rakuten-api/scripts/probe_search_add.py

"""Rakuten 探针：搜低价商品 → 实测 add-to-cart → 跨 cart/checkout

策略（顺序）：
1. 用 Playwright 打开 search.rakuten.co.jp 限定最低价筛选
2. 从结果里逐个取商品 URL
3. 用 Playwright 实际打开商品页，等 "買い物かごに入れる" 按钮
4. 点按钮（不是手工 POST，确保 form 字段都是当前版本）
5. 跟到 cart 后让 SPA 渲染够久
6. 落证据，尝试清理（删除购物车里这件）

避免 httpx（已被 TLS 指纹拦）。全程 Playwright。
"""
from __future__ import annotations

import asyncio
import json
import re
import sys
from pathlib import Path

sys.path.insert(0, str(Path(__file__).resolve().parent.parent))

from app.shared.config import get_settings  # noqa: E402
from app.trading.core import auth_site  # noqa: E402

PROBE_DIR = Path(__file__).resolve().parent.parent / ".probe" / "checkout"
PROBE_DIR.mkdir(parents=True, exist_ok=True)


def save(name: str, content: str) -> Path:
    path = PROBE_DIR / name
    path.write_text(content, encoding="utf-8")
    print(f"  saved -> {path} ({len(content)} bytes)")
    return path


async def collect_item_urls(page, limit: int = 10) -> list[str]:
    """从搜索结果抽商品 URL（PC 与 SP 都试）"""
    html = await page.content()
    # 候选：/shop-code/item-code/、item.rakuten.co.jp、div[data-itemid]
    urls: list[str] = []
    for m in re.findall(r'https://item\.rakuten\.co\.jp/[\w-]+/[\w-]+/?', html):
        if m not in urls:
            urls.append(m)
        if len(urls) >= limit:
            break
    return urls


async def try_add_item(page, item_url: str) -> tuple[bool, str]:
    """尝试加购一个商品。返回 (是否成功, 落地 URL)。"""
    print(f"  open item page: {item_url}")
    try:
        await page.goto(item_url, wait_until="domcontentloaded", timeout=30_000)
    except Exception as exc:
        return False, f"goto failed: {exc}"

    # 等 __INITIAL_STATE__ 出现
    try:
        await page.wait_for_function(
            "() => window.__INITIAL_STATE__ && window.__INITIAL_STATE__.purchase",
            timeout=10_000,
        )
    except Exception:
        return False, "no __INITIAL_STATE__ (PC 模板/反爬)"

    # 找加购按钮
    candidates = [
        'input[type="submit"][value*="買い物かご"]',
        'input[type="submit"][value*="購入"]',
        'button:has-text("買い物かご")',
        'button:has-text("カートに入れる")',
        '[data-testid*="add-cart"]',
        'button.add-cart',
    ]
    btn = None
    for sel in candidates:
        loc = page.locator(sel).first
        try:
            await loc.wait_for(state="visible", timeout=2000)
            btn = loc
            print(f"  found button via {sel!r}")
            break
        except Exception:
            continue
    if btn is None:
        return False, "no add-to-cart button visible"

    # 监听导航
    try:
        async with page.expect_navigation(timeout=15_000, wait_until="domcontentloaded"):
            await btn.click()
    except Exception as exc:
        return False, f"click/navigation failed: {exc}"

    final_url = page.url
    body = await page.content()
    # 成功标志：URL 含 cart/add-item 或 confirm；或正文含 レジに進む
    success = (
        "add-item" in final_url
        or "cart" in final_url
        or "レジに進む" in body
        or "買い物かごに追加" in body
    )
    # 失败标志：URL 含 /error
    if "/error" in final_url:
        return False, f"error page: {final_url}"
    return success, final_url


async def main() -> int:
    settings = get_settings()
    state_path = settings.auth_state_path / "rakuten_state.json"
    if not state_path.exists():
        print(f"找不到 {state_path}")
        return 1

    from playwright.async_api import async_playwright

    async with async_playwright() as pw:
        browser = await pw.chromium.launch(
            headless=True,
            channel=settings.browser_channel or None,
            args=["--no-first-run", "--disable-blink-features=AutomationControlled"],
        )
        context = await browser.new_context(
            storage_state=state_path,
            user_agent=auth_site.RAKUTEN_USER_AGENT,
            locale="ja-JP",
            timezone_id="Asia/Tokyo",
            viewport={"width": 390, "height": 844},
            is_mobile=True,
            has_touch=True,
        )
        page = await context.new_page()

