      "name": "WHO Global Nutrition Report",
      "url": "https://www.who.int/data/nutrition/nlis/info/malnutrition-in-women",
      "type": "html",
      "language": "en",
      "required": false,
      "cache_days": 180,
      "notes": "WHO nutrition data and statistics"
    },
    {
      "id": "cdc_nutrition_facts",
      "category": "nutrition",
      "name": "CDC Nutrition Facts",
      "url": "https://www.cdc.gov/nutrition/index.html",
      "type": "html",
      "language": "en",
      "required": true,
      "cache_days": 90,
      "notes": "CDC nutrition guidelines and resources"
    },
    {
      "id": "who_physical_activity",
      "category": "fitness",
      "name": "WHO Physical Activity Recommendations",
      "url": "https://www.who.int/news-room/fact-sheets/detail/physical-activity",
      "type": "html",
      "language": "en",
      "required": true,
      "cache_days": 90,
      "notes": "WHO guidelines on physical activity"
    },
    {
      "id": "cdc_mental_health",
      "category": "mental_health",
      "name": "CDC Mental Health Resources",
      "url": "https://www.cdc.gov/mentalhealth/index.htm",
      "type": "html",
      "language": "en",
      "required": true,
      "cache_days": 90,
      "notes": "CDC mental health information"
    },
    {
      "id": "who_mental_health",
      "category": "mental_health",
      "name": "WHO Mental Health Fact Sheet",
      "url": "https://www.who.int/news-room/fact-sheets/detail/mental-health-strengthening-our-response",
      "type": "html",
      "language": "en",
      "required": true,
      "cache_days": 90,
      "notes": "WHO mental health guidelines"
    },
    {
      "id": "cdc_sleep_health",
      "category": "general",
      "name": "CDC Sleep and Health",
      "url": "https://www.cdc.gov/sleep/index.html",
      "type": "html",
      "language": "en",
      "required": true,
      "cache_days": 90,
      "notes": "CDC sleep health information"
    },
    {
      "id": "who_obesity_overweight",
      "category": "general",
      "name": "WHO Obesity and Overweight",
      "url": "https://www.who.int/news-room/fact-sheets/detail/obesity-and-overweight",
      "type": "html",
      "language": "en",
      "required": true,
      "cache_days": 90,
      "notes": "WHO guidelines on obesity and overweight"
    }
  ],
  "manual_downloads": [
    {
      "category": "nutrition_vn",
      "name": "Khuyến nghị dinh dưỡng cho người Việt Nam",
      "source": "Bộ Y Tế Việt Nam",
      "url": "https://moh.gov.vn",
      "instructions": "Truy cập moh.gov.vn → Văn bản → Tìm kiếm 'dinh dưỡng' → Download các file PDF liên quan",
      "required": true
    },
    {
      "category": "nutrition_vn",
      "name": "Bảng thành phần dinh dưỡng thực phẩm Việt Nam",
      "source": "Viện Dinh Dưỡng Quốc Gia",
      "url": "http://viendinhduong.vn",
      "instructions": "Truy cập viendinhduong.vn → Tài liệu → Download bảng thành phần dinh dưỡng",
      "required": true
    },
    {
      "category": "traditional_medicine",
      "name": "Y học cổ truyền Việt Nam",
      "source": "Viện Y Học Cổ Truyền",
      "url": "http://vietim.vn",
      "instructions": "Truy cập vietim.vn → Tài liệu → Download tài liệu về dược liệu và bài thuốc",
      "required": false
    }
  ],
  "notes": [
    "All sources are official government/health organization websites",
    "HTML sources will be converted to markdown for ingestion",
    "Manual downloads are required for Vietnamese-specific content",
    "Update cache_days to control how often documents are re-downloaded"
  ]
}