Web Scraping with Nokogiri and Mechanize - ruby

I am parsing prada.com and would like to scrape data in the div class "nextItem" and get its name and price. Here is my code:
require 'rubygems'
require 'mechanize'
require 'nokogiri'
require 'open-uri'
agent = Mechanize.new
page = agent.get('http://www.prada.com/en/US/e-store/department/woman/handbags.html?cmp=from_home')
fp = File.new('prada_prices','w')
html_doc = Nokogiri::HTML(page)
page = html_doc.xpath("//ol[#class='nextItem']")
page.each do {|i| fp.write(i.text + "\n")}
end
I get an error and no output. What I think I am doing is instantiating a mechanize object and calling it agent.
Then creating a page variable and assigning it the url provided.
Then creating a variable that is a nokogiri object with the mechanize url passed in
Then searching the url for all class references that are titled nextItem
Then printing all the data contained there
Can someone show me where I might have went wrong?

Since Prada's website dynamically loads its content via JavaScript, it will be hard to scrape its content. See "Scraping dynamic content in a website" for more information.
Generally speaking, with Mechanize, after you get a page:
page = agent.get(page_url)
you can easily search items with CSS selectors and scrape for data:
next_items = page.search(".fooClass")
next_items.each do |item|
price = item.search(".fooPrice").text
end
Then simply handle the strings or generate hashes as you desire.

Here are the wrong parts:
Check again the block syntax - use {} or do/end but not both in the same time.
Mechanize#get returns a Mechanize::Page which act as a Nokogiri document, at least it has search, xpath, css. Use them instead of trying to coerce the document to a Nokogiri::HTML object.
There is no need to require 'open-uri', and require 'nokogiri' when you are not using them directly.
Finally check maybe more about Ruby's basics before continuing with web scraping.
Here is the code with fixes:
require 'rubygems'
require 'mechanize'
agent = Mechanize.new
page = agent.get('http://www.prada.com/en/US/e-store/department/woman/handbags.html?cmp=from_home')
fp = File.new('prada_prices','w')
page = page.search("//ol[#class='nextItem']").each do |i|
fp.write(i.text + "\n")
end
fp.close

Related

want to get taobao's list of URL of products on search result page without taobao API

I want to get taobao's list of URL of products on search result page without taobao API.
I tried following Ruby script.
require "open-uri"
require "rubygems"
require "nokogiri"
url='https://world.taobao.com/search/search.htm?_ksTS=1517338530524_300&spm=a21bp.7806943.20151106.1&search_type=0&_input_charset=utf-8&navigator=all&json=on&q=%E6%99%BA%E8%83%BD%E6%89%8B%E8%A1%A8&cna=htqfEgp0pnwCATyQWEDB%2FRCE&callback=__jsonp_cb&abtest=_AB-LR517-LR854-LR895-PR517-PR854-PR895'
charset = nil
html = open(url) do |f|
charset = f.charset
f.read
end
doc = Nokogiri::HTML.parse(html, nil, charset)
p doc.xpath('//*[#id="list-itemList"]/div/div/ul/li[1]/div/div[1]/div/a/#href').each{|i| puts i.text}
# => 0
I want to get list of URL like https://click.simba.taobao.com/cc_im?p=%D6%C7%C4%DC%CA%D6%B1%ED&s=328917633&k=525&e=lDs3%2BStGrhmNjUyxd8vQgTvfT37ERKUkJtUYVk0Fu%2FVZc0vyfhbmm9J7EYm6FR5sh%2BLS%2FyzVVWDh7%2FfsE6tfNMMXhI%2B0UDC%2FWUl0TVvvELm1aVClOoSyIIt8ABsLj0Cfp5je%2FwbwaEz8tmCoZFXvwyPz%2F%2ByQnqo1aHsxssXTFVCsSHkx4WMF4kAJ56h9nOp2im5c3WXYS4sLWfJKNVUNrw%2BpEPOoEyjgc%2Fum8LOuDJdaryOqOtghPVQXDFcIJ70E1c5A%2F3bFCO7mlhhsIlyS%2F6JgcI%2BCdFFR%2BwwAwPq4J5149i5fG90xFC36H%2B6u9EBPvn2ws%2F3%2BHHXRqztKxB9a0FyA0nyd%2BlQX%2FeDu0eNS7syyliXsttpfoRv3qrkLwaIIuERgjVDODL9nFyPftrSrn0UKrE5HoJxUtEjsZNeQxqovgnMsw6Jeaosp7zbesM2QBfpp6NMvKM5e5s1buUV%2F1AkICwRxH7wrUN4%2BFn%2FJ0%2FIDJa4fQd4KNO7J5gQRFseQ9Z1SEPDHzgw%3D however I am getting 0
What should I do?
I don't know taobao.com but the page seems like its running lots of javascript. So perhaps the content can actually not be retrieved with a client without javascript capabilities. So instead of open-uri, you could try the gem selenium-webdriver:
https://rubygems.org/gems/selenium-webdriver/versions/2.53.4

Parsing large HTML files with Nokogiri

I'm trying to parse http://www.pro-medic.ru/index.php?ht=246&perpage=all with Nokogiri, but unfortunately I can't get all items from the page.
My simple test code is:
require 'open-uri'
require 'nokogiri'
html = Nokogiri::HTML open('http://www.pro-medic.ru/index.php?ht=246&perpage=all')
p html.css('ul.products-grid-compact li .goods_container').count
It returns only 83 items but the real count is about 186.
I thought that the problem could be in open, but it seems that function reads the HTML page correctly.
Has anybody faced the same problem?
The file seems to exceed Nokogiri's parser limits. You can relax the limits by adding the HUGE flag:
require 'open-uri'
require 'nokogiri'
url = 'http://www.pro-medic.ru/index.php?ht=246&perpage=all'
html = Nokogiri::HTML(open(url)) do |config|
config.options |= Nokogiri::XML::ParseOptions::HUGE
end
html.css('ul.products-grid-compact li .goods_container').count
#=> 186
Note that |= is a bitwise OR assignment operator, don't confuse it with the logical operator ||=
According to Parse Options, you can also set this flag via config.huge

How do I print XPath value?

