一、為什么要自己爬店鋪詳情？

選品：直播團(tuán)隊(duì)需要「店鋪評(píng)分、粉絲數(shù)、上新頻率」快速篩選靠譜供應(yīng)商

競(jìng)品：同一類目，對(duì)手店鋪突然漲粉 10w，第一時(shí)間預(yù)警

數(shù)據(jù)訓(xùn)練：店鋪簡(jiǎn)介 + 評(píng)分 → 做多模態(tài)情感分類

投資：提前發(fā)現(xiàn)「高評(píng)分+低粉絲」?jié)摿Φ辏劰?yīng)鏈合作

官方 taobao.shop.get 需要企業(yè)資質(zhì)，個(gè)人 99% 被卡；網(wǎng)頁(yè)端「店鋪主頁(yè)」公開可見，走網(wǎng)頁(yè)派依舊是最低成本方案。下面用純 PHP 把「店鋪主頁(yè) → 基礎(chǔ)信息 → 商品列表 → 落庫(kù) → 飛書播報(bào)」一次擼完。

二、技術(shù)選型（全部開源）

三、0 環(huán)境搭建（Linux / Win / mac 通用）

bash

# 1. 安裝 PHP 8.2+ 擴(kuò)展
sudo dnf install php php-cli php-curl php-dom php-mbstring php-pdo php-mysqlnd

# 2. 安裝 Composer
curl -sS https://getcomposer.org/installer | php
sudo mv composer.phar /usr/local/bin/composer

# 3. 創(chuàng)建項(xiàng)目
mkdir taobao-shop-php && cd taobao-shop-php
composer init --no-interaction --require="php:>=8.2"
composer install

四、核心流程：6 步閉環(huán)（全部代碼可跑）

① 找入口：店鋪主頁(yè) + 簽名算法（2025-10 有效）

店鋪主頁(yè)：

https://shop{m}.taobao.com/shop/shop_index.htm?shop_id={shop_id}

店鋪內(nèi)所有商品接口（JSONP）：

https://shop.m.taobao.com/shop/shopItemSearch.jsonp?shopId={shop_id}¤tPage={page}&pageSize=20&callback=jsonp123

返回：

JavaScript

jsonp123({"total":1523,"items":[{...}]})

簽名邏輯（與詳情頁(yè)同款）：

php

function sign(string $raw): string {
    return strtoupper(md5($raw));
}

調(diào)用前拼字符串：t + "&12574478&" + data + "&"，其中 t 為毫秒時(shí)間戳。

② 網(wǎng)絡(luò)層：GuzzleHttp 7 + 連接池

php

http = new Client([
            'timeout' => 10,
            'headers' => [
                'User-Agent' => 'Mozilla/5.0 (Windows NT 10.0; Win64; x64)',
                'Referer' => 'https://shop.taobao.com/'
            ]
        ]);
    }

    public function fetchIndex(int $shopId): array {
        $url = "https://shop.taobao.com/shop/shop_index.htm?shop_id={$shopId}";
        $html = $this->http->get($url)->getBody()->getContents();
        return $this->parseIndex($html, $shopId);
    }

    public function fetchItems(int $shopId, int $page = 1): array {
        $this->rateLimit(); // 限速
        $callback = 'jsonp' . microtime(true);
        $data = json_encode(['shopId' => $shopId, 'currentPage' => $page], JSON_UNESCAPED_SLASHES);
        $t = (string) (microtime(true) * 1000);
        $sign = sign($t . "&12574478&" . $data . "&");

        $url = "https://shop.m.taobao.com/shop/shopItemSearch.jsonp?" . http_build_query([
            'shopId' => $shopId,
            'currentPage' => $page,
            'pageSize' => 20,
            'callback' => $callback,
            't' => $t,
            'sign' => $sign
        ]);

        $jsonp = $this->http->get($url)->getBody()->getContents();
        $json = preg_replace('/^jsonpd+(|)$/m', '', $jsonp);
        return $this->parseItems(json_decode($json, true), $shopId);
    }

    private function parseIndex(string $html, int $shopId): array {
        $doc = new DOMDocument();
        @$doc->loadHTML($html);
        $xpath = new DOMXPath($doc);
        return [
            'shop_id' => $shopId,
            'shop_name' => trim($xpath->query("http://h1[@class='shop-name']")->item(0)?->nodeValue ?? ''),
            'shop_score' => trim($xpath->query("http://span[@class='shop-score']")->item(0)?->nodeValue ?? ''),
            'shop_fans' => trim($xpath->query("http://span[@class='shop-fans']")->item(0)?->nodeValue ?? ''),
        ];
    }

    private function parseItems(array $root, int $shopId): array {
        $items = [];
        foreach ($root['items'] ?? [] as $i) {
            $items[] = [
                'shop_id' => $shopId,
                'item_id' => $i['itemId'],
                'title' => $i['title'],
                'price' => $i['price'],
                'pic_url' => $i['picUrl'],
                'sold' => $i['sold'] ?? 0,
                'comment_count' => $i['commentCount'] ?? 0,
                'created_at' => date('Y-m-d H:i:s', $i['created'] / 1000)
            ];
        }
        return $items;
    }

    private function rateLimit(): void {
        usleep(1000000 / $this->qps); // 微秒
    }
}

