ferretvsrvest
Ferret is a web scraping system. It aims to simplify data extraction from the web for UI testing, machine learning, analytics and more. ferret allows users to focus on the data. It abstracts away the technical details and complexity of underlying technologies using its own declarative language. It is extremely portable, extensible, and fast.
Features
- Declarative language
- Support of both static and dynamic web pages
- Embeddable
- Extensible
Ferret is always implemented in Python through pyfer
rvest is a popular R library for web scraping and parsing HTML and XML documents. It is built on top of the xml2 and httr libraries and provides a simple and consistent API for interacting with web pages.
One of the main advantages of using rvest is its simplicity and ease of use. It provides a number of functions that make it easy to extract information from web pages, even for those who are not familiar with web scraping. The html_nodes and html_node functions allow you to select elements from an HTML document using CSS selectors, similar to how you would select elements in JavaScript.
rvest also provides functions for interacting with forms, including html_form, set_values, and submit_form functions. These functions make it easy to navigate through forms and submit data to the server, which can be useful when scraping sites that require authentication or when interacting with dynamic web pages.
rvest also provides functions for parsing XML documents. It includes xml_nodes and xml_node functions, which also use CSS selectors to select elements from an XML document, as well as xml_attrs and xml_attr functions to extract attributes from elements.
Another advantage of rvest is that it provides a way to handle cookies, so you can keep the session alive while scraping a website, and also you can handle redirections with handle_redirects
Example Use
// Example scraper for Google in Ferret:
LET google = DOCUMENT("https://www.google.com/", {
driver: "cdp",
userAgent: "Mozilla/5.0 (Macintosh; Intel Mac OS X 10_14_6) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/76.0.3809.87 Safari/537.36"
})
HOVER(google, 'input[name="q"]')
WAIT(RAND(100))
INPUT(google, 'input[name="q"]', @criteria, 30)
WAIT(RAND(100))
CLICK(google, 'input[name="btnK"]')
WAITFOR EVENT "navigation" IN google
WAIT_ELEMENT(google, "#res")
LET results = ELEMENTS(google, X("//*[text() = 'Search Results']/following-sibling::*/*"))
FOR el IN results
RETURN {
title: INNER_TEXT(el, 'h3')?,
description: INNER_TEXT(el, X("//em/parent::*")),
url: ELEMENT(el, 'a')?.attributes.href
}
library("rvest")
# Rvest can use basic HTTP client to download remote HTML:
tree <- read_html("http://webscraping.fyi/lib/r/rvest")
# or read from string:
tree <- read_html('
<div class="products">
<a href="/product/1">Cat Food</a>
<a href="/product/2">Dog Food</a>
</div>
')
# to parse HTML trees with rvest we use r pipes (the %>% symbol) and html_element function:
# we can use css selectors:
print(tree %>% html_element(".products>a") %>% html_text())
# "[1] "\nCat Food\nDog Food\n""
# or XPath:
print(tree %>% html_element(xpath="//div[@class='products']/a") %>% html_text())
# "[1] "\nCat Food\nDog Food\n""
# Additionally rvest offers many quality of life functions:
# html_text2 - removes trailing and leading spaces and joins values
print(tree %>% html_element("div") %>% html_text2())
# "[1] "Cat Food Dog Food""
# html_attr - selects element's attribute:
print(tree %>% html_element("div") %>% html_attr('class'))
# "products"