Scraping with Beautiful Soup does not update values properly - python

I try to web-scrape weather website but the data does not update properly. The code:
from urllib.request import urlopen
from bs4 import BeautifulSoup
url = 'https://www.wunderground.com/dashboard/pws/KORPISTO1'
while True:
soup = BeautifulSoup(urlopen(url), 'html.parser')
data = soup.find("div", {"class": "weather__text"})
print(data.text)
I am looking at 'WIND & WIND GUST' in 'CURRENT CONDITIONS' section. It prints the first values correctly (for example 1.0 / 2.2 mph) but after that the values update very slowly (at times 5+ minutes pass by) even though they change every 10-20-30 seconds in the website.
And when the values update in Python they are still different from the current values in the website.

You could try this alternate method: since the site actually retrieves the data from another url, you could just directly make the request and scrape the site only every hour or so to update the request url.
from urllib.request import urlopen
from bs4 import BeautifulSoup
import json
from datetime import datetime, timedelta
#def getReqUrl...
reqUrl = getReqUrl()
prevTime, prevAt = '', datetime.now()
while True:
ures = json.loads(urlopen(reqUrl).read())
if 'observations' not in asd:
reqUrl = getReqUrl()
ures = json.loads(urlopen(reqUrl).read())
#to see time since last update
obvTime = ures['observations'][0]['obsTimeUtc']
td = (datetime.now() - prevAt).seconds
wSpeed = ures['observations'][0]['imperial']['windSpeed']
wGust = ures['observations'][0]['imperial']['windGust']
print('',end=f'\r[+{td}s -> {obvTime}]: {wGust} ° / {wSpeed} °mph')
if prevTime < obvTime:
prevTime = obvTime
prevAt = datetime.now()
print('')
Even when making the request directly, the "observation time" in the retrieved data jumps around sometimes, which is why I'm only printing on a fresh line when obvTime increases - without that, it looks like this. (If that's preferred you can just print normally without the '',end='\r... format, and the second if block is no longer necessary either).
The first if block is for refreshing the reqUrl (because it expires after a while), which is when I actually scrape the wunderground site, because the url is inside one of their script tags:
def getReqUrl():
url = 'https://www.wunderground.com/dashboard/pws/KORPISTO1'
soup = BeautifulSoup(urlopen(url), 'html.parser')
appText = soup.select_one('#app-root-state').text
nxtSt = json.loads(appText.replace('&q;','"'))['wu-next-state-key']
return [
ns for ns in nxtSt.values()
if 'observations' in ns['value'] and
len(ns['value']['observations']) == 1
][0]['url'].replace('&a;','&')
or, since I know how the url starts, more simply like:
def getReqUrl():
url = 'https://www.wunderground.com/dashboard/pws/KORPISTO1'
soup = BeautifulSoup(urlopen(url), 'html.parser')
appText = soup.select_one('#app-root-state').text
rUrl = 'https://api.weather.com/v2/pws/observations/current'
rUrl = rUrl + appText.split(rUrl)[1].split('&q;')[0]
return rUrl.replace('&a;','&')

try:
import requests
from bs4 import BeautifulSoup
url = 'https://www.wunderground.com/dashboard/pws/KORPISTO1'
headers = {'User-Agent': 'Mozilla/5.0 (X11; Ubuntu; Linux x86_64; rv:76.0) Gecko/20100101 Firefox/76.0'}
session = requests.Session()
r = session.get(url, timeout=30, headers=headers) # print(r.status_code)
soup = BeautifulSoup(r.content, 'html.parser')
#'WIND & WIND GUST' in 'CURRENT CONDITIONS' section
wind_gust = [float(i.text) for i in soup.select_one('.weather__header:-soup-contains("WIND & GUST")').find_next('div', class_='weather__text').select('span.wu-value-to')]
print(wind_gust)
[1.8, 2.2]
wind = wind_gust[0]
gust = wind_gust[1]
print(wind)
1.8
print(gust)
2.2