③ 并發(fā)池：Guzzle Pool + 進(jìn)度條

php

public function fetchAllItems(int $shopId, int $maxPage = 200): array {
    $first = $this->fetchItems($shopId, 1);
    $totalPage = min(ceil(($first['total'] ?? 0) / 20), $maxPage);

    $requests = function () use ($shopId, $totalPage) {
        for ($p = 2; $p <= $totalPage; $p++) {
            yield new Request('GET', $this-?>buildItemUrl($shopId, $p));
        }
    };

    $items = [];
    $pool = new Pool($this->http, $requests(), [
        'concurrency' => 15, // 令牌桶
        'fulfilled' => function ($response, $index) use (&$items, $shopId) {
            $jsonp = $response->getBody()->getContents();
            $json = preg_replace('/^jsonpd+(|)$/m', '', $jsonp);
            $items[] = $this->parseItems(json_decode($json, true), $shopId);
        },
        'rejected' => function ($reason, $index) {
            Log::error("Page $index failed: $reason");
        },
    ]);
    $pool->promise()->wait();
    return array_merge(...$items);
}

④ 落庫(kù)：Laravel Eloquent 批量 + Redis 去重

sql

CREATE TABLE tb_shop_detail (
  id BIGINT AUTO_INCREMENT PRIMARY KEY,
  shop_id BIGINT NOT NULL,
  shop_name VARCHAR(100) NOT NULL,
  shop_score VARCHAR(20) NOT NULL,
  shop_fans VARCHAR(20) NOT NULL,
  item_id BIGINT NOT NULL,
  title VARCHAR(200) NOT NULL,
  price VARCHAR(30) NOT NULL,
  pic_url VARCHAR(500) NOT NULL,
  sold INT DEFAULT 0,
  comment_count INT DEFAULT 0,
  created_at DATETIME NOT NULL,
  UNIQUE KEY uk_item (item_id),
  INDEX idx_shop (shop_id)
) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4;

模型：

php

	

	批量插入：

	php

	
use IlluminateSupportFacadesDB;
use AppModelsShopDetail;

function bulkSave(array $rows): int
{
    $new = 0;
    foreach (array_chunk($rows, 1000) as $chunk) {
        $exists = Redis::command('sadd', ['item_id_set', ...array_column($chunk, 'item_id')]);
        $filtered = array_filter($chunk, fn($i) => $exists[$i['item_id']] ?? false);
        if ($filtered) {
            ShopDetail::insert($filtered);
            $new += count($filtered);
        }
    }
    return $new;
}

	

	⑥ Docker 定時(shí)：每天 8 點(diǎn)飛書播報(bào)

	Dockerfile

	dockerfile

	
FROM php:8.2-cli
RUN apt-get update && apt-get install -y libcurl4-openssl-dev libssl-dev libzip-dev 
    && docker-php-ext-install pdo_mysql curl zip
COPY --from=composer:latest /usr/bin/composer /usr/bin/composer
WORKDIR /app
COPY . .
RUN composer install --no-dev
CMD ["php","crawl.php"]

	

	crontab

	0 8 * * * docker run --rm -v /mnt/nas/shop:/app/storage taobao-shop-php

	飛書推送（精簡(jiǎn)版）

	php

	
function report(int $shopId, int $new): void {
    $body = json_encode([
        'msg_type' => 'text',
        'content'  => ['text' => "店鋪 $shopId 新增 $new 條商品，已落庫(kù)~"]
    ]);
    file_get_contents('https://open.feishu.cn/open-apis/bot/v2/hook/xxx', false, stream_context_create([
        'http' => ['method' => 'POST', 'header' => 'Content-Type: application/json', 'content' => $body]
    ]));
}

	

	五、踩坑 & 反爬錦囊

	JSONP 殼：正則為 ^jsonpd+(|)$，剝完再 json_decode

	Referer：必須 https://shop.taobao.com/，否則 403

	限速：?jiǎn)?IP 15 QPS 穩(wěn)過，> 200/10min 必出滑塊

	代理池：青果云 1G ≈ 0.8 元，能跑 8 萬(wàn)頁(yè)

	重復(fù)：Redis item_id_set 秒級(jí)去重，內(nèi)存省 90 %

	六、結(jié)語(yǔ)

	從店鋪主頁(yè)、JSONP 簽名、Guzzle 并發(fā)、Eloquent 落庫(kù)，到 Docker 定時(shí) + 飛書群播報(bào)，一條完整的 PHP 閉環(huán)就打通了。

	全部代碼可直接扔進(jìn) PhpStorm / VSCode 跑通，改一行 shopId 就能薅任意店鋪。

	祝各位運(yùn)營(yíng)、產(chǎn)品、算法大佬爬得開心，爆單更開心！


	審核編輯 黃宇