# Crawling Practice

> https://colab.research.google.com/drive/13xFkK9OyCTzGn1XTrlPL0t6tgGfUPA5s?authuser=2

- Reference: [https://news.naver.com/main/list.naver?mode=LSD&mid=sec&sid1=101&date=20200506&page=1](https://news.naver.com/main/list.naver?mode=LSD&mid=sec&sid1=101&date=20200506&page=1)


- To check HTML sauce code, press 'command'+'option'+'I'
![image.png](https://cdn.hashnode.com/res/hashnode/image/upload/v1659536796151/tJXRUM95J.png align="left")

- Url of an article is under `<li>`, `<dl>`.
![image.png](https://cdn.hashnode.com/res/hashnode/image/upload/v1659537009728/vWWK-8XNc.png align="left")
> To do crawling the URL, `soup.select('.type06_headline li dl')`


### Steps of the practice
1. Save article's URL as list form from Naver news page.
2. Crawling the articles by using titles and bodies with newspaper3k package.
3. Update dataframe once each crawling is done.
4. Once the crawling is done, save dataframe file having all the data as csv file.


- Install newspaper3k

```
!pip install newspaper3k
```
![image.png](https://cdn.hashnode.com/res/hashnode/image/upload/v1659783748479/8rhQokhd0.png align="left")

- Import packages

```
# Import package 'requests'
import requests

# Import module 'article' from package 'newspaper'
from newspaper import Article

# Import package 'pandas' and call it as pd.
import pandas as pd

# Import module 'BeautifulSoup' from package 'bs4'
from bs4 import BeautifulSoup
```

- to do crawling the website, have a look at the URL. Change the URL as a form of variables.

![image.png](https://cdn.hashnode.com/res/hashnode/image/upload/v1659783952020/_BbCE-MSF.png align="left")

> 'Page_num' variable indicates a page number.

>'Code' variable indicates a category.

>'Date' variable indicates a date.

>Since URL is a string variable, integer variables need to be changed to string by using `str()`.


```
page_num = 1
code = 101 
date = 20200506

url = 'https://news.naver.com/main/list.nhn?mode=LSD&mid=sec&sid1='+str(code)+'&date='+str(date)+'&page='+str(page_num) 
print(url)
```

**Result**
```
https://news.naver.com/main/list.nhn?mode=LSD&mid=sec&sid1=101&date=20200506&page=1
```

- Now we would like to get a URL for each an article. To do this, we need to get the entire HTML sauce code for the page by using `requests.get(url)`

```
#save the entire HTML of the URL as news
news = requests.get(url)
# print the HTML tag
news.content
```

> It shows an error message since Naver, the website is blocking unidentified user crawling the website.

- To avoid this, we need to pretend as a proper user by running the code below.

```
headers = {'user-agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/89.0.4389.114 Safari/537.36'}
news = requests.get(url, headers=headers)
```

- After that run the code below. And it shows the entire HTML code.

```
news = requests.get(url, headers=headers)
news.content
```
![image.png](https://cdn.hashnode.com/res/hashnode/image/upload/v1659786332153/cfAx0gfgh.png align="left")


- Now make a code with BeautifulSoup to return a URL by typing page number, category, and date.

```
# number of pages, category, date
def make_urllist(page_num, code, date): 
# make an empty list to save URLs of each article
  urllist= []

  # 1 ~ page_num까지 정해진 페이지만큼 반복.
  for i in range(1, page_num + 1):

		# 함수의 입력으로 된 변수들로 주소를 조합
    headers = {'user-agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/89.0.4389.114 Safari/537.36'}
    url = 'https://news.naver.com/main/list.nhn?mode=LSD&mid=sec&sid1='+str(code)+'&date='+str(date)+'&page='+str(i)   
    news = requests.get(url, headers=headers)

    # BeautifulSoup 모듈을 사용하여 HTML 페이지를 분석
    soup = BeautifulSoup(news.content, 'html.parser')

    # case1
    news_list = soup.select('.newsflash_body .type06_headline li dl')

    # case2
    news_list.extend(soup.select('.newsflash_body .type06 li dl'))
        
    # 뉴스 리스트에 있는 각 뉴스로부터 a 태그인 <a href ='주소'> 에서 '주소'만을 가져온다.
    for line in news_list:
        urllist.append(line.a.get('href'))
  return urllist
```