강상위 / my-broadcasting

Go to a project
Toggle navigation Toggle navigation pinning
김승훈
8ec7e6bc 1 parent 9ad7029f

add crawling_time,samename,more_data

Inline Side-by-side
Showing 3 changed files with 183 additions and 12 deletions
var async = require('async');
var webdriver = require('selenium-webdriver');
var options = {desiredCapabilities: {browserName: 'chrome'}};
var JASON = require('jason')
const {Builder, By, Key, until} = require('selenium-webdriver');
var iconv = require('iconv-lite')
var driver = new webdriver.Builder().
withCapabilities(webdriver.Capabilities.chrome()).
build();
let url1 = "https://search.naver.com/search.naver?sm=top_hty&fbm=1&ie=utf8&query=%EB%9F%B0%EB%8B%9D%EB%A7%A8"
driver
.get(url1).then(() => {
driver
.findElements(webdriver.By.className('brcs_detail'))
.then(broad_info => {
broad_info[0].findElement(webdriver.By.tagName('dl'))
.then(more_data => {
more_data.findElement(webdriver.By.tagName('dd'))
.then(one_line => {
one_line.findElement(webdriver.By.className("inline"))
.then(isbroad => {
isbroad.getText()
.then(isbroad_string => {
console.log(isbroad_string)
broadcasting=isbroad_string.substring(0,4)
isbroad = isbroad_string.substring(18,21)
br_date = isbroad_string.substring(23,24)
br_date = br_date + "요일"
br_time = isbroad_string.substring(26,34)
console.log(broadcasting,isbroad,br_date,br_time)
if(isbroad == "방영중") {
driver.findElement(webdriver.By.id('nx_query')).clear().then(( )=> {
//var buf = iconv.encode("편성표", "euc-kr")
//console.log(buf)
//var encodestr='';
//for(var i=0; i<buf.length;i++) {
// encodestr+='%'+buf[i].toString('16')
// }
//encodestr = encodestr.toUpperCase();
//console.log(iconv.encode(encodestr, 'EUC-KR').toString())
//console.log(encodestr)
var buf = iconv.encode('편성표','euckr')
var param = buf.toString('binary')
console.log(param)
driver.findElement(webdriver.By.id('nx_query')).sendKeys(broadcasting,br_date,param)
driver.findElement(webdriver.By.className('spnew ico_search')).click()
})
}
else {
}
})
})
})
})
})
})
......@@ -4,6 +4,8 @@ var options = {desiredCapabilities: {browserName: 'chrome'}};
var JASON = require('jason')
const {Builder, By, Key, until} = require('selenium-webdriver');
var ajason = new Object()
var aname = []
var asrc=[]
......@@ -11,7 +13,11 @@ var alink=[]
var driver = new webdriver.Builder().
withCapabilities(webdriver.Capabilities.chrome()).
build();
let url1 = "https://people.search.naver.com/search.naver?sm=sbx_hty&where=nexearch&ie=utf8&query=%EC%9C%A0%EC%9E%AC%EC%84%9D&x=0&y=0"
let url1 = "https://people.search.naver.com/search.naver?sm=tab_hty&where=nexearch&query=%EC%9C%A0%EC%9E%AC%EC%84%9D&ie=utf8&x=0&y=0"
var driver_1 = new webdriver.Builder().
withCapabilities(webdriver.Capabilities.chrome()).
build();
driver
.get(url1).then(() => {
driver
......@@ -25,21 +31,13 @@ driver
img_src[0].getAttribute("src")
.then(src => {
if(src) {
asrc.push(src)
}
else {
asrc.push("none")
}
asrc.push(src)
console.log(src)
}).then(() => {
console.log(asrc,alink,aname)
console.log(asrc)
})
})
})
whoclass.findElements(webdriver.By.tagName("strong"))
.then(name_temp => {
name_temp[0].getText().then(name => {
......@@ -54,6 +52,86 @@ driver
findname[0].getAttribute("href")
.then(link => {
alink.push(link)
}).then(() => {
driver_1.get(alink[0]).then(() => {
driver_1
.findElement(webdriver.By.id('pagination_76'))
.then(paginationBtn => {
paginationBtn.findElements(webdriver.By.className('bt_next'))
.then(elemsBtn => {
var cnt = 1;
function getContentsAndClickNext (callback) {
console.log('higetcontests', cnt);
cnt++;
driver_1
.findElement(webdriver.By.id('listUI_76'))
.then(contentsUI => {
contentsUI
.findElements(webdriver.By.tagName('li'))
.then(elems => {
elems.forEach(elem => {
elem
.getText()
.then(text => {
console.log(text);
// 내 추측 : stale해지면, boolean 값으로 driver 어딘가에 true설정되는것 같아
// wait이라는 함수를 써서, 이게 false 가 될 때까지 기다린 다음에 getText()를 해.
// ok?
})
.catch(error => {
// 문제가 있으면 이곳을 의심해볼것! 변수 error를 체크해보기!
// console.log('really?? exception!');
})
})
elemsBtn[0].click();
sleep(50).then(() => {
// Do something after the sleep!
});
callback();
})
.catch(error => {
console.log("asdfasdf!!', e");
})
})
.catch(error => {
console.log('fuck!@#', error);
})
}
async.whilst(
function() {
console.log('whilist result', cnt < 5);
return cnt < 5;
},
getContentsAndClickNext,
function(e) {
console.log('Exception', e)
}
);
})
.catch(error => {
console.log('Exception 4444', error);
}); // 다음 컨텐츠로 가는 버튼(2개임. 하나는 평소용 두번째는 더이상 갈 페이지가 없을 때 쓰는 버튼)
})
.catch(error => {
console.log('Exception 555', error);
}); // 방송 컨텐츠 페이지네이션 버튼
// sleep time expects milliseconds
function sleep (time) {
return new Promise((resolve) => setTimeout(resolve, time));
}
// Usage!
sleep(8000).then(() => {
// Do something after the sleep!
});
});
})
})
})
......
var async = require('async');
var webdriver = require('selenium-webdriver');
var options = {desiredCapabilities: {browserName: 'chrome'}};
var JASON = require('jason')
const {Builder, By, Key, until} = require('selenium-webdriver');
var iconv = require('iconv-lite')
var findtitle = "런닝맨"
var driver = new webdriver.Builder().
withCapabilities(webdriver.Capabilities.chrome()).
build();
let url1 = "https://search.naver.com/search.naver?sm=tab_hty.top&where=nexearch&query=SBS+%EC%9D%BC%EC%9A%94%EC%9D%BC+%ED%8E%B8%EC%84%B1%ED%91%9C&oquery=SBS+%EC%9B%94%EC%9A%94%EC%9D%BC+%ED%8E%B8%EC%84%B1%ED%91%9C&tqi=UsDTSlpySD0ssv33OfVssssssXK-499490"
driver
.get(url1).then(() => {
driver
.findElements(webdriver.By.className('cont_inner type_day _scheduleArea'))
.then(contentsearch_section=> {
contentsearch_section[0].findElements(webdriver.By.tagName('tbody'))
.then(tbody => {
tbody[0].findElements(webdriver.By.tagName('tr'))
.then(trs => {
trs.forEach(tr => {
tr.getText()
.then(content => {
if(content.indexOf(findtitle) != -1) {
console.log(content)
}
})
})
})
})
})
})