文章目录

  说到GooglePlay,自定义的数据结构,解析起来真的是让人感觉到窒息。而且基本是每间隔一段时间就会稍微的发现变动,解析规则基本持久不了太久可能就会失效,不过都是一些细微的变动,不值一提~

GooglePlay是没有对外提供任何API的,想要爬取相关的数据就需要通过Web端的方式,Git上面也有国外的大佬开源了google-play-scraper,Python跟JS版本的我记得都有,直接导包调用

但是稳定性不够好,也是基于Web端去爬取解析的,一旦结构发生变化,作者维护不够及时的话,自然也就无法使用

之后参考开源的项目,自己重新实现了数据抽取那一块逻辑,并将解析服务部署在了境外服务器上,然后通过远端调用的方式去解析

使用Python爬取GooglePlay并从复杂的自定义数据结构中实现解析-LMLPHP

在我之前的Cloudflare反爬虫防护绕过文章中,就是以一个第三方的APK下载网站为示例进行的讲解,GooglePlay的APK包如果是在官方网站下载的话会比较麻烦,直接可以用我提到的那个第三方网站取下载就行

主要爬取内容就是APP应用的相关描述信息,如下图所示:

使用Python爬取GooglePlay并从复杂的自定义数据结构中实现解析-LMLPHP

在点击关于此应用的时候,抓包可以看到重点数据集都嵌套在了HTML源代码中,使用JS函数定义了AF_initDataCallBackdata参数就是加载的数据

使用Python爬取GooglePlay并从复杂的自定义数据结构中实现解析-LMLPHP

尝试将数据拷贝出来,不过发现太长了,截图都放不下,图片都给整裂开了!反正就是巨长再加上多层嵌套,需要拆解慢慢去写解析规则!

接下来,先实现对data数据的提取,实现代码如下所示:

async def extract_json_block(html, block_id):
    prefix = re.compile(r"AF_init[dD]ata[cC]all[bB]ack\s*\({[^{}]*key:\s*'" + re.escape(block_id) + ".*?data:")
    suffix = re.compile(r"}\s*\)\s*;")

    try:
        block = prefix.split(html)[1]
        block = suffix.split(block)[0]
    except IndexError:
        raise PlayStoreException("Could not extract block %s" % block_id)

    block = block.strip()
    block = re.sub(r"^function\s*\([^)]*\)\s*{", "", block)
    block = re.sub("}$", "", block)
    block = re.sub(r", sideChannel: {$", "", block)

    return block

使用Python爬取GooglePlay并从复杂的自定义数据结构中实现解析-LMLPHP

如上就是提取出来的数据,拿到数据以后,如果你是小白新手或许会因此放弃,实在是让人头大,现在开始实现解析核心逻辑代码,如下所示:

app_detail_ds_block = 'ds:7'
app_details_mapping = {
    'title': [app_detail_ds_block, 1, 2, 0, 0],
    'developer_name': [app_detail_ds_block, 1, 2, 68, 0],
    'developer_link': [app_detail_ds_block, 1, 2, 68, 1, 4, 2],
    'price_inapp': [app_detail_ds_block, 1, 2, 19, 0],
    'category': [app_detail_ds_block, 1, 2, 79, 0, 0, 1, 4, 2],
    'video_link': [app_detail_ds_block, 1, 2, 100, 1, 2, 0, 2],
    'icon_link': [app_detail_ds_block, 1, 2, 95, 0, 3, 2],
    'num_downloads_approx': [app_detail_ds_block, 1, 2, 13, 1],
    'num_downloads': [app_detail_ds_block, 1, 2, 13, 2],
    'published_date': [app_detail_ds_block, 1, 2, 10, 0],
    'published_timestamp': [app_detail_ds_block, 1, 2, 10, 1, 0],
    'pegi': [app_detail_ds_block, 1, 2, 9, 0],
    'pegi_detail': [app_detail_ds_block, 1, 2, 9, 2, 1],
    'os': [app_detail_ds_block, 1, 2, 140, 1, 1, 0, 0, 1],
    'rating': [app_detail_ds_block, 1, 2, 51, 0, 1],
    'description': [app_detail_ds_block, 1, 2, 72, 0, 1],
    'price': [app_detail_ds_block, 1, 2, 57, 0, 0, 0, 0, 1, 0, 2],
    'num_of_reviews': [app_detail_ds_block, 1, 2, 51, 2, 1],
    'developer_email': [app_detail_ds_block, 1, 2, 69, 1, 0],
    'developer_address': [app_detail_ds_block, 1, 2, 69, 2, 0],
    'developer_website': [app_detail_ds_block, 1, 2, 69, 0, 5, 2],
    'developer_privacy_policy_link': [app_detail_ds_block, 1, 2, 99, 0, 5, 2],
    'data_safety_list': [app_detail_ds_block, 1, 2, 136, 1],
    'updated_on': [app_detail_ds_block, 1, 2, 145, 0, 0],
    'app_version': [app_detail_ds_block, 1, 2, 140, 0, 0, 0]
}

