0.1

.gitignore

@@ -0,0 +1,3 @@
+build
+.git
+node_modules

.npmignore

@@ -0,0 +1,2 @@
+.git
+node_modules

README.md

@@ -0,0 +1 @@
+# 网页抓取器，带发邮件

dom.js

@@ -0,0 +1,111 @@
+var jsdom = require("jsdom");
+var async = require("async");
+var fs = require('fs');
+var log4js = require('log4js');
+var request = require('request');
+
+var send_mail = require('./mail.js');
+var house_name_array = require('./read_txt.js');
+
+
+
+
+
+
+
+var logger = log4js.getLogger('cheese');
+var jquery = fs.readFileSync("./deps/jquery.1.7.1.js").toString();
+
+
+
+var HOUSE_URL = 'http://www.szfcweb.com/szfcweb/DataSerach/CanSaleHouseSelectIndex.aspx'
+
+var get_count = function(){
+
+
+logger.info("send start");
+
+
+	jsdom.env({
+	  html:HOUSE_URL,
+	  src:[jquery],
+	  done:function (errors, window) {
+		  	if(errors) return logger.error(errors);
+		  	var jq = window.$;
+		  	var href_array =[];
+		  	var result_array = [];
+
+		  	var form_obj={}
+		  	jq('#aspnetForm input').each(function(){
+		  		form_obj[jq(this).attr('name')] = jq(this).val();
+		  	})
+
+		  	//console.log(form_obj)
+
+		  	//获得a标签的href数组
+		  	/*
+				解析a标签的jquery代码
+		  	*/
+		  		var async_array = [];
+		  		var reg = /共&nbsp\d+&nbsp条/
+
+				house_name_array.forEach(function(v){
+					async_array.push(function(callback){
+
+						var obj = JSON.parse(JSON.stringify(form_obj));
+						obj["ctl00$MainContent$txt_Pro"] = v.name
+						obj["ctl00$MainContent$ddl_houseclass"] = v.is_zz ? 'HC001' : 'HC002';
+						var r = request.post(HOUSE_URL,function(e,res,body){
+							if(e) return callback(e);
+							var zz_r = reg.exec(body);
+							if(zz_r){
+								zz_r = /\d+/.exec(zz_r[0]);
+								result_array.push({
+									name:v.name,
+									count:zz_r[0]
+								});
+
+							}
+							else{
+								result_array.push({
+									name:v.name,
+									count:-1
+								});
+							}
+							callback();
+						}).form(obj)
+
+					})
+				})
+
+
+				async.series(async_array,
+				// optional callback
+				function(err, results){
+					if(err) return logger.error(err);
+			    	var str = "";
+			    	result_array.forEach(function(v,i){
+			    		str += v.name +'\t'+v.count+'\n<br/>';
+			    	})
+
+			    	send_mail(str, function(err,response){ //去发送邮件
+			    		if(err) return logger.error(err);
+			    		//console.log(response)
+			    		logger.info("send end");
+			    	})
+
+				});
+
+	  }//end done function
+
+	});//end jsdom
+
+
+
+
+};
+
+
+
+
+exports.loop_fn = get_count;

house.txt

@@ -0,0 +1,9 @@
+逸庭
+玲珑湾
+独墅
+海悦
+熙岸
+------------
+东方之门大厦
+万宝商业广场
+乐嘉大厦

logs/cheese.log

@@ -0,0 +1,2 @@
+[2013-06-29 01:02:13.145] [INFO] cheese - send start
+[2013-06-29 01:02:23.809] [INFO] cheese - send end

loop.js

@@ -0,0 +1,36 @@
+var cluster = require('cluster');
+var loop_fn = require('./dom.js').loop_fn;
+var log4js = require('log4js');
+var loop_time = 1000*60*60*24;
+
+log4js.configure({
+  appenders: [
+    //{ type: 'console' },
+    { type: 'file', filename: './logs/cheese.log', category: 'cheese' }
+  ]
+});
+var logger = log4js.getLogger('cheese');
+
+
+if (cluster.isMaster) {
+  // Fork workers.
+
+    cluster.fork();
+
+
+  cluster.on('exit', function(worker, code, signal) {
+  	logger.fatal('worker ' + worker.process.pid + ' died');
+  	cluster.fork();
+  });
+} else {
+    loop_fn();
+	setInterval(function(){
+		loop_fn();
+	},loop_time)
+}
+
+
+
+
+
+

