Quantcast
Channel: CNode:Node.js专业中文社区
Viewing all articles
Browse latest Browse all 14821

用TypeScript开发爬虫程序

$
0
0

全局安装typescript: npm install -g typescript目前版本2.0.3,这个版本不再需要使用typings命令了。但是vscode捆绑的版本是1.8的,需要一些配置工作,看本文的处理办法。

测试tsc命令: tsc创建要写的程序项目文件夹: mkdir test-typescript-spider进入该文件夹: cd test-typescript-spider初始化项目: npm init安装superagent和cheerio模块: npm i --save superagent cheerio安装对应的类型声明模块: npm i -s @types/superagent --savenpm i -s @types/cheerio --save安装项目内的typescript(必须走这一步): npm i --save typescript用vscode打开项目文件夹。在该文件夹下创建tsconfig.json文件,并复制以下配置代码进去:

{
    "compilerOptions": {
        "target": "ES6",
        "module": "commonjs",
        "noEmitOnError": true,
        "noImplicitAny": true,
        "experimentalDecorators": true,
        "sourceMap": false,
     // "sourceRoot": "./",
        "outDir": "./out"
    },
    "exclude": [
        "node_modules"
    ]
}

在vscode打开“文件”-“首选项”-“工作区设置” 在settings.json中加入(如果不做这个配置,vscode会在打开项目的时候提示选择哪个版本的typescript):

{
   "typescript.tsdk": "node_modules/typescript/lib"
}

创建api.ts文件,复制以下代码进去:

import superagent = require('superagent');
import cheerio = require('cheerio');

export const remote_get = function(url: string) {

    const promise = new Promise<superagent.Response>(function (resolve, reject) {
        superagent.get(url)
            .end(function (err, res) {
                if (!err) {
                    resolve(res);
                } else {
                    console.log(err)
                    reject(err);
                }
            });
    });
    return promise;
}

创建app.ts文件,书写测试代码:

import api = require('./api');
const go = async () => {
    let res = await api.remote_get('http://www.baidu.com/');
    console.log(res.text);
}
go();

执行命令: tsc然后: node out/app观察输出是否正确。

现在尝试抓取http://cnodejs.org/的第一页文章链接。 修改app.ts文件,代码如下:

import api = require('./api');
import cheerio = require('cheerio');

const go = async () => {
    const res = await api.remote_get('http://cnodejs.org/');
    const $ = cheerio.load(res.text);
    let urls: string[] = [];
    let titles: string[] = [];
    $('.topic_title_wrapper').each((index, element) => {
        titles.push($(element).find('.topic_title').first().text().trim());
        urls.push('http://cnodejs.org/' + $(element).find('.topic_title').first().attr('href'));
    })
    console.log(titles, urls);
}
go();

观察输出,文章的标题和链接都已获取到了。

现在尝试深入抓取文章内容

import api = require('./api');
import cheerio = require('cheerio');

const go = async () => {
    const res = await api.remote_get('http://cnodejs.org/');
    const $ = cheerio.load(res.text);
    $('.topic_title_wrapper').each(async (index, element) => {
        let url = ('http://cnodejs.org' + $(element).find('.topic_title').first().attr('href'));
        const res_content = await api.remote_get(url);
        const $_content = cheerio.load(res_content.text);
        console.log($_content('.topic_content').first().text());
    })

}
go();

可以发现因为访问服务器太迅猛,导致出现很多次503错误。 解决: 添加helper.ts文件:

export const wait_seconds = function (senconds: number) {
    return new Promise(resolve => setTimeout(resolve, senconds * 1000));
}

修改api.ts文件为:

import superagent = require('superagent');
import cheerio = require('cheerio');

export const get_index_urls = function () {
    const res = await remote_get('http://cnodejs.org/');
    const $ = cheerio.load(res.text);
    let urls: string[] = [];
    $('.topic_title_wrapper').each(async (index, element) => {
        urls.push('http://cnodejs.org' + $(element).find('.topic_title').first().attr('href'));
    });
    return urls;
}
export const get_content = async function (url: string) {
    const res = await remote_get(url);
    const $ = cheerio.load(res.text);
    return $('.topic_content').first().text();
}

export const remote_get = function (url: string) {

    const promise = new Promise<superagent.Response>(function (resolve, reject) {

        superagent.get(url)
            .end(function (err, res) {
                if (!err) {
                    resolve(res);
                } else {
                    console.log(err)
                    reject(err);
                }
            });
    });
    return promise;
}

修改app.ts文件为:

import api = require('./api');
import helper = require('./helper');
import cheerio = require('cheerio');

const go = async () => {
    const res = await api.remote_get('http://cnodejs.org/');
    const $ = cheerio.load(res.text);
    let urls = await api.get_index_urls();
    for (let i = 0; i < urls.length; i++) {
        await helper.wait_seconds(1);
        let text = await api.get_content(urls[i]);
        console.log(text);
    }
}
go();