Related

Scraping a website with a particular format using Python

I am trying to use Python to scrape the US News Ranking for universities, and I'm struggling. I normally use Python "requests" and "BeautifulSoup".
The data is here:
https://www.usnews.com/education/best-global-universities/rankings
Using right click and inspect shows a bunch of links and I don't even know which one to pick. I followed an example from the web that I found but it just gives me empty data:
import requests
import urllib.request
import time
from bs4 import BeautifulSoup
import pandas as pd
import math
from lxml.html import parse
from io import StringIO
url = 'https://www.usnews.com/education/best-global-universities/rankings'
urltmplt = 'https://www.usnews.com/education/best-global-universities/rankings?page=2'
css = '#resultsMain :nth-child(1)'
npage = 20
urlst = [url] + [urltmplt + str(r) for r in range(2,npage+1)]
def scrapevec(url, css):
doc = parse(StringIO(url)).getroot()
return([link.text_content() for link in doc.cssselect(css)])
usng = []
for u in urlst:
print(u)
ts = [re.sub("\n *"," ", t) for t in scrapevec(u,css) if t != ""]
This doesn't work as t is an empty array.
I'd really appreciate any help.
The MWE you posted is not working at all: urlst is never defined and cannot be called. I strongly suggest you to look for basic scraping tutorials (with python, java, etc.): there is plenty and in general is a good starting.
Below you can find a snippet of a code that prints the universities' names listed on page 1 - you'll be able to extend the code to all the 150 pages through a for loop.
import requests
from bs4 import BeautifulSoup
newheaders = {
'User-Agent': 'Mozilla/5.0 (X11; Linux i686 on x86_64)'
}
baseurl = 'https://www.usnews.com/education/best-global-universities/rankings'
page1 = requests.get(baseurl, headers = newheaders) # change headers or get blocked
soup = BeautifulSoup(page1.text, 'lxml')
res_tab = soup.find('div', {'id' : 'resultsMain'}) # find the results' table
for a,univ in enumerate(res_tab.findAll('a', href = True)): # parse universities' names
if a < 10: # there are 10 listed universities per page
print(univ.text)
Edit: now the example works, but as you say in your question, it only returns empty lists. Below an edited version of the code that returns a list of all universities (pp. 1-150)
import requests
from bs4 import BeautifulSoup
def parse_univ(url):
newheaders = {
'User-Agent': 'Mozilla/5.0 (X11; Linux i686 on x86_64)'
}
page1 = requests.get(url, headers = newheaders) # change headers or get blocked
soup = BeautifulSoup(page1.text, 'lxml')
res_tab = soup.find('div', {'id' : 'resultsMain'}) # find the results' table
res = []
for a,univ in enumerate(res_tab.findAll('a', href = True)): # parse universities' names
if a < 10: # there are 10 listed universities per page
res.append(univ.text)
return res
baseurl = 'https://www.usnews.com/education/best-global-universities/rankings?page='
ll = [parse_univ(baseurl + str(p)) for p in range(1, 151)] # this is a list of lists
univs = [item for sublist in ll for item in sublist] # unfold the list of lists
Re-edit following QHarr suggestion (thanks!) - same output, shorter and more "pythonic" solution
import requests
from bs4 import BeautifulSoup
def parse_univ(url):
newheaders = {
'User-Agent': 'Mozilla/5.0 (X11; Linux i686 on x86_64)'
}
page1 = requests.get(url, headers = newheaders) # change headers or get blocked
soup = BeautifulSoup(page1.text, 'lxml')
res_tab = soup.find('div', {'id' : 'resultsMain'}) # find the results' table
return [univ.text for univ in res_tab.select('[href]', limit=10)]
baseurl = 'https://www.usnews.com/education/best-global-universities/rankings?page='
ll = [parse_univ(baseurl + str(p)) for p in range(1, 151)] # this is a list of lists
univs = [item for sublist in ll for item in sublist]