mail.js

@@ -0,0 +1,47 @@
+var tips = '选中复制的数字到excel,然后选中数字，选择 数据->分列，分隔符选择空格，下一步，搞定！！';
+var accouont = 'loushenghai1981@hotmail.com';
+var pwd = '123456lsh'
+var qq_acc = require('./qq_acc.json');
+
+
+
+var nodemailer = require("nodemailer");
+
+// create reusable transport method (opens pool of SMTP connections)
+var smtpTransport = nodemailer.createTransport("SMTP",{
+    service: "QQ",
+    auth: {
+        user: qq_acc.user,
+        pass: qq_acc.pass
+    }
+});
+
+
+
+var send_mail = function(string,callback){
+	var string = string || '';
+	var callback = callback || function(){};
+	//console.log(string)
+
+	// setup e-mail data with unicode symbols
+	var mailOptions = {
+	    from: "loushenghai1981 ✔ <"+qq_acc.user+">", // sender address
+	    to: "655028@qq.com, 53822985@qq.com", // list of receivers
+	    subject: "szhouse ✔", // Subject line
+	    //text: string, // plaintext body
+	    html: string+"<br/><br/><br/><b>"+tips+" ✔</b>" // html body
+	}
+
+	// send mail with defined transport object
+	smtpTransport.sendMail(mailOptions, function(error, response){
+	    callback(error,response);
+	    smtpTransport.close();
+	    // if you don't want to use this transport object anymore, uncomment following line
+	    //smtpTransport.close(); // shut down the connection pool, no more messages
+	});
+
+
+}
+
+
+module.exports = send_mail;

package.json

@@ -0,0 +1,34 @@
+{
+  "author": {
+    "name": "doublespout",
+    "email": "snoopyxdy@163.com",
+    "url": "http://snoopyxdy.blog.163.com"
+  },
+  "name": "szhouse",
+  "description": "szhouse reques",
+  "keywords": [
+    "node",
+    "szhouse"
+  ],
+  "version": "0.1.0",
+  "url": "https://github.com/DoubleSpout/szhouse",
+  "homepage": "https://github.com/DoubleSpout/szhouse",
+  "repository": {
+    "type": "git",
+    "url": "https://github.com/DoubleSpout/szhouse.git"
+  },
+  "main": "index.js",
+  "engines": {
+    "node":">=0.10.0"
+  },
+  "dependencies":{
+        "jsdom":"0.6.5"
+  },
+  "scripts": {
+    "tset":"node ./test/main.js",
+    "async":"0.2.9",
+    "log4js":"0.6.6",
+    "nodemailer":"0.4.4",
+    "request":"2.21.0"
+  }
+}

qq_acc.json

@@ -0,0 +1,4 @@
+{
+	"user":"",
+	"pass":""
+}

read_txt.js

@@ -0,0 +1,37 @@
+String.prototype.trim= function(){ 
+    // 用正则表达式将前后空格  
+    // 用空字符串替代。  
+    return this.replace(/(^\s*)|(\s*$)/g, "");  
+}
+
+var pos_str = '---';
+var fs = require('fs');
+var txt_content = fs.readFileSync("./house.txt").toString();
+var all_array = txt_content.split('\n').map(function(v,i){
+	if(~v.indexOf('--')) return pos_str
+	return v.trim();
+})
+var pos = all_array.indexOf(pos_str);
+
+var zhu_zhai = all_array.slice(0,pos);
+var fei_zhu_zhai = all_array.slice(pos+1);
+
+var house_name_array = [];
+
+zhu_zhai.forEach(function(v){
+	house_name_array.push({
+		name:v,
+		is_zz:1
+	})
+})
+
+fei_zhu_zhai.forEach(function(v){
+	house_name_array.push({
+		name:v,
+		is_zz:0
+	})
+})
+
+
+module.exports = house_name_array;
+