观察输出可以看到,程序实现了隔一秒再请求下一个内容页。

现在尝试把抓取到的东西存到数据库中。 安装mongoose模块: npm i mongoose --savenpm i -s @types/mongoose --save然后建立Scheme。先创建models文件夹: mkdir models在models文件夹下创建index.ts:

import * as mongoose from 'mongoose';

mongoose.connect('mongodb://127.0.0.1/cnodejs_data', {
    server: { poolSize: 20 }
}, function (err) {
    if (err) {
        process.exit(1);
    }
});

// models
export const Article = require('./article');

在models文件夹下创建IArticle.ts:

interface IArticle {
    title: String;
    url: String;
    text: String;
}
export = IArticle;

在models文件夹下创建Article.ts:

import mongoose = require('mongoose');
import IArticle = require('./IArticle');
interface IArticleModel extends IArticle, mongoose.Document { }

const ArticleSchema = new mongoose.Schema({
    title: { type: String },
    url: { type: String },
    text: { type: String },
});

const Article = mongoose.model<IArticleModel>("Article", ArticleSchema);
export = Article;

修改api.ts为:

import superagent = require('superagent');
import cheerio = require('cheerio');
import models = require('./models');
const Article = models.Article;

export const get_index_urls = async function () {
    const res = await remote_get('http://cnodejs.org/');

    const $ = cheerio.load(res.text);
    let urls: string[] = [];
    $('.topic_title_wrapper').each((index, element) => {
        urls.push('http://cnodejs.org' + $(element).find('.topic_title').first().attr('href'));
    });
    return urls;

}
export const fetch_content = async function (url: string) {
    const res = await remote_get(url);

    const $ = cheerio.load(res.text);
    let article = new Article();
    article.text = $('.topic_content').first().text();
    article.title = $('.topic_full_title').first().text().replace('置顶', '').replace('精华', '').trim();
    article.url = url;
    console.log('获取成功:' + article.title);
    article.save();

}
export const remote_get = function (url: string) {

    return new Promise<superagent.Response>((resolve, reject) => {
        superagent.get(url)
            .end(function (err, res) {
                if (!err) {
                    resolve(res);
                } else {
                    reject(err);
                }
            });
    });
}

修改app.ts为:

import api = require('./api');
import helper = require('./helper');
import cheerio = require('cheerio');

(async () => {

    try {
        let urls = await api.get_index_urls();
        for (let i = 0; i < urls.length; i++) {
            await helper.wait_seconds(1);
            await api.fetch_content(urls[i]);
        }
    } catch (err) {
        console.log(err);
    }

    console.log('完毕!');

})();

执行 tscnode out/app观察输出,并去数据库检查一下 可以发现入库成功了!

补充:remote_get方法的改进版,实现错误重试和加入代理服务器. 放弃了superagent库,用的request库,仅供参考:

//config.retries = 3;
let current_retry = config.retries || 0;
export const remote_get = async function (url: string, proxy?: string) {
    //每次请求都先稍等一下
    await wait_seconds(2);
    if (!proxy) {
        proxy = '';
    }
    const promise = new Promise<string>(function (resolve, reject) {
        console.log('get: ' + url + ',  using proxy: ' + proxy);
        let options: request.CoreOptions = {
            headers: {
                'Cookie': '',
                'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/53.0.2785.143 Safari/537.36',
                'Referer': 'https://www.baidu.com/'
            },
            encoding: 'utf-8',
            method: 'GET',
            proxy: proxy,
            timeout: 3000,
        }
        request(url, options, async function (err, response, body) {
            console.log('got:' + url);
            if (!err) {
                body = body.toString();
                current_retry = config.retries || 0;
                console.log('bytes:' + body.length);
                resolve(body);
            } else {
                console.log(err);
                if (current_retry <= 0) {
                    current_retry = config.retries || 0;
                    reject(err);
                } else {
                    console.log('retry...(' + current_retry + ')')
                    current_retry--;
                    try {
                        let body = await remote_get(url, proxy);
                        resolve(body);
                    } catch (e) {
                        reject(e);
                    }
                }
            }
        });
    });
    return promise;
}

另外,IArticle.ts和Article.ts合并为一个文件,可能更好,可以参考我另一个model的写法:

import mongoose = require('mongoose');

interface IProxyModel {
    uri: string;
    ip: string;
    port:string;
    info:string;
}
export interface IProxy extends IProxyModel, mongoose.Document { }

const ProxySchema = new mongoose.Schema({
    uri: { type: String },//
    ip: { type: String },//
    port: { type: String },//
    info: { type: String },//
});
export const Proxy = mongoose.model<IProxy>("Proxy", ProxySchema);

导入的时候这么写就行了:

import { IProxy, Proxy } from './models';

其中Proxy可以用来做new、find、where之类的操作:

let x = new Proxy();
let xx = await Proxy.find({});
let xxx = await Proxy.where('aaa',123).exec();

而IProxy用于实体对象的传递,例如

function xxx(p:IProxy){
}

Viewing all articles
Browse latest Browse all 14821

Trending Articles