Not able to scrape the all the reviews

I am trying to scrape this website and trying to get the reviews but I am facing an issue,
The page loads only 50 reviews.
To load more you have to click "Show More Reviews" and I don't know how to get all the data as there is no page link, also "Show more Reviews" doesn't have a URL to explore, the address remains the same.
url =
"https://www.capterra.com/p/134048/HiMama-Preschool-Child-Care-App/#reviews"
import requests
from bs4 import BeautifulSoup
import json
import pandas as pd
a = []
url = requests.get(url)
html = url.text
soup = BeautifulSoup(html, "html.parser")
table = soup.findAll("div", {"class":"review-comments"})
#print(table)
for x in table:
a.append(x.text)
df = pd.DataFrame(a)
df.to_csv("review.csv", sep='\t')
I know this is not pretty code but I am just trying to get the review text first.
kindly help. As I am little new to this.
Looking at the website, the "Show more reviews" button makes an ajax call and returns the additional info, all you have to do is find it's link and send a get request to it (which I've done with some simple regex):
import requests
import re
from bs4 import BeautifulSoup
headers = {
"user-agent": "Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) snap Chromium/74.0.3729.169 Chrome/74.0.3729.169 Safari/537.36"
}
url = "https://www.capterra.com/p/134048/HiMama-Preschool-Child-Care-App/#reviews"
Data = []
#Each page equivalant to 50 comments:
MaximumCommentPages = 3
with requests.Session() as session:
info = session.get(url)
#Get product ID, needed for getting more comments
productID = re.search(r'"product_id":(\w*)', info.text).group(1)
#Extract info from main data
soup = BeautifulSoup(info.content, "html.parser")
table = soup.findAll("div", {"class":"review-comments"})
for x in table:
Data.append(x)
#Number of pages to get:
#Get additional data:
params = {
"page": "",
"product_id": productID
}
while(MaximumCommentPages > 1): # number 1 because one of them was the main page data which we already extracted!
MaximumCommentPages -= 1
params["page"] = str(MaximumCommentPages)
additionalInfo = session.get("https://www.capterra.com/gdm_reviews", params=params)
print(additionalInfo.url)
#print(additionalInfo.text)
#Extract info for additional info:
soup = BeautifulSoup(additionalInfo.content, "html.parser")
table = soup.findAll("div", {"class":"review-comments"})
for x in table:
Data.append(x)
#Extract data the old fashioned way:
counter = 1
with open('review.csv', 'w') as f:
for one in Data:
f.write(str(counter))
f.write(one.text)
f.write('\n')
counter += 1
Notice how I'm using a session to preserve cookies for the ajax call.
Edit 1: You can reload the webpage multiple times and call the ajax again to get even more data.
Edit 2: Save data using your own method.
Edit 3: Changed some stuff, now gets any number of pages for you, saves to file with good' ol open()

Using requests in a loop provides different HTML than using it outside of the loop

I'm attempting to use Python to scrape historic exchange rates from a website.
If I scrape the site manually:
url = "https://www.x-rates.com/historical/?from=USD&amount=1&date=2018-07-12"
page = requests.get(url, timeout=5)
soup = BeautifulSoup(page.content, "html.parser")
table = soup.find("tbody")
The result is the correct HTML I'm looking for.
However, if I use the same block inside this loop:
for d in date_generated:
date = str(d).replace("00:00:00", "")
url = "https://www.x-rates.com/historical/?from=USD&amount=1&date=" + date
page = requests.get(url, timeout=5)
soup = BeautifulSoup(page.content, "html.parser")
table = soup.find("tbody")
for i,x in zip(table.find_all("a"), table.find_all("td", class_="")):
time.sleep(3)
request += 1
elapsed_time = time.time() - start_time
print(i.text.strip(), x.text.strip())
I don't get the correct HTML content. The generated URL's work, and I get a 200 status_code but a NoneType object is returned at the second loop when it attempts to use the table, indicating that it never got the HTML I wanted in the first place (But I do get it if I run it out of the loop).
For those asking, here's how the dates are generated.
start_time = time.time()
start = datetime.datetime.strptime("2018-07-07", "%Y-%m-%d")
end = datetime.datetime.strptime("2018-07-12", "%Y-%m-%d")
date_generated = [start + datetime.timedelta(days=p) for p in range(0, (end-
start).days)]
for the sake of this answer lets say that the value of d is equal to 2018-07-07 00:00:00.
Nonetheless, when you run:
date = str(d).replace("00:00:00", "")
you forget the space between 2018-07-07 and 00:00:00.
So you should replace this line with:
date = str(d).replace(" 00:00:00", "")
Hope this solves your problem :)
Websites will limit the number of requests. In order to solve this, you need to specify request headers. Try this:
for d in date_generated:
date = str(d).replace("00:00:00", "")
url = "https://www.x-rates.com/historical/?from=USD&amount=1&date=" + date
head = {'user-agent': ('Mozilla/5.0 (X11; Linux x86_64)'
'AppleWebKit/537.36 (KHTML, like Gecko)'
'Chrome/66.0.3359.139 Safari/537.36'),
'referer': None}
head['referer'] = url
page = requests.get(url, timeout=5, headers=head)
You might be overloading the website, sending too many requests as once and being put on a timeout. Try placing a time.sleep(1) in the loop, as #hootnot suggested. If (1) isn't enough, try a longer timeout.

