xhs_crawler_claude/app/user_resolver.py

"""user_resolver.py

将“数字小红书号（例如 11432840954）”解析成 profile_id（/user/profile/{profile_id} 里的那串）。
- 使用 Playwright 打开搜索结果页（更接近真实浏览器，风险更低）
- 解析 HTML：根据 '小红书号：<keyword>' 定位，再提取附近的 profile href
- 强制缓存：data/profile_map.json，避免重复搜索造成风控压力
"""

from __future__ import annotations

import json
import os
import random
import re
from html import unescape
from typing import Dict, Optional, Tuple

from playwright.async_api import Browser

from app import config

PROFILE_MAP_FILE = os.path.join(config.DATA_DIR, "profile_map.json")


def _load_map() -> Dict[str, str]:
    if not os.path.exists(PROFILE_MAP_FILE):
        return {}
    try:
        with open(PROFILE_MAP_FILE, "r", encoding="utf-8") as f:
            return json.load(f) or {}
    except Exception:
        return {}


def _save_map(m: Dict[str, str]) -> None:
    tmp = PROFILE_MAP_FILE + ".tmp"
    with open(tmp, "w", encoding="utf-8") as f:
        json.dump(m, f, ensure_ascii=False, indent=2)
    os.replace(tmp, PROFILE_MAP_FILE)


def extract_profile_id_from_search_html(html: str, keyword: str) -> Optional[str]:
    """从 search_result HTML 中解析出对应 keyword 的 profile_id。"""
    if not html:
        return None

    html = unescape(html)
    marker = f"小红书号：{keyword}"
    idx = html.find(marker)
    if idx == -1:
        return None

    start = max(0, idx - 5000)
    end = min(len(html), idx + 5000)
    chunk = html[start:end]

    m = re.search(r'href="\/user\/profile\/([0-9a-fA-F]+)', chunk)
    if not m:
        return None
    return m.group(1)


def parse_line_to_keyword_or_profile(line: str) -> Tuple[Optional[str], Optional[str], Optional[str]]:
    """解析一行 txt，返回 (keyword数字, profile_id, label备注) 三者之一。

    允许格式：
    - 11432840954
    - 11432840954  张三
    - https://www.xiaohongshu.com/user/profile/<profile_id> ...
    - /user/profile/<profile_id> ...
    - profile_id(十六进制串)
    - 支持行尾 # 备注
    """
    raw = line.strip()
    if not raw:
        return None, None, None

    # 整行注释
    if raw.startswith("#") or raw.startswith("//"):
        return None, None, None

    # 拆备注：支持 # 以及连续空格
    label = None
    if "#" in raw:
        raw, label = raw.split("#", 1)
        raw = raw.strip()
        label = label.strip() or None

    parts = raw.split()
    core = parts[0].strip() if parts else ""
    if len(parts) > 1 and not label:
        label = " ".join(parts[1:]).strip() or None

    # URL/相对路径中直接提取 profile_id
    m = re.search(r"/user/profile/([0-9a-fA-F]+)", core)
    if m:
        return None, m.group(1), label

    # 可能直接就是 profile_id（十六进制串，长度>=8）
    if re.fullmatch(r"[0-9a-fA-F]{8,}", core):
        return None, core, label

    # 纯数字：小红书号
    if re.fullmatch(r"\d{5,20}", core):
        return core, None, label

    return None, None, label


async def resolve_keyword_to_profile_id(browser: Browser, keyword: str, *, slow_mode: bool = True) -> Tuple[Optional[str], str]:
    """将数字 keyword 解析成 profile_id，并写入缓存。返回 (profile_id, message)。"""
    m = _load_map()
    if keyword in m:
        return m[keyword], "cache"

    # 使用新的 context（不强制登录）
    context = await browser.new_context()
    page = await context.new_page()

    try:
        url = f"{config.XIAOHONGSHU_BASE_URL}/search_result?keyword={keyword}&source=unknown"
        await page.goto(url, wait_until="domcontentloaded", timeout=20000)

        # 轻微等待/滚动（更像真人浏览）
        if slow_mode:
            await page.wait_for_timeout(int(random.uniform(800, 1600)))
            # 小幅滚动一下
            try:
                await page.mouse.wheel(0, random.randint(200, 600))
            except Exception:
                pass
            await page.wait_for_timeout(int(random.uniform(400, 1000)))

        html = await page.content()
        pid = extract_profile_id_from_search_html(html, keyword)
        if not pid:
            return None, "not_found"

        m[keyword] = pid
        _save_map(m)
        return pid, "ok"
    except Exception as e:
        return None, f"error:{e}"
    finally:
        try:
            await page.close()
        except Exception:
            pass
        try:
            await context.close()
        except Exception:
            pass