I am trying to extract a table using html_table and the rvest package
library(rvest)
test <- html("http://www.privacyrights.org/data-breach/new?title=")
test %>% html_table(html_nodes("table.data-breach-table")[[1]])
however, I keep getting an error
Error in UseMethod("html_nodes"): no applicable method for
'html_nodes' applied to an object of class "character"
If you are going to nest parenthesized calls anyway, why bother with piping?
html_table(html_nodes(test, "table.data-breach-table")[[1]])
Otherwise go full pipe and use magrittr as well:
library(magrittr)
test %>%
html_nodes("table.data-breach-table") %>%
extract2(1) %>%
html_table()
NOTE:
the URL you are using does not have the table you want anyway
you should be using the newest rvest and using read_html
As far as why it wasn't working, you were piping test incorrectly and html_nodes was operating on the table… string instead of the parsed HTML document it expects.
Since you're trying to scrape breaches, this may be of help:
library(rvest)
library(dplyr)
library(pbapply)
urls <- sprintf("http://www.privacyrights.org/data-breach?title=&page=%d", 1:94)
pblapply(urls, function(URL) {
pg <- read_html(URL)
tab <- html_nodes(pg, "table")[3]
rows <- html_nodes(tab, "tr:not(.data-breach-bottom)")
bind_rows(lapply(seq(2, length(rows)-2, by=2), function(i) {
tds_1 <- html_nodes(rows[i], "td")
tds_2 <- html_text(html_nodes(rows[i+1], "td"), trim=TRUE)
data_frame(date_public=html_text(tds_1[1], TRUE),
name_loc=html_text(tds_1[2], TRUE),
entity=html_text(tds_1[3], TRUE),
type=html_text(tds_1[4], TRUE),
recs=html_text(tds_1[5], TRUE),
descr=tds_2[1])
}))
}) -> things
It's from an older gitst of mine. You'll need to add a randomized sleep delay to that if you do plan on scraping all their breaches.
NOTE also that it's skewed data and be very aware of it's limitations as you attempt to use it (I do data breach research for a living).
Related
I'm trying to extract a bit of information under the node /html/head/script[16] from a website (here) but am unable to do so.
nykaa <- "https://www.nykaa.com/biotique-bio-kelp-protein-shampoo-for-falling-hair-intensive-hair-growth-treatment-conf/p/357142?categoryId=1292&productId=357142&ptype=product&skuId=39934"
obj <- read_html(nykaa)
extracted_json <- obj %>%
html_nodes(xpath = "/html/head/script[16]") %>%
html_text(trim = TRUE)
Currently, my output for the above code is null. But I would like to extract the data under the above mentioned node in an organized manner.
You can use regex to grab the javascript object inside that script tag and then pass to jsonlite and parse. You need to root around a bit to get what you want from that but it is all there
library(rvest)
library(magrittr)
library(stringr)
library(jsonlite)
p <- read_html('https://www.nykaa.com/biotique-bio-kelp-protein-shampoo-for-falling-hair-intensive-hair-growth-treatment-conf/p/357142?categoryId=1292&productId=357142&ptype=product&skuId=39934') %>% html_text()
all_data <- jsonlite::parse_json(str_match_all(p,'window\\.__PRELOADED_STATE__ = (.*)')[[1]][,2])
For this website: https://www.coinopsy.com/dead-coins/, I'm using R and the rvest package to scrape names, summary, etc., that kind of info, to make my own form. I've done this with other websites and it was really successful, but this one is odd.
I used SelectorGadget, which is useful, in my previous jobs, to figure out the css nodes' names, but html_nodes and html_text return empty character, I don't know if it's because the website is structured under a totally different format!
An example of the css code:
td class="all sorting_1">a class="coin_name" href="007coin">007Coin /a>/td>
a class="coin_name" href="007coin">007Coin /a>
url <- "https://www.coinopsy.com/dead-coins/"
webpage <- read_html(url)
Item_html <- html_nodes(webpage,'.coin_name')
Item <- html_text(Item_html)
> Item
character(0)
Can someone help me out on this issue?
If you disable javascript in the browser you will see that that content is not loaded. If you then inspect the html you will see the data is stored in a script tag; presumably loaded into the table when javascript runs in the browser. Javascript doesn't run with the method you are using. You can extract the javascript array of arrays from the response html. Then parse into a dataframe. I am new to R so looking into how this can be done in this case. I will include a full example with python at the end. I will update if my research yields something. Otherwise, you can regex out contents from returned string in data.
library(rvest)
library(stringr)
library(magrittr)
url = 'https://www.coinopsy.com/dead-coins/'
r <- read_html(url) %>%
html_node('body') %>%
html_text() %>%
toString()
data <- str_match_all(r,'var table_data = (.*?);')
data <- data[[1]][,2] # string representation of list of lists
#step to convert string to object
#step to convert object to dataframe
In python there is the ast library which makes the conversion easy and the result of the below is the table you see on the page.
