config.example.yaml

# 全球新闻抓取和分析系统配置文件示例
# 复制此文件为 config.yaml 并填入你的配置

# Gmail配置
email:
  smtp_server: "smtp.gmail.com"
  smtp_port: 587
  sender_email: "your-email@gmail.com"  # 请修改为你的Gmail地址
  sender_password: "your-app-password"  # 请使用Gmail应用专用密码，不是登录密码
  recipient_email: "your-email@gmail.com"  # 接收邮件的地址（保留向后兼容）
  recipient_emails:  # 支持多个接收邮箱（推荐使用）
    - "your-email@gmail.com"
    - "another@example.com"  # 可以添加更多邮箱
    - "team@company.com"

# 大模型配置 (支持OpenAI和Anthropic)
llm:
  provider: "openai"  # 可选: openai, anthropic, openai-compatible
  api_key: "your-api-key-here"  # 请填入你的API Key
  model: "gpt-4o"  # OpenAI: gpt-4o, gpt-4-turbo; Anthropic: claude-3-5-sonnet-20241022
  base_url: ""  # OpenAI兼容接口的base_url，如使用OpenAI官方可留空
  max_tokens: 4000
  temperature: 0.7
  timeout_seconds: 1800  # 本地模型处理长上下文时允许最多等待30分钟
  max_retries: 1  # 超时后只重试一次，避免任务长时间堆积

# News API配置
news_api:
  enabled: true
  api_key: "your-newsapi-key"  # 从 https://newsapi.org 获取
  countries: ["us", "gb", "de", "fr", "ru", "sg", "th", "id", "jp", "in"]
  categories: ["business", "technology", "general"]
  page_size: 10
  max_articles_per_country: 10

# 免费、无需 API Key 的全球新闻索引。仅保留白名单媒体，并作为 RSS 的补充。
gdelt:
  enabled: true
  timespan: "1d"
  max_records_per_query: 20
  queries:
    - "(economy OR market OR inflation OR \"interest rate\")"
    - "(semiconductor OR chip OR AI OR memory OR storage)"
    - "(bank OR credit OR liquidity OR consumer OR retail)"
    - "(US election OR White House OR Congress OR foreign policy OR Middle East)"
  trusted_domains:
    - "reuters.com"
    - "bbc.com"
    - "ft.com"
    - "apnews.com"
    - "cnbc.com"
    - "npr.org"
    - "theguardian.com"
    - "federalreserve.gov"

# Google News 搜索 RSS：免费无需密钥，仅接受媒体名称白名单。
google_news:
  enabled: true
  language: "en-US"
  country: "US"
  ceid: "US:en"
  max_records_per_query: 20
  queries:
    - "global economy markets"
    - "semiconductor AI memory storage"
    - "banking credit consumer retail"
    - "US election White House Congress Middle East foreign policy"
    - "site:imf.org/en/News/Articles IMF"
  trusted_sources:
    - "Reuters"
    - "BBC"
    - "Financial Times"
    - "Associated Press"
    - "CNBC"
    - "NPR"
    - "The Guardian"
    - "International Monetary Fund | IMF"

# IMF 的 RSS 页面会返回 HTML/403；保留官网归档作为独立的尽力抓取入口。
imf_archive:
  enabled: true
  url: "https://www.imf.org/en/whats-new-archive"
  language: "en"
  lookback_days: 30
  max_records: 10
  region: "international"
  priority: 5
  sectors: ["finance", "policy"]

# RSS Feed配置
rss_feeds:
  enabled: true
  sources:
    # 全球综合
    - name: "BBC World News"
      url: "http://feeds.bbci.co.uk/news/world/rss.xml"
      region: "global"
      priority: 5
    - name: "BBC Business"
      url: "http://feeds.bbci.co.uk/news/business/rss.xml"
      region: "global"
      priority: 5
    - name: "Al Jazeera"
      url: "https://www.aljazeera.com/xml/rss/all.xml"
      region: "global"
      priority: 5
    
    # 美国
    - name: "NPR News"
      url: "https://feeds.npr.org/1001/rss.xml"
      region: "americas"
      priority: 5
    - name: "CNN Top Stories"
      url: "http://rss.cnn.com/rss/cnn_topstories.rss"
      region: "americas"
      priority: 4
    - name: "CNN World"
      url: "http://rss.cnn.com/rss/cnn_world.rss"
      region: "americas"
      priority: 4
    