Unable to expand more... python

I can scrape all the reviews from the web page.But I am not getting full content.Only half review content i can scrape.I need to scrape the full content.
from bs4 import BeautifulSoup import requests import re
s = requests.Session()
def get_soup(url):
headers = {'User-Agent': 'Mozilla/5.0 (X11; Ubuntu; Linux x86_64; rv:57.0) Gecko/20100101 Firefox/57.0'}
r = s.get(url, headers=headers)
#with open('temp.html', 'wb') as f:
# f.write(r.content)
# webbrowser.open('temp.html')
if r.status_code != 200:
print('status code:', r.status_code)
else:
return BeautifulSoup(r.text, 'html.parser')
def parse(url, response):
if not response:
print('no response:', url)
return
# get number of reviews
# num_reviews = response.find('span', class_='reviews_header_count').text
# num_reviews = num_reviews[1:-1] # remove `( )`
# num_reviews = num_reviews.replace(',', '') # remove `,`
# num_reviews = int(num_reviews)
# print('num_reviews:', num_reviews, type(num_reviews))
num_reviews = (20)
# num_reviews = num_reviews[1:-1] # remove `( )`
# num_reviews = num_reviews.replace(',', '') # remove `,`
# num_reviews = int(num_reviews)
print('num_reviews:', num_reviews, type(num_reviews))
# create template for urls to pages with reviews
url = url.replace('Hilton_New_York_Grand_Central-New_York_City_New_York.html', 'or{}-Hilton_New_York_Grand_Central-New_York_City_New_York.html')
print('template:', url)
# add requests to list
for offset in range(0, num_reviews, 5):
print('url:', url.format(offset))
url_ = url.format(offset)
parse_reviews(url_, get_soup(url_))
#return # for test only - to stop after first page
def parse_reviews(url, response):
print('review:', url)
if not response:
print('no response:', url)
return
for idx, review in enumerate(response.find_all('div', class_='review-container')):
item = {
'hotel_name': response.find('h1', class_='heading_title').text,
'review_title': review.find('span', class_='noQuotes').text,
'review_body': review.find('p', class_='partial_entry').text,
'review_date': review.find('span', class_='relativeDate')['title'],#.text,#[idx],
# 'num_reviews_reviewer': review.find('span', class_='badgetext').text,
'reviewer_name': review.find('span', class_='scrname').text,
'bubble_rating': review.select_one('div.reviewItemInline span.ui_bubble_rating')['class'][1][7:],
}
#~ yield item
results.append(item)
for key,val in item.items():
print(key, ':', val)
print('----')
#return # for test only - to stop after first review
start_urls = [
'https://www.tripadvisor.in/Hotel_Review-g60763-d93339-Reviews-Hilton_New_York_Grand_Central-New_York_City_New_York.html',
#'https://www.tripadvisor.com/Hotel_Review-g60795-d102542-Reviews-Courtyard_Philadelphia_Airport-Philadelphia_Pennsylvania.html',
#'https://www.tripadvisor.com/Hotel_Review-g60795-d122332-Reviews-The_Ritz_Carlton_Philadelphia-Philadelphia_Pennsylvania.html', ]
results = []
for url in start_urls:
parse(url, get_soup(url))
import pandas as pd
df = pd.DataFrame(results) # <--- convert list to DataFrame df.to_csv('output.csv')
I am getting an output sample in csv file from review like:
I went on a family trip and it was amazing, I hope to come back soon. The room was small but what can you expect from New York. It was close to many things and the staff was perfect.I will come back again soon.More...
I just want to expand that more. I need a help..I really have no clue to do it.Please help.
I have written one more code but unable to pull the id from next page.Code is given below
import re
import urllib
#import webbrowser``
s = requests.Session()
headers = {'User-Agent': 'Mozilla/5.0 (X11; Ubuntu; Linux x86_64; rv:57.0) Gecko/20100101 Firefox/57.0'}
for i in range(0,10,5):
url = ("https://www.tripadvisor.in/Hotel_Review-g60763-d93339-Reviews-or{}-Hilton_New_York_Grand_Central-New_York_City_New_York.html").format(i)
print(url)
r = s.get(url,headers=headers)
html = BeautifulSoup(r.text, 'html.parser')
pattern = re.compile(r"UID_(\w+)\-SRC_(\w+)")
id = soup.find("div", id=pattern)["id"]
uid = pattern.match(id).group(2)
print(uid)
url1 ="https://www.tripadvisor.in/ShowUserReviews-g60763-d93339-r"+str(uid)+"-Hilton_New_York_Grand_Central-New_York_City_New_York.html#CHECK_RATES_CONT"
print(url1)
url2 = ('"' + url1 + '"')`enter code here`
print(url2)
The site uses ajax to expand the review content. The full content is not downloaded until the More link is clicked.
One way to access the content would be to figure out the ajax request format and then issue a HTTP request for the same. That might be difficult, perhaps not.
Another, easier, way is by noticing that the review title is a clickable link which loads the full review in a new page. You can therefore scrape the URL for each review and send a similar GET request. Then scrape the data from the response.