import requests
import re
import ast
import pandas as pd
r = requests.get('https://www.coinopsy.com/dead-coins/')
p = re.compile(r'var table_data = (.*?);') #p1 = re.compile(r'(\[".*?"\])')
data = p.findall(r.text)[0]
listings = ast.literal_eval(data)
df = pd.DataFrame(listings)
print(df)
Edit:
Currently I can't find a library which does the conversion I mentioned. Below is ugly way of combining and feels inefficient. I would welcome suggestions on improvements (though that may be for code review later). I'm still looking at this so will update.
library(rvest)
library(stringr)
library(magrittr)
url = 'https://www.coinopsy.com/dead-coins/'
headers <- c("Column To Drop","Name","Summary","Project Start Date","Project End Date","Founder","urlId")
# https://www.coinopsy.com/dead-coins/bigone-token/ where bigone-token is urlId
r <- read_html(url) %>%
html_node('body') %>%
html_text() %>%
toString()
data <- str_match_all(r,'var table_data = (.*?);')
data <- data[[1]][,2]
z <- substr(data, start = 2, stop = nchar(data)-1) %>% str_match_all(., "\\[(.*?)\\]")
z <- z[[1]][,2]
for(i in seq(1,length(z))){
if(i==1){
df <- rapply(as.list(strsplit(z[i], ",")[[1]][2:7]), function(x) trimws(sub("'(.*?)'", "\\1", x)))
}else{
df <- rbind(df,rapply(as.list(strsplit(z[i], ",")[[1]][2:7]), function(x) trimws(sub("'(.*?)'", "\\1", x))))
}
}
maybe it will help someone, I had the same problem, the solution was that at the beginning I have to specify the label to which the script is to be directed followed by the ".". In your case you want to address a class named coin_name, when specifying that class in the html_nodes function you don't specify the tag, same as I did. To solve it, I only had to specify the label, which in your case is the "a" label, and it would look like this.
Item_html <- html_nodes(webpage,'a.coin_name')
That way the html_nodes function would not return empty.
I know you already solved it but I hope someone can help you.
I'm trying to scrape tabulated data on previous US statewide election results, and I think ballotpedia.org is a good place to be getting this data from - as URLs are in a consistent format for all states.
Here's the code I set up to test it:
library(dplyr)
library(rvest)
# STEP 1 - URL COMPONENTS TO SCRAPE FROM
senate_base_url <- "https://ballotpedia.org/United_States_Senate_elections_in_"
senate_state_urls <- gsub(" ", "_", state.name)
senate_year_urls <- c(",_2012", ",_2014", ",_2016")
# TEST
test_url <- paste0(senate_base_url, senate_state_urls[10], senate_year_urls[2])
this results in the following URL: https://ballotpedia.org/United_States_Senate_elections_in_Georgia,_2014
Using the 'selectorgadget' chrome plugin, I selected the table in question containing the election result, and tried parsing it into R as follows:
test_data <- read_html(test_url)
test_data <- test_data %>%
html_node(xpath = '//*[#id="collapsibleTable0"]') %>%
html_table()
However, I'm getting the following error:
Error in UseMethod("html_table") :
no applicable method for 'html_table' applied to an object of class "xml_missing"
Furthermore, the R object test_data yields a list with 2 empty elements.
Can anyone tell me what I'm doing wrong here? Is the html_table() function the wrong one? Using html_text() simply returns an NA character vector. Any help would be greatly appreciated, thanks very much :).
Your xpath statement is incorrect, thus the html_node function is returning a null value.
Here is a solution using the html tags. "Look for a table tag within a center tag"
library(rvest)
test_data <- read_html(test_url)
test_data <- test_data %>% html_nodes("center table") %>% html_table()
Or to retrieve the fully collapsed table use the html tag with class name:
collapsedtable<-test_data %>% html_nodes("table.collapsible") %>%
html_table(fill=TRUE)
this works for me:
library(httr)
library(XML)
r <- httr::GET("https://ballotpedia.org/United_States_Senate_elections_in_Georgia,_2014")
XML::readHTMLTable(rawToChar(r$content))[[2]]
I imported the csv file that I want to use in r. Here, I am trying to call one of the columns from the csv file. This column has a list of urls titled "URLs". Then, I want the code which I have to scrap data from each url. In short, I want to use more efficient way than listing all the urls in c() function since I have about 200 links.