async def find_item_from_json_mapping(google_app_detail_request_result, app_detail_mapping):
	ds_json_block = app_detail_mapping[0]
	json_block_raw = await extract_json_block(google_app_detail_request_result, ds_json_block)
	json_block = json.loads(json_block_raw)
	return await get_nested_item(json_block, app_detail_mapping[1:])

app_details_mapping则是解析数据的核心,索引基本上变动较小!因为数据都是在list中多级嵌套,所以需要花费一点精力时间去分析,app_detail_ds_block前段时间我记得是ds:5,这个倒是会偶尔变动

class PlayStoreException(BaseException):

    def __init__(self, *args):
        if args:
            self.message = args[0]
        else:
            self.message = None

    def __str__(self):
        if self.message:
            return "PlayStoreException, {0}".format(self.message)
        else:
            return "PlayStoreException raised"

class GooglePlayStoreScraper(object):

    def __init__(self):
        self.PLAYSTORE_URL = "https://play.google.com"
        self.PROXIES = {'http': 'http://127.0.0.1:7890', 'https': 'http://127.0.0.1:7890'}

    async def _app_connection(self, url, sleeptime=2, retry=0):
        for _ in range(retry + 1):
            try:
                async with aiohttp.ClientSession() as session:
                    async with session.get(url, proxy=self.PROXIES) as response:
                        return await response.text()
            except aiohttp.ClientError:
                if sleeptime > 0:
                    await asyncio.sleep(sleeptime)
        raise PlayStoreException(f"Could not connect to : {url}")

    async def get_nested_item(self, item_holder, list_of_indexes):
        index = list_of_indexes[0]
        if len(list_of_indexes) > 1:
            return await get_nested_item(item_holder[index], list_of_indexes[1:])
        else:
            return item_holder[index]

    async def get_app_details(self, app_id, country="nl", lang="nl"):
        url = f"{self.PLAYSTORE_URL}/store/apps/details?id={quote_plus(app_id)}&hl={lang}&gl={country}"
        request_result = await self._app_connection(url, retry=1)

        app = {'id': app_id, 'link': url}
        
        for k, v in app_details_mapping.items():
            try:
                app[k] = await self.find_item_from_json_mapping(request_result, v)
            except PlayStoreException:
                raise PlayStoreException(f"Could not parse Play Store response for {app_id}")
            except Exception as e:
                self._log_error(country, f'App Detail error for {app_id} on detail {k}: {str(e)}')
                app.setdefault('errors', []).append(k)

        app['developer_link'] = self.PLAYSTORE_URL + app.get('developer_link', '')
        app['category'] = app.get('category', '').replace('/store/apps/category/', '')
        
        if 'data_safety_list' in app:
            app['data_safety_list'] = ', '.join(item[1] for item in app['data_safety_list'] if len(item) > 1)
        
        soup = BeautifulSoup(request_result, 'html.parser')
        list_of_categories = ', '.join(', '.join(category.text for category in element.find_all('span')) for element in soup.find_all('div', {'class': 'Uc6QCc'}))
        app['list_of_categories'] = list_of_categories if list_of_categories else app.setdefault('errors', []) + ['list_of_categories']

        if 'errors' in app:
            plural = 's' if len(app['errors']) > 1 else ''
            app['errors'] = f"Detail{plural} not found for key{plural}: {', '.join(app['errors'])}"

        return app

在完成上面爬虫程序核心入口的实现以后,基本上用采集到数据解析都已经完成,只需要调用get_app_details函数,传人需要爬取的目标APP的包名即可爬取并解析数据,如下所示:

使用Python爬取GooglePlay并从复杂的自定义数据结构中实现解析-LMLPHP

  好了,到这里又到了跟大家说再见的时候了。创作不易,帮忙点个赞再走吧。你的支持是我创作的动力,希望能带给大家更多优质的文章

12-21 22:20