【发布时间】:2017-11-15 20:39:38
【问题描述】:
我正在构建一个小型网络scraper,我偶然发现了以下问题:我的应用程序需要抓取网站的不同部分并将信息放入数据库。有时它会给出疯狂的结果,例如重复条目,或者它从函数 getPhoto() 返回 undefined。但是,如果我只调用该函数(并且不运行脚本的其余部分),它会返回正确的结果!
我有一个 for 循环,它遍历不同的 URL。它进入每个 URL 并抓取以下信息:1. 标题,2.description,3. 内部链接,4. 调用根据标题生成图像的函数(getPhoto(...)),5. 保存结果到数据库。一切都发生在服务器上(我使用的是 Cron 作业,没有客户端交互)
for (i = 0; i < AllLinks.length; i++) {
if (AllLinks[i] != undefined && AllLinks[i] != null && sepLink[2] == "www.fly4free.pl") {
var t2 = {
travelTitle: null,
travelTitle2: null,
travelTitle3: null,
travelDescription: null,
travelDescription2: null,
travelDescription3: null,
travelBuy: null,
travelBuy2: null,
travelImage: null
};
var TravelLink1 = AllLinks[i];
result = HTTP.get(AllLinks[i], {});
$ = cheerio.load(result.content);
t2.travelTitle = $('.article__title').text();
t2.travelDescription = $('.article__content').find('p').first().text();
if ($("img[src$='//www.fly4free.pl/wp-content/uploads/2016/09/lotJm.png']").parent().attr('href') != null) {
t2.travelBuy = $("img[src$='//www.fly4free.pl/wp-content/uploads/2016/09/lotJm.png']").parent().attr('href'); // Link to buy
}
if (t2.travelBuy) {
if (t2.travelBuy.split('https://').pop().split('http://').pop() != null) {
t2.travelBuy2 = t2.travelBuy.split('https://').pop().split('http://').pop(); // link ready for DB
} else {
t2.travelBuy2 = t2.travelBuy;
}
}
t2.travelTitle3 = convertCurrencyInText(t2.travelTitle, 'PLN');
t2.travelDescription3 = convertCurrencyInText(t2.travelDescription, 'PLN');
translate(t2.travelTitle3, {from: 'pl', to: 'en'}).then(res => {
t2.travelTitle2 = res.text; // title for DB
if (t2.travelTitle2) { t2.travelImage = getPhoto(t2.travelTitle2); }
translate(t2.travelDescription3, {from: 'pl', to: 'en'}).then(response => {
t2.travelDescription2 = response.text; // description for DB
if (t2.travelDescription2 != null && t2.travelTitle2 != null && t2.travelBuy2 != null && TravelLink1 != null && t2.travelImage != null) {
Links.insert({ title: t2.travelTitle2, description:t2.travelDescription2, image: t2.travelImage, buyLink:t2.travelBuy2, link: TravelLink1, datetime: new Date() });
}
}).catch(err => {
console.error(err);
});
}).catch(err => {
console.error(err);
});
}
}
“AllLinks”包含不同的 URL。我在抓取此 URL 时遇到问题:http://www.fly4free.pl/na-wakacje-do-toskanii-tanie-loty-do-pizy-z-gdanska-za-170-pln/
getPhoto() 函数
function getPhoto(title) {
var travelPlace = nlp(title).match('to *').out('text').replace('to','').trim();
if (travelPlace) {var travelPlace2 = travelPlace.split(' '); }
if (travelPlace2) {var travelPlace3 = travelPlace2[0] + "+" + travelPlace2[1]; }
if (travelPlace3) {
var URL = "https://pixabay.com/api/?key="+API_KEY+"&q="+travelPlace3+"&category=travel&orientation=horizontal";
var images = (HTTP.get(URL, {}));
if (images.data.totalHits > 0) {
var imageLink = images.data.hits[0].webformatURL;
return imageLink;
} else if (images.data.totalHits == 0) {
var URL = "https://pixabay.com/api/?key="+API_KEY+"&q="+travelPlace2[0]+"&category=travel&orientation=horizontal";
var images = (HTTP.get(URL, {}));
if (images.data.totalHits > 0) {
var imageLink = images.data.hits[0].webformatURL;
return imageLink;
}
}
} else if (nlp(title).places().data().length > 0) {
var result = nlp(title).places().data()[0].text.replace(/[^a-zA-Z ]/g, "").trim();
var URL = "https://pixabay.com/api/?key="+API_KEY+"&q="+result+"&category=travel&orientation=horizontal";
var images = (HTTP.get(URL, {}));
if (images.data.totalHits > 0) {
var imageLink = images.data.hits[0].webformatURL;
return imageLink;
}
} else {
var title2 = title.replace(/[^a-zA-Z ]/g, "").split(" ");
if (title2) {
for(i = 0; i < title2.length; i++) {
if (cities[title2[i]] == 1) {
var URL = "https://pixabay.com/api/?key="+API_KEY+"&q="+title2[i]+"&category=travel&orientation=horizontal";
var images = (HTTP.get(URL, {}));
if (images.data.totalHits > 0) {
var imageLink = images.data.hits[0].webformatURL;
return imageLink;
}
} else {
var URL = "https://pixabay.com/api/?key="+API_KEY+"&q=travel&category=travel&orientation=horizontal";
var images = (HTTP.get(URL, {}));
if (images.data.totalHits > 0) {
var imageLink = images.data.hits[0].webformatURL;
return imageLink;
}
}
}
}
}
}
我尝试通过控制台记录结果 - 有时我从 getPhoto() 获得正确的图像,但从 t2.travelBuy 获得未定义的链接,有时反之亦然。你能告诉我我做错了什么吗?我看到有些人在这类问题上使用 Promises 或 async/await 函数。你认为这对我有帮助吗?我应该如何更改我的代码才能抓取网站而不会出现“未定义”?
“翻译”来自“google-translate-api”包
【问题讨论】:
-
老实说并没有阅读你的全部内容,但简短的回答是:这听起来像是一个异步问题。要么返回一个promise,要么把你没有得到正确数据的函数放在设置数据的函数的回调中。
标签: javascript function meteor web-scraping serverside-javascript