https://www.nytimes.com/2018/04/07/health/health-care-mergers-doctors.html?rref=collection%2Fsectioncollection%2Fhealth
https://www.nytimes.com/2018/04/11/well/move/why-exercise-alone-may-not-be-the-key-to-weight-loss.html?rref=collection%2Fsectioncollection%2Fhealth
https://www.nytimes.com/2018/04/07/health/antidepressants-withdrawal-prozac-cymbalta.html?rref=collection%2Fsectioncollection%2Fhealth
https://www.nytimes.com/2018/04/09/well/why-you-should-get-the-new-shingles-vaccine.html?rref=collection%2Fsectioncollection%2Fhealth
https://www.nytimes.com/2018/04/09/health/fda-essure-bayer-contraceptive-implant.html?rref=collection%2Fsectioncollection%2Fhealth
https://www.nytimes.com/2018/04/09/health/hot-pepper-thunderclap-headaches.html?rref=collection%2Fsectioncollection%2Fhealth
The error appears when running this: article <- links %>% map(read_html).
It gives me this message:
(Error in UseMethod("read_xml") :
no applicable method for 'read_xml' applied to an object of class "factor")
Here is the code:
setwd("C:/Users/Majed/Desktop")
d <- read.csv("NYT.csv")
d
links<- d$URLs
article <- links %>% map(read_html)
title <-
article %>% map_chr(. %>% html_node("title") %>% html_text())
content <-
article %>% map_chr(. %>% html_nodes(".story-body-text") %>% html_text() %>% paste(., collapse = ""))
article_table <- data.frame("Title" = title, "Content" = content)
Pay attention to the meaning of your error message: read_html expects a character string, but you're giving it a factor. read.csv converts strings to factors, unless you include the argument stringsAsFactors = F. read_csv from readr is a good alternative if you, like me, forget that you don't want strings automatically turned into factors.
I can't reproduce the problem without your data, but try converting the URLs to strings:
links <- as.character(d$URLs)
article <- links %>% map(read_html)
I am web scraping a page at
http://catalog.ihsn.org/index.php/catalog#_r=&collection=&country=&dtype=&from=1890&page=1&ps=100&sid=&sk=&sort_by=nation&sort_order=&to=2017&topic=&view=s&vk=
From this url, I have built up a dataframe through the following code:
dflist <- map(.x = 1:417, .f = function(x) {
Sys.sleep(5)
url <- ("http://catalog.ihsn.org/index.php/catalog#_r=&collection=&country=&dtype=&from=1890&page=1&ps=100&sid=&sk=&sort_by=nation&sort_order=&to=2017&topic=&view=s&vk=")
read_html(url) %>%
html_nodes(".title a") %>%
html_text() %>%
as.data.frame()
}) %>% do.call(rbind, .)
I have repeated the same code in order to get all the data I was interested in and it seems to work perfectly, although is of course a little slow due to the Sys.sleep() thing.
My issue has raised once I have tried to scrape the single projects descriptions that should be included in the dataframe.
For instance, the first project description is at
http://catalog.ihsn.org/index.php/catalog/7118/study-description
the second project description is at
http://catalog.ihsn.org/index.php/catalog/6606/study-description
and so forth.
My problem is that I can't find a dynamic way to scrape all the projects' pages and insert them in the data frame, being the number in the URLs not progressive nor at the end of the link.
To make things clearer, this is the structure of the website I am scraping:
1.http://catalog.ihsn.org/index.php/catalog#_r=&collection=&country=&dtype=&from=1890&page=1&ps=100&sid=&sk=&sort_by=nation&sort_order=&to=2017&topic=&view=s&vk=
1.1. http://catalog.ihsn.org/index.php/catalog/7118
1.1.a http://catalog.ihsn.org/index.php/catalog/7118/related_materials
1.1.b http://catalog.ihsn.org/index.php/catalog/7118/study-description
1.1.c. http://catalog.ihsn.org/index.php/catalog/7118/data_dictionary
I have scraped successfully level 1. but cannot level 1.1.b. (study-description) , the one I am interested in, since the dynamic element of the URL (in this case: 7118) is not consistent in the website's above 6000 pages of that level.
You have to extract the deeper urls from the .title a and then scrape those as well. Here's a small example on how to do that using rvest and the tidyverse
library(tidyverse)
library(rvest)
scraper <- function(x) {
Sys.sleep(5)
url <- sprintf("http://catalog.ihsn.org/index.php/catalog#_r=&collection=&country=&dtype=&from=1890&page=%s&ps=100&sid=&sk=&sort_by=nation&sort_order=&to=2017&topic=&view=s&vk=", x)
html <- read_html(url)
tibble(title = html_nodes(html, ".title a") %>% html_text(trim = TRUE),
project_url = html_nodes(html, ".title a") %>% html_attr("href"))
}
result <- map_df(1:2, scraper) %>%
mutate(study_description = map(project_url, ~read_html(sprintf("%s/study-description", .x)) %>% html_node(".xsl-block") %>% html_text()))
This isn't complete as to all the things you want to do, but should show you an approach.