    # 欧洲
    - name: "Financial Times"
      url: "https://www.ft.com/?format=rss"
      region: "europe"
      priority: 5
    - name: "Deutsche Welle"
      url: "https://rss.dw.com/xml/rss-en-all"
      region: "europe"
      priority: 4
    - name: "France 24"
      url: "https://www.france24.com/en/rss"
      region: "europe"
      priority: 4
    - name: "The Guardian World"
      url: "https://www.theguardian.com/world/rss"
      region: "europe"
      priority: 5
    
    # 亚洲和东南亚
    - name: "Channel NewsAsia"
      url: "https://www.channelnewsasia.com/api/v1/rss-outbound-feed?_format=xml"
      region: "asia"
      priority: 5
    - name: "Nikkei Asia"
      url: "https://asia.nikkei.com/rss/feed/nar"
      region: "asia"
      priority: 5
    - name: "The Straits Times"
      url: "https://www.straitstimes.com/news/world/rss.xml"
      region: "asia"
      priority: 4
    - name: "Bangkok Post"
      url: "https://www.bangkokpost.com/rss/data/news.xml"
      region: "asia"
      priority: 3
    - name: "South China Morning Post"
      url: "https://www.scmp.com/rss/91/feed"
      region: "asia"
      priority: 5
    
    # 俄罗斯
    - name: "TASS"
      url: "https://tass.com/rss/v2.xml"
      region: "russia"
      priority: 3
    - name: "Moscow Times"
      url: "https://www.themoscowtimes.com/rss/news"
      region: "russia"
      priority: 4
    
    # 国际组织
    - name: "UN News"
      url: "https://news.un.org/feed/subscribe/en/news/all/rss.xml"
      region: "international"
      priority: 4
    - name: "Federal Reserve Press"
      url: "https://www.federalreserve.gov/feeds/press_all.xml"
      region: "americas"
      priority: 5
      sectors: ["finance", "policy"]

# 新闻抓取配置
fetcher:
  max_articles: 50  # 每次抓取的最大文章数
  fetch_interval_hours: 24  # 抓取间隔（小时）
  timeout_seconds: 30  # 请求超时时间
  retry_attempts: 3  # 失败重试次数
  retry_delay_seconds: 5  # 重试延迟
  user_agent: "Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36"

# 内容过滤配置
content_filter:
  min_content_length: 100  # 最小内容长度
  exclude_keywords:  # 排除包含这些关键词的文章
    - "horoscope"
    - "celebrity gossip"
    - "entertainment"
    - "sports"
  include_keywords:  # 优先包含这些关键词的文章
    - "economy"
    - "market"
    - "trade"
    - "gdp"
    - "inflation"
    - "policy"
    - "geopolitical"
    - "sanctions"
    - "investment"
    - "stock"
    - "AI"
    - "IT"
    - "semiconductor"
    - "chip"
    - "memory"
    - "storage"
    - "HBM"
    - "bank"
    - "credit"
    - "interest rate"
    - "consumer"
    - "retail"
    - "证监会"
    - "人民银行"
    - "工信部"
    - "election"
    - "White House"
    - "Congress"
    - "President"
    - "外交政策"
    - "美国大选"
    - "白宫"
    - "国会"
  sector_keywords:
    tech: ["AI", "semiconductor", "chip", "memory", "storage", "HBM", "算力", "芯片", "半导体", "存储"]
    finance: ["bank", "banking", "credit", "interest rate", "liquidity", "银行", "券商", "保险", "降准", "降息"]
    consumer: ["consumer", "retail", "consumer spending", "消费", "零售", "社零", "以旧换新", "食品饮料"]
    policy: ["policy", "regulation", "Federal Reserve", "PBOC", "CSRC", "政策", "监管", "证监会", "人民银行", "工信部"]
    us_politics: ["election", "White House", "Congress", "President", "campaign", "sanctions", "foreign policy", "美国大选", "白宫", "国会", "选举", "外交政策"]

