#!/usr/bin/env python3
"""Get full news page snapshots for article context."""
import subprocess
import re
import time

AGENT_BROWSER = "/home/hermes/.hermes/node/lib/node_modules/agent-browser/bin/agent-browser-linux-x64"

def run(cmd, timeout=15):
    result = subprocess.run(cmd, shell=True, capture_output=True, text=True, timeout=timeout)
    return result.stdout

for news_id in [1101, 1103]:
    run(f'{AGENT_BROWSER} close --all 2>/dev/null', timeout=3)
    print(f"\n=== News ID {news_id} ===")
    run(f'{AGENT_BROWSER} open "https://en.boardgamearena.com/news?id={news_id}" --args "--no-sandbox"', timeout=20)
    time.sleep(2)
    snap = run(f'{AGENT_BROWSER} snapshot', timeout=15)
    
    with open(f'/home/hermes/workspace/bga_news_{news_id}.txt', 'w') as f:
        f.write(snap)
    
    # Extract paragraphs after the heading
    # Find the game description text - typically StaticText paragraphs
    texts = re.findall(r'StaticText\s+\"([^\"]{30,})\"', snap)
    
    print(f"  Title: {re.findall(r'heading\s+\"([^\"]+)\"\s+\[level=3', snap)}")
    print(f"  Date: {re.findall(r'StaticText\s+\"(\w+\s+\d+(?:st|nd|rd|th)\s+\d{4})\"', snap)[:5]}")
    print(f"  Long texts ({len(texts)} found):")
    for i, t in enumerate(texts[:15]):
        if len(t) > 30:
            print(f"    [{i}] ({len(t)} chars) {t[:200]}...")
    
    run(f'{AGENT_BROWSER} close --all 2>/dev/null', timeout=3)
    time.sleep(1)

print("\nDone.")