#!/usr/bin/env python3
"""
知乎问题页抓取脚本 v2 (bypass zse-ck, 支持滚动加载全部回答)
用法:
  python3 zhihu_question_fetch2.py <question_url或question_id> [output.txt] [max_answers]
示例:
  python3 zhihu_question_fetch2.py 356351510 /tmp/q.txt 10
原理: playwright 真实浏览器, 先访问首页种 cookie(d_c0), 再访问问题页,
等待 zse-ck JS 验证, 滚动加载更多回答, 提取: 问题标题 / 问题描述 / 回答列表。
"""
import sys, re, json, argparse, time
from playwright.sync_api import sync_playwright

CHROME = "/root/.cache/ms-playwright/chromium-1217/chrome-linux64/chrome"
UA = ("Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 "
      "(KHTML, like Gecko) Chrome/126.0.0.0 Safari/537.36")

def extract_question_id(url_or_id: str) -> str:
    m = re.search(r"/question/(\d+)", url_or_id)
    if m:
        return m.group(1)
    if url_or_id.isdigit():
        return url_or_id
    raise ValueError("无法解析 question id: " + url_or_id)

def fetch(url_or_id: str, timeout_ms: int = 60000, max_answers: int = 12) -> dict:
    qid = extract_question_id(url_or_id)
    url = f"https://www.zhihu.com/question/{qid}"
    result = {"question_id": qid, "url": url, "title": "", "detail": "",
              "answers": [], "error": None}
    with sync_playwright() as p:
        browser = p.chromium.launch(
            executable_path=CHROME,
            headless=True,
            args=["--no-sandbox", "--disable-dev-shm-usage",
                  "--disable-blink-features=AutomationControlled"],
        )
        ctx = browser.new_context(user_agent=UA, viewport={"width": 1280, "height": 900}, locale="zh-CN")
        page = ctx.new_page()
        try:
            page.goto("https://www.zhihu.com/", timeout=30000, wait_until="domcontentloaded")
            page.wait_for_timeout(5000)
        except Exception as e:
            result["error"] = f"home: {e}"
        try:
            page.goto(url, timeout=timeout_ms, wait_until="domcontentloaded")
            page.wait_for_timeout(12000)

            # 问题标题
            for sel in ["h1.QuestionHeader-title", "h1", ".QuestionHeader-title"]:
                el = page.query_selector(sel)
                if el:
                    t = el.inner_text().strip()
                    if t:
                        result["title"] = t
                        break
            # 问题描述
            for sel in [".QuestionRichText", ".QuestionHeader-detail .RichText",
                        ".QuestionHeader-detail"]:
                el = page.query_selector(sel)
                if el:
                    d = el.inner_text().strip()
                    if d:
                        result["detail"] = d
                        break

            # 滚动加载更多回答: 最多滚 25 次, 每次 3000ms
            last_count = 0
            for i in range(25):
                page.mouse.wheel(0, 3000)
                page.wait_for_timeout(1800)
                cards = page.query_selector_all(".List-item")
                cur = len(cards)
                if cur != last_count:
                    last_count = cur
                if cur >= max_answers + 3:
                    break
            # 再滚几次确保底部
            for i in range(3):
                page.mouse.wheel(0, 4000)
                page.wait_for_timeout(1500)

            # 展开"阅读全文"
            for sel in ["button:has-text('阅读全文')", "button:has-text('展开')",
                        ".RichContent-collapsedText"]:
                try:
                    btns = page.query_selector_all(sel)
                    for b in btns[:max_answers + 5]:
                        try:
                            b.click(timeout=2000)
                            page.wait_for_timeout(1000)
                        except Exception:
                            pass
                except Exception:
                    pass

            # 按回答卡片提取
            seen = set()
            cards = page.query_selector_all(".List-item .AnswerCard, .List-item")
            for card in cards:
                if len(result["answers"]) >= max_answers:
                    break
                author = ""
                a = card.query_selector(".AuthorInfo-name, .UserLink-link, .AuthorInfo")
                if a:
                    author = a.inner_text().strip()
                rich = card.query_selector(".RichText, .RichContent-inner")
                if not rich:
                    continue
                body = rich.inner_text().strip()
                if len(body) < 30:
                    continue
                key = body[:80]
                if key in seen:
                    continue
                seen.add(key)
                result["answers"].append({"author": author, "text": body})
            if not result["answers"]:
                for rich in page.query_selector_all(".RichText")[:max_answers]:
                    body = rich.inner_text().strip()
                    if len(body) >= 30:
                        result["answers"].append({"author": "", "text": body})
        except Exception as e:
            result["error"] = f"page: {e}"
        browser.close()
    return result

if __name__ == "__main__":
    ap = argparse.ArgumentParser()
    ap.add_argument("url_or_id", help="知乎问题链接或 question_id")
    ap.add_argument("output", nargs="?", default=None, help="输出文件路径(默认打印到stdout)")
    ap.add_argument("max_answers", nargs="?", type=int, default=12, help="最多抓取回答数")
    args = ap.parse_args()
    r = fetch(args.url_or_id, max_answers=args.max_answers)
    if r["error"]:
        print("WARN:", r["error"], file=sys.stderr)
    lines = []
    if r["title"]:
        lines.append("【问题】" + r["title"])
    if r["detail"]:
        lines.append("【问题描述】\n" + r["detail"])
    lines.append(f"【共抓取 {len(r['answers'])} 个回答】")
    for i, ans in enumerate(r["answers"], 1):
        lines.append(f"\n===== 回答 {i}" + (f" | {ans['author']}" if ans["author"] else "") + " =====")
        lines.append(ans["text"])
    text = "\n".join(lines)
    if text:
        if args.output:
            with open(args.output, "w", encoding="utf-8") as f:
                f.write(text)
            print(f"已保存 {len(text)} 字符, {len(r['answers'])} 个回答 -> {args.output}", file=sys.stderr)
        else:
            print(text)
    else:
        print("FAILED: 未提取到内容, title=", r["title"], file=sys.stderr)
        sys.exit(1)