LinkedIn Agent

class LinkedInAgent(BaseBrowserAgent):
    """Scrapes LinkedIn profiles via Cloud SDK skill, falls back to browser-use.
    
    Primary: Cloud SDK LinkedIn Company Posts skill (no cookies needed, $0.01/run)
    Fallback: browser-use Agent + Google snippet extraction
    """
    
    agent_name = "linkedin"
    
    def __init__(self, settings: Settings, *, inbox_pool=None):
        super().__init__(settings, inbox_pool=inbox_pool)
        self._cloud = CloudSkillRunner(settings)

# RESEARCH: Checked linkedin-api (2k stars, unofficial), linkedin-scraper (archived)
# DECISION: Browser Use Cloud SDK skills (LinkedIn Company Posts $0.01/run, deterministic)
# FALLBACK: browser-use Agent + Google snippet extraction
# ALT: linkedin-api (gets rate-limited fast, account bans)

def _extract_json(raw: str) -> dict:
    """Robustly extract JSON from browser-use output."""
    cleaned = raw.strip()
    if "```json" in cleaned:
        cleaned = cleaned.split("```json", 1)[1].split("```", 1)[0]
    elif "```" in cleaned:
        cleaned = cleaned.split("```", 1)[1].split("```", 1)[0]
    cleaned = cleaned.strip()
    
    for text in [cleaned, raw]:
        try:
            return json.loads(text)
        except (json.JSONDecodeError, ValueError):
            pass
        start = text.find("{")
        end = text.rfind("}") + 1
        if start >= 0 and end > start:
            try:
                return json.loads(text[start:end])
            except (json.JSONDecodeError, ValueError):
                pass
    return {}

def _parse_linkedin_output(
    raw_output: str, person_name: str
) -> dict:
    """Parse browser-use or Cloud SDK output into structured profile data."""
    data = _extract_json(raw_output)
    
    profile_url = data.get("profile_url", "")
    display_name = data.get("full_name", person_name)
    headline = data.get("headline", "")
    location = data.get("location")
    about = data.get("about", "")
    current_company = data.get("current_company", "")
    current_title = data.get("current_title", "")
    experience = data.get("experience", [])
    education = data.get("education", [])
    skills = data.get("skills", [])
    connections = data.get("connections_count")
    recent_posts = data.get("recent_posts", [])
    
    raw_data = {
        "headline": headline,
        "about": about,
        "current_company": current_company,
        "current_title": current_title,
        "experience": experience,
        "education": education,
        "skills": skills,
        "connections_count": connections,
        "recent_posts": recent_posts,
        "browser_use_output": raw_output,
    }
    
    followers_count = None
    if connections:
        try:
            cleaned = str(connections).replace(",", "").replace("+", "").strip()
            followers_count = int(cleaned)
        except (ValueError, TypeError):
            pass
    
    profile = SocialProfile(
        platform="linkedin",
        url=profile_url if profile_url else f"https://linkedin.com/search?q={person_name}",
        display_name=display_name,
        bio=headline if headline else about[:200] if about else None,
        followers=followers_count,
        location=location,
        raw_data=raw_data,
    )
    
    snippets: list[str] = []
    if headline:
        snippets.append(f"LinkedIn: {display_name} — {headline}")
    if about:
        snippets.append(f"About: {about[:300]}")
    if experience:
        exp_strs = [
            f"{e.get('title', '?')} at {e.get('company', '?')}" for e in experience[:3]
        ]
        snippets.append(f"Experience: {'; '.join(exp_strs)}")
    if education:
        edu_strs = [
            f"{e.get('degree', '')} {e.get('field', '')} @ {e.get('school', '?')}"
            for e in education[:2]
        ]
        snippets.append(f"Education: {'; '.join(edu_strs)}")
    if skills:
        snippets.append(f"Skills: {', '.join(skills[:10])}")
    if recent_posts:
        for post in recent_posts[:2]:
            text = post.get("text", "")
            if text:
                snippets.append(f"Post: {text[:150]}")
    if not snippets:
        snippets.append(raw_output[:500])
    
    return {"profile": profile, "snippets": snippets}