I want to print the contents of an XPath node. Here is what I have:
require "mechanize"
agent = Mechanize.new
agent.get("http://store.steampowered.com/promotion/snowglobefaq")
puts agent.xpath("//*[#id='item_52b3985a70d58']/div[4]")
This returns: <main>: undefined method xpath for #<Mechanize:0x2fa18c0> (NoMethodError).
I just started using Mechanize and have no idea what I'm doing, however, I've used Watir and thought this would work but it didn't.
You an use Nokogiri to parse the page after retrieving it. Here is the example code:
m = Mechanize.new
result = m.get("http://google.com")
html = Nokogiri::HTML(result.body)
divs = html.xpath('//div').map { |div| div.content } # here you can do whatever is needed with the divs
# I've mapped their content into an array
There are two things wrong:
The ID doesn't exist on that page. Try this to see the list of tag IDs available:
require "open-uri"
require 'nokogiri'
doc = Nokogiri::HTML(open("http://store.steampowered.com/promotion/snowglobefaq"))
puts doc.search('[id*="item"]').map{ |n| n['id'] }.sort
The correct chain of methods is agent.page.xpath.
Because there is no sample HTML showing exactly which tag you want, we can't help you much.

Using Nokogiri to scrape a value from Yahoo Finance?

I wrote a simple script:
require 'rubygems'
require 'nokogiri'
require 'open-uri'
url = "http://au.finance.yahoo.com/q/bs?s=MYGN"
doc = Nokogiri::HTML(open(url))
name = doc.at_css("#yfi_rt_quote_summary h2").text
market_cap = doc.at_css("#yfs_j10_mygn").text
ebit = doc.at("//*[#id='yfncsumtab']/tbody/tr[2]/td/table[2]/tbody/tr/td/table/tbody/tr[11]/td[2]/strong").text
puts "#{name} - #{market_cap} - #{ebit}"
The script grabs three values from Yahoo finance. The problem is that the ebit XPath returns nil. The way I got the XPath was using the Chrome developer tools and copy and pasting.
This is the page I'm trying to get the value from http://au.finance.yahoo.com/q/bs?s=MYGN and the actual value is 483,992 in the total current assets row.
Any help would be appreciated, especially if there is a way to get this value with CSS selectors.
Nokogiri supports:
require 'nokogiri'
require 'open-uri'
doc = Nokogiri::HTML(open("http://au.finance.yahoo.com/q/bs?s=MYGN"))
ebit = doc.at('strong:contains("Total Current Assets")').parent.next_sibling.text.gsub(/[^,\d]+/, '')
puts ebit
# >> 483,992
I'm using the <strong> tag as an place-marker with the :contains pseudo-class, then backing up to the containing <td>, moving to the next <td> and grabbing its text, then finally stripping the white-space using gsub(/[^,\d]+/, '') which removes everything that isn't a number or a comma.
Nokogiri supports a number of jQuery's JavaScript extensions, which is why :contains works.
This seems to work for me
doc.css("table.yfnc_tabledata1 tr[11] td[2]").text.tr(",","").to_i
#=> 483992
Or as a string
doc.css("table.yfnc_tabledata1 tr[11] td[2]").text.strip.gsub(/\u00A0/,"")
#=> "483,992"

Why is Mechanize not following the link

I am trying to follow a link with Mechanize but it does not seem to be working, syntax appears to be correct, am I referencing this incorrectly or do I need to do something else?
Problem area
agent.page.links_with(:text => 'VG278H')[2].click
Full Code
require 'rubygems'
require 'mechanize'
require 'open-uri'
agent = Mechanize.new
agent.get ("http://icecat.biz/en/")
#Show all form fields belonging to the first form
form = agent.page.forms[0].fields
#Enter VG278H into the text box lookup_text, submit the data
agent.page.forms[0]["lookup_text"] = "VG278H"
agent.page.forms[0].submit #Results of this is stored in Mechanize agent.page object
#Call agent.page with our results and assign them to a variable page
page = agent.page
agent.page.links_with(:text => 'VG278H')[2].click
doc = page.parser
puts doc
You should grab a copy of Charles (http://www.charlesproxy.com/) or something that allows you to watch what happens when you submit the form from your browser. Anyway, your problem is that this part:
agent.page.forms[0]["lookup_text"] = "VG278H"
agent.page.forms[0].submit
is returning an html fragment that looks like this:
<meta http-equiv="Content-Type" content="text/html; charset=utf-8"><script>self.location.href="http://icecat.us/index.cgi?language=en&new_search=1&lookup_text=VG278H"</script>
So you actually need to call this directly or scrap out the the self.location.href and have your agent perform a get:
page = agent.get("http://icecat.us/index.cgi?language=en&new_search=1&lookup_text=VG278H")
If you were going to do that, this works:
require 'rubygems'
require 'mechanize'
require 'open-uri'
agent = Mechanize.new
agent.get ("http://icecat.biz/en/")
page = agent.get("http://icecat.us/index.cgi?language=en&new_search=1&lookup_text=VG278H")
page = page.links_with(:text => 'VG278H')[2].click
doc = page.parser
puts doc
Happy scraping

Resources