youlu-openclaw-workspace/scripts/news_digest/send_digest.py

#!/usr/bin/env python3
"""
Daily News Digest - Fetch news and send email summary
Runs news_digest, formats output, sends via himalaya
"""

import json
import subprocess
import sys
from datetime import datetime
from pathlib import Path

# Config
SCRIPT_DIR = Path(__file__).resolve().parent
DB_PATH = SCRIPT_DIR / "news_digest.db"
CONFIG_PATH = SCRIPT_DIR / "config.json"

# Email config
EMAIL_TO = "lu@luyx.org"
EMAIL_FROM = "youlu@luyanxin.com"

def run_news_digest():
    """Run news_digest and get articles with summaries"""
    # Fetch new articles using run.sh (handles uv dependencies)
    fetch_cmd = [
        str(SCRIPT_DIR / "run.sh"),
    ]

    # Run fetch (this saves articles to DB and generates summaries)
    try:
        result = subprocess.run(fetch_cmd, capture_output=True, text=True, cwd=SCRIPT_DIR, timeout=300)
    except subprocess.TimeoutExpired:
        print("Fetch timed out after 300s", file=sys.stderr)
        sys.exit(1)
    if result.returncode != 0:
        print(f"Fetch failed: {result.stderr}", file=sys.stderr)
        return None
    
    # Now query with summary field
    query_cmd = [
        str(SCRIPT_DIR / "run.sh"),
        "--no-fetch",
        "-f", "id,title,url,summary,feed_name,category"
    ]
    
    try:
        result = subprocess.run(query_cmd, capture_output=True, text=True, cwd=SCRIPT_DIR, timeout=30)
    except subprocess.TimeoutExpired:
        print("Query timed out after 30s", file=sys.stderr)
        sys.exit(1)
    if result.returncode != 0:
        print(f"Query failed: {result.stderr}", file=sys.stderr)
        return None

    # Parse JSON output
    try:
        articles = json.loads(result.stdout)
        return articles
    except json.JSONDecodeError as e:
        print(f"Failed to parse JSON: {e}", file=sys.stderr)
        print(f"Raw output: {result.stdout[:500]}", file=sys.stderr)
        return []

def format_email(articles):
    """Format articles into email body"""
    if not articles:
        return None
    
    today = datetime.now().strftime('%Y-%m-%d')
    
    lines = [
        f"📰 每日新闻摘要 ({today})",
        "=" * 50,
        ""
    ]
    
    # Group by category
    by_category = {}
    for a in articles:
        cat = a.get('category', 'Other')
        if cat not in by_category:
            by_category[cat] = []
        by_category[cat].append(a)
    
    # Sort categories
    category_order = ['Tech', 'China', 'World', 'Other']
    ordered_cats = category_order + [c for c in by_category if c not in category_order]

    for cat in ordered_cats:
        if cat not in by_category:
            continue
        
        cat_name = {
            'Tech': '💻 科技',
            'China': '🇨🇳 国内',
            'World': '🌍 国际',
            'Other': '📄 其他'
        }.get(cat, cat)
        
        lines.append(f"\n{cat_name}")
        lines.append("-" * 30)
        
        for article in by_category[cat]:
            title = article.get('title', '无标题')
            url = article.get('url', '')
            summary = article.get('summary', '')
            source = article.get('feed_name', '未知来源')
            
            lines.append(f"\n• {title}")
            if summary:
                # Clean up summary (remove excessive newlines, preserve content)
                summary_clean = summary.replace('\n', ' ').strip()
                lines.append(f"  {summary_clean}")
            lines.append(f"  来源: {source} | {url}")
    
    lines.append("\n" + "=" * 50)
    lines.append("由 RSS News Digest 自动生成")
    
    return '\n'.join(lines)

def send_email(subject, body):
    """Send email via himalaya"""
    # Construct raw email message with headers
    message = f"""From: {EMAIL_FROM}
To: {EMAIL_TO}
Subject: {subject}
Content-Type: text/plain; charset=utf-8

{body}"""
    
    cmd = [
        "himalaya",
        "message", "send",
        "-a", "Youlu"  # Account name from config.toml
    ]

    try:
        result = subprocess.run(cmd, capture_output=True, text=True, input=message, timeout=30)
    except subprocess.TimeoutExpired:
        return False, "himalaya timed out after 30s"
    return result.returncode == 0, result.stderr

def main():
    # Fetch and get articles
    articles = run_news_digest()
    
    if articles is None:
        # Fetch failed, don't send email
        sys.exit(1)
    
    if not articles:
        # No new articles, send a "no news" email or skip
        # Let's skip sending email if no articles
        print("No articles to send")
        return
    
    # Format email
    body = format_email(articles)
    if not body:
        print("Failed to format email")
        sys.exit(1)
    
    # Send email
    today = datetime.now().strftime('%Y-%m-%d')
    subject = f"📰 每日新闻摘要 - {today}"
    
    success, error = send_email(subject, body)
    if success:
        print(f"Email sent successfully: {len(articles)} articles")
    else:
        print(f"Failed to send email: {error}", file=sys.stderr)
        sys.exit(1)

if __name__ == "__main__":
    main()