无法使用 Scrapy 生成数据

Question

I am trying to extract data from a site that has terrible html formatting, as all the info I want is in the same div and split with line breaks. I am new to web scraping in general, so please bear with me.

https://wsldata.com/directory/record.cfm?LibID=48

In order to get the parts I need, I use:

details_raw = response.xpath('/html/body/div/table/tbody/tr/td/div/div/text()').getall()

returns

['\r\n',
 '\r\n',
 '\r\n',
 '\r\n      \r\n      ',
 '\r\n\t\t\t',
 '\r\n      ',
 '\r\n      ',
 '\r\n      ',
 '\r\n\t\t\t\r\n\t\t\t',
 '\r\n      \r\n      ',
 '\r\n\t\t\tDirector',
 '\r\n       Ext: 5442',
 '\r\n      ',
 '\r\n      ',
 '\r\n\t\t\t\r\n\t\t\t',
 '\r\n      \r\n      ',
 '\r\n\t\t\tAssistant Library Director',
 '\r\n       Ext: 5433',
 '\r\n      ',
 '\r\n      ',
 '\r\n\t\t\t\r\n\t\t\t',
 '\r\n      \r\n      ',
 '\r\n\t\t\tYouth Services Librarian',
 '\r\n      ',
 '\r\n      ',
 '\r\n      ',
 '\r\n\t\t\t\r\n\t\t\t',
 '\r\n      \r\n      ',
 '\r\n\t\t\tTechnical Services Librarian',
 '\r\n       Ext: 2558',
 '\r\n      ',
 '\r\n      ',
 '\r\n\t\t\t\r\n\t\t\t',
 '\r\n      \r\n      ',
 '\r\n\t\t\tOutreach Librarian',
 '\r\n      ',
 '\r\n      ',
 '\r\n      ',
 '\r\n\t\t\t\r\n\t\t\t',
 '\r\n      \r\n      ',
 '\r\n\t\t\tFoundation Executive Director',
 '\r\n       Ext: 5456',
 '\r\n      ',
 '\r\n      ',
 '\r\n\t\t\t\r\n\t\t\t',
 '\r\n      \r\n',
 '\r\n',
 ' \xa0|\xa0 ',
 '\r\n']

I have managed to bring that to a desired format using the following code

import scrapy
import re

class LibspiderSpider(scrapy.Spider):
    name = 'libspider'
    allowed_domains = ['wsldata.com']
    start_urls = ['https://wsldata.com/directory/record.cfm?LibID=48']
    # Note that start_urls contains multiple links, I just simplified it here to reduce cluttering
    
    def parse(self, response):        
           
        details_raw = response.xpath('/html/body/div/table/tbody/tr/td/div/div/text()').getall()
        
        details_clean = []
        titles = []
        details = []

        for detail in details_raw:
            detail = re.sub(r'\t', '', detail)
            detail = re.sub(r'\n', '', detail)
            detail = re.sub(r'\r', '', detail)
            detail = re.sub(r'  ', '', detail)
            detail = re.sub(r' \xa0|\xa0 ', '', detail)
            detail = re.sub(r'|', '', detail)
            detail = re.sub(r' E', 'E', detail)
            if detail == '':
                pass
            elif detail == '|':
                pass
            else:
                details_clean.append(detail)
                if detail[0:3] != 'Ext':
                    titles.append(detail)

        for r in range(len(details_clean)):
            if r == 0:
                details.append(details_clean[r])  
            else:
                if details_clean[r-1][0:3] != 'Ext' and details_clean[r][0:3] != 'Ext':
                    details.append('-')
                    details.append(details_clean[r])
                else:
                    details.append(details_clean[r])
                    
        output = []
        for t in range(len(details)//2):  
            info = {
                "title": details[(t*2)],
                "phone": details[(t*2+1)],
            }
            output.append(info)

The block of code after the response.xpath line is used to clean my input to a nicer output. When testing the code outside of scrapy, using the weird input I showed on the top of post, I get:

[{'title': 'Director', 'phone': 'Ext: 5442'}, {'title': 'Assistant Library Director', 'phone': 'Ext: 5433'}, {'title': 'Youth Services Librarian', 'phone': '-'}, {'title': 'Technical Services Librarian', 'phone': 'Ext: 2558'}, {'title': 'Outreach Librarian', 'phone': '-'}, {'title': 'FoundationExecutive Director', 'phone': 'Ext: 5456'}]

When I try to implement this code into scrapy's parse(), my log doesn't show any items scraped and I obviously get an empty json.

yield is not present in the above code, as I have tried multiple ways to implement it and none of them worked. Am I missing a connection between scrapy's response and yield or is what I am trying to do not possible and should just extract the weird list and work it off scrapy like so:

    def parse(self, response):        
           
        details_raw = response.xpath('/html/body/div/table/tbody/tr/td/div/div/text()').getall()
        yield{
            'details_in' : details_raw
        }

which extracts:

[
{"details_in": ["\r\n", "\r\n", "\r\n", "\r\n      \r\n      ", "\r\n\t\t\t", "\r\n      ", "\r\n      ", "\r\n      ", "\r\n\t\t\t\r\n\t\t\t", "\r\n      \r\n      ", "\r\n\t\t\tDirector", "\r\n       Ext: 5442", "\r\n      ", "\r\n      ", "\r\n\t\t\t\r\n\t\t\t", "\r\n      \r\n      ", "\r\n\t\t\tAssistant Library Director", "\r\n       Ext: 5433", "\r\n      ", "\r\n      ", "\r\n\t\t\t\r\n\t\t\t", "\r\n      \r\n      ", "\r\n\t\t\tYouth Services Librarian", "\r\n      ", "\r\n      ", "\r\n      ", "\r\n\t\t\t\r\n\t\t\t", "\r\n      \r\n      ", "\r\n\t\t\tTechnical Services Librarian", "\r\n       Ext: 2558", "\r\n      ", "\r\n      ", "\r\n\t\t\t\r\n\t\t\t", "\r\n      \r\n      ", "\r\n\t\t\tOutreach Librarian", "\r\n      ", "\r\n      ", "\r\n      ", "\r\n\t\t\t\r\n\t\t\t", "\r\n      \r\n      ", "\r\n\t\t\tFoundation Executive Director", "\r\n       Ext: 5456", "\r\n      ", "\r\n      ", "\r\n\t\t\t\r\n\t\t\t", "\r\n      \r\n", "\r\n", " \u00a0|\u00a0 ", "\r\n"]},
{"details_in": ["\r\n", "\r\n", "\r\n", "\r\n      \r\n      ", "\r\n\t\t\tBranch Librarian", "\r\n      ", "\r\n      ", "\r\n      ", "\r\n\t\t\t\r\n\t\t\t", "\r\n      \r\n", "\r\n", " \u00a0|\u00a0 ", "\r\n"]},
...
...
]

Answer 1

If you want to remove those lines from the list you can use this (instead of regex):

>>> lst=['\r\n',
...  '\r\n',
...  '\r\n',
...  '\r\n      \r\n      ',
...  '\r\n\t\t\t',
...  '\r\n      ',
...  '\r\n      ',
...  '\r\n      ',
...  '\r\n\t\t\t\r\n\t\t\t',
...  '\r\n      \r\n      ',
...  '\r\n\t\t\tDirector',
...  '\r\n       Ext: 5442',
...  '\r\n      ',
...  '\r\n      ',
...  '\r\n\t\t\t\r\n\t\t\t',
...  '\r\n      \r\n      ',
...  '\r\n\t\t\tAssistant Library Director',
...  '\r\n       Ext: 5433',
...  '\r\n      ',
...  '\r\n      ',
...  '\r\n\t\t\t\r\n\t\t\t',
...  '\r\n      \r\n      ',
...  '\r\n\t\t\tYouth Services Librarian',
...  '\r\n      ',
...  '\r\n      ',
...  '\r\n      ',
...  '\r\n\t\t\t\r\n\t\t\t',
...  '\r\n      \r\n      ',
...  '\r\n\t\t\tTechnical Services Librarian',
...  '\r\n       Ext: 2558',
...  '\r\n      ',
...  '\r\n      ',
...  '\r\n\t\t\t\r\n\t\t\t',
...  '\r\n      \r\n      ',
...  '\r\n\t\t\tOutreach Librarian',
...  '\r\n      ',
...  '\r\n      ',
...  '\r\n      ',
...  '\r\n\t\t\t\r\n\t\t\t',
...  '\r\n      \r\n      ',
...  '\r\n\t\t\tFoundation Executive Director',
...  '\r\n       Ext: 5456',
...  '\r\n      ',
...  '\r\n      ',
...  '\r\n\t\t\t\r\n\t\t\t',
...  '\r\n      \r\n',
...  '\r\n',
...  ' \xa0|\xa0 ',
...  '\r\n']
>>> newlst = [i.strip() for i in lst if i.strip()]
>>> newlst
['Director', 'Ext: 5442', 'Assistant Library Director', 'Ext: 5433', 'Youth Services Librarian', 'Technical Services Librarian', 'Ext: 2558', 'Outreach Librarian', 'Foundation Executive Director', 'Ext: 5456', '|']

You can achieve the result you want by using the correct xpath selectors :

import scrapy


class LibspiderSpider(scrapy.Spider):
    name = 'libspider'
    allowed_domains = ['wsldata.com']
    start_urls = ['https://wsldata.com/directory/record.cfm?LibID=48']

    def parse(self, response):
        details_raw = response.xpath('//div[@class="main"]//div[@style="margin:16px 8px;"]')
        if details_raw:
            details_raw = details_raw[:-1]

        for detail in details_raw:
            item = dict()
            item['title'] = detail.xpath('./following-sibling::br[1]/following::text()').get(default='').strip()
            item['phone'] = detail.xpath('./following-sibling::br[2]/following::text()').get(default='-').strip()
            yield item

The xpath selectors look like this because like you said it's:

a site that has terrible html formatting

I'm sure that you can find another xpath selectors that will fit your needs, but this one isn't terrible =).