# 美股和 A 股行业行情。A 股代码使用腾讯日线代码与东财 secid 的组合。
markets:
  enabled: true
  historical_days: 90
  timeout_seconds: 20
  retry_attempts: 2
  retry_delay_seconds: 2
  assets:
    - name: "标普500ETF"
      symbol: "SPY"
      provider: "yahoo"
      market: "US"
      sector: "broad"
    - name: "纳斯达克100ETF"
      symbol: "QQQ"
      provider: "yahoo"
      market: "US"
      sector: "tech"
    - name: "美国半导体ETF"
      symbol: "SMH"
      provider: "yahoo"
      market: "US"
      sector: "tech"
    - name: "美国金融ETF"
      symbol: "XLF"
      provider: "yahoo"
      market: "US"
      sector: "finance"
    - name: "美国必需消费ETF"
      symbol: "XLP"
      provider: "yahoo"
      market: "US"
      sector: "consumer"
    - name: "美国可选消费ETF"
      symbol: "XLY"
      provider: "yahoo"
      market: "US"
      sector: "consumer"
    - name: "VIX恐慌指数"
      symbol: "^VIX"
      provider: "yahoo"
      market: "US"
      sector: "risk"
    - name: "沪深300"
      symbol: "000300"
      tencent_symbol: "sh000300"
      secid: "1.000300"
      provider: "eastmoney"
      market: "CN"
      sector: "broad"
    - name: "科创50"
      symbol: "000688"
      tencent_symbol: "sh000688"
      secid: "1.000688"
      provider: "eastmoney"
      market: "CN"
      sector: "tech"
    - name: "半导体ETF"
      symbol: "512480.SS"
      provider: "yahoo"
      market: "CN"
      sector: "tech"
    - name: "银行ETF"
      symbol: "512800.SS"
      provider: "yahoo"
      market: "CN"
      sector: "finance"
    - name: "消费ETF"
      symbol: "159928.SZ"
      provider: "yahoo"
      market: "CN"
      sector: "consumer"

# 默认用权威新闻与市场相对强弱生成行业情绪摘要。
# 股吧抓取为可选功能，请确认数据源条款后才启用并配置代码。
sentiment:
  enabled: true
  guba_enabled: false
  timeout_seconds: 15
  guba_codes:
    tech: []
    finance: []
    consumer: []

# 分析配置
analysis:
  focus_areas:
    - "全球经济趋势"
    - "地缘政治风险"
    - "贸易关系变化"
    - "货币政策"
    - "全球股市影响因素"
    - "中国A股市场分析"
    - "大宗商品价格"
    - "科技发展趋势"
    - "AI行业发展"
    - "新能源汽车"
    - "半导体产业链"
  output_language: "zh-CN"  # 输出语言
  include_predictions: true  # 是否包含预测
  prediction_timeframe: "未来1-3个月"  # 预测时间范围
  include_a_share_analysis: true  # 是否包含A股专项分析
  short_term_timeframes: ["未来5个交易日", "未来1-4周"]
  us_market_focus:
    - "科技/AI/半导体与存储"
    - "金融与利率传导"
    - "必需及可选消费"
  a_share_focus:
    - "沪深300指数趋势"
    - "创业板/科创板机会"
    - "行业板块轮动"
    - "政策导向行业"
    - "出口导向企业影响"
    - "资金流向分析"
    - "芯片/AI/存储产业链"
    - "银行/券商/保险"
    - "必需消费与可选消费"

# 日志配置
logging:
  level: "INFO"  # DEBUG, INFO, WARNING, ERROR, CRITICAL
  file: "logs/news_analyzer.log"
  max_file_size_mb: 10
  backup_count: 5
  console_output: true

# 调度配置
scheduler:
  enabled: true
  run_time: "08:00"  # 每天运行时间 (24小时制)
  timezone: "Asia/Shanghai"  # 时区

# 数据存储配置
storage:
  save_raw_news: true
  news_cache_dir: "data/news_cache"
  reports_dir: "data/reports"
  mobile_reports_dir: "data/mobile_reports"
  market_cache_dir: "data/market_cache"
  historical_news_days: 14
  historical_news_max_per_day: 25
  max_cache_days: 360  # 缓存保留天数