        # 搜索：100~300 円筛选
        print("=== 步骤 1：搜低价商品 ===")
        search_url = "https://search.rakuten.co.jp/search/mall/?f=2&fs=1000&max=300&min=100&s=1&p=1"
        # 备用关键词：邮费便宜 / 免邮的便宜商品
        await page.goto(search_url, wait_until="domcontentloaded", timeout=30_000)
        await page.wait_for_timeout(2000)
        save("search-result.html", await page.content())
        item_urls = await collect_item_urls(page, limit=10)
        print(f"  候选商品 {len(item_urls)} 个")
        for u in item_urls[:5]:
            print(f"    {u}")

        # 试加购，最多试 5 个
        chosen: str | None = None
        for i, url in enumerate(item_urls[:5]):
            print(f"\n=== 步骤 2.{i+1}：试加购 {url} ===")
            ok, info = await try_add_item(page, url)
            print(f"  result: ok={ok} info={info}")
            if ok:
                chosen = url
                save("03-add-success-landing.html", await page.content())
                break
            # 失败：保存页面用于复盘
            save(f"03-add-fail-{i+1}.html", await page.content())

        if not chosen:
            print("\n✗ 5 个商品全部加购失败。")
            await browser.close()
            return 1

        print(f"\n=== 步骤 3：访问 cart，让 SPA 充分渲染 ===")
        # 先访问 cart
        await page.goto(auth_site.RAKUTEN_CART_URL, wait_until="domcontentloaded", timeout=30_000)
        # 等 SPA 把 cart 数据渲染出来：尝试等 "レジに進む" 出现，最多 15 秒
        try:
            await page.wait_for_selector(
                'a:has-text("レジに進む"), button:has-text("レジに進む"), [data-testid*="checkout"]',
                timeout=15_000,
            )
            print("  'レジに進む' 出现了！")
        except Exception:
            print("  'レジに進む' 15s 内没出现，继续抓现状")

        cart_html = await page.content()
        save("04-cart-after-add.html", cart_html)
        print(f"  cart len={len(cart_html)} url={page.url}")

        # 找 レジに進む href
        checkout_links = re.findall(r'href=["\']([^"\']*(?:checkout|step1|order/step)[^"\']*)["\']', cart_html, re.IGNORECASE)
        print(f"  checkout links: {sorted(set(checkout_links))[:3]}")

        # 看 cart count API
        resp = await context.request.get(
            "https://cart-api.step.rakuten.co.jp/rms/mall/cart/count/all/jsonp/?sid=1010",
            headers={"Referer": "https://sp.cart.step.rakuten.co.jp/cart"},
        )
        print(f"  count API: status={resp.status} body={(await resp.text())[:200]!r}")

        print("\n=== 步骤 4：尝试清理购物车 ===")
        # 找削除按钮
        del_selectors = [
            'a:has-text("削除")',
            'button:has-text("削除")',
            '[data-testid*="delete"]',
            '[data-testid*="remove"]',
        ]
        for sel in del_selectors:
            loc = page.locator(sel).first
            try:
                await loc.wait_for(state="visible", timeout=3000)
                print(f"  找到删除按钮 {sel!r}，点击")
                await loc.click()
                await page.wait_for_timeout(2000)
                break
            except Exception:
                continue
        else:
            print("  没找到删除按钮，请手动清理购物车")

        save("05-cart-after-cleanup.html", await page.content())

        await browser.close()

    print(f"\n探针输出目录：{PROBE_DIR}")
    print(f"测试用商品：{chosen}")
    return 0


if __name__ == "__main__":
    raise SystemExit(asyncio.run(main()))