Python - save requests or BeautifulSoup object locally

I have some code that is quite long, so it takes a long time to run. I want to simply save either the requests object (in this case "name") or the BeautifulSoup object (in this case "soup") locally so that next time I can save time. Here is the code:
from bs4 import BeautifulSoup
import requests
url = 'SOMEURL'
name = requests.get(url)
soup = BeautifulSoup(name.content)
Since name.content is just HTML, you can just dump this to a file and read it back later.
Usually the bottleneck is not the parsing, but instead the network latency of making requests.
from bs4 import BeautifulSoup
import requests
url = 'https://google.com'
name = requests.get(url)
with open("/tmp/A.html", "w") as f:
f.write(name.content)
# read it back in
with open("/tmp/A.html") as f:
soup = BeautifulSoup(f)
# do something with soup
Here is some anecdotal evidence for the fact that bottleneck is in the network.
from bs4 import BeautifulSoup
import requests
import time
url = 'https://google.com'
t1 = time.clock();
name = requests.get(url)
t2 = time.clock();
soup = BeautifulSoup(name.content)
t3 = time.clock();
print t2 - t1, t3 - t2
Output, from running on Thinkpad X1 Carbon, with a fast campus network.
0.11 0.02
Storing requests locally and restoring them as Beautifoul Soup object latter on
If you are iterating through pages of web site you can store each page with request explained here.
Create folder soupCategory in same folder where your script is.
Use any latest user agent for headers
headers = {'user-agent':'Mozilla/5.0 (Macintosh; Intel Mac OS X 10_13_6) AppleWebKit/605.1.15 (KHTML, like Gecko) Version/13.0 Safari/605.1.15'}
def getCategorySoup():
session = requests.Session()
retry = Retry(connect=7, backoff_factor=0.5)
adapter = HTTPAdapter(max_retries=retry)
session.mount('http://', adapter)
session.mount('https://', adapter)
basic_url = "https://www.somescrappingdomain.com/apartments?adsWithImages=1&page="
t0 = time.time()
j=0
totalPages = 1525 # put your number of pages here
for i in range(1,totalPages):
url = basic_url+str(i)
r = requests.get(url, headers=headers)
pageName = "./soupCategory/"+str(i)+".html"
with open(pageName, mode='w', encoding='UTF-8', errors='strict', buffering=1) as f:
f.write(r.text)
print (pageName, end=" ")
t1 = time.time()
total = t1-t0
print ("Total time for getting ",totalPages," category pages is ", round(total), " seconds")
return
Latter on you can create Beautifoul Soup object as #merlin2011 mentioned with:
with open("/soupCategory/1.html") as f:
soup = BeautifulSoup(f)

Categories