如何使用R语言为本地电视门户网站指定评论点赞？

Great question! Let's break this down step by step—since rvest is perfect for scraping static content, but clicking that green like arrow triggers a dynamic backend request (not just a static page update), we'll need two practical approaches to get this done: either reverse-engineer the site's like API, or simulate a real browser click with automation.

Approach 1: Reverse-Engineer the Like API (More Efficient) #

This method mimics the backend request your browser sends when you click "like", using httr to replicate it directly in R.

Step 1: Locate the Target Comment from User "eto" #

First, use rvest to find the comment and extract its unique ID (the site needs this to know which comment to like):

library(rvest)
library(stringr)

url1 <- "http://rs.n1info.com/Comment/StandardComment/637905/Sci-Tech/Facebook-ce-dati-prednost-verodostojnim-medijskim-kucama.html"
webpage <- read_html(url1, encoding = "UTF-8")

# Grab all comment blocks (adjust the CSS selector if the site uses a different class)
comments <- html_nodes(webpage, ".comment-item")

# Loop through comments to find the one by "eto"
target_comment <- NULL
for (comment in comments) {
  # Extract the username (tweak the selector to match the site's username element)
  username <- html_text(html_node(comment, ".comment-author-name")) %>% str_trim()
  if (username == "eto") {
    target_comment <- comment
    break
  }
}

# Extract the comment's unique ID (check the comment block's attributes—look for data-id, data-comment-id, etc.)
if (!is.null(target_comment)) {
  comment_id <- html_attr(target_comment, "data-id")
  cat("Found target comment ID:", comment_id, "\n")
} else {
  stop("Couldn't locate a comment from user 'eto'")
}

Step 2: Capture the Like Request with Browser Dev Tools #

1. Open the comment page in Chrome/Firefox
2. Right-click the green like arrow → Inspect to open Dev Tools
3. Switch to the Network tab, then click the like arrow
4. Look for an XHR/fetch request (usually named something like VoteUp or LikeComment)
5. Note these details:
   • The request URL (e.g., http://rs.n1info.com/Comment/VoteUp)
   • Request method (POST/GET)
   • Required parameters (e.g., commentId=637905 or a CSRF token)
   • Any cookies or headers sent with the request

Step 3: Replicate the Request in R with httr #

Use a session to maintain cookies (some sites require active session cookies to allow likes):

library(httr)

# Start a session to persist cookies from the initial page load
session <- session(url1)

# Build the request using the details from Dev Tools
vote_api_url <- "http://rs.n1info.com/Comment/VoteUp" # Replace with your actual API URL
request_body <- list(
  commentId = comment_id,
  # Add any other required parameters (e.g., csrfToken = "xxx")
)

# Send the POST request (use GET if the site uses that method instead)
response <- session_post(
  session,
  vote_api_url,
  body = request_body,
  encode = "form" # Use "json" if the site sends JSON data instead of form data
)

# Check if the like was successful
if (http_status(response)$category == "Success") {
  cat("Successfully liked the comment from 'eto'!\n")
} else {
  cat("Like failed. Response message:", content(response, "text"), "\n")
}

Approach 2: Simulate Browser Clicks with RSelenium (Beginner-Friendly) #

If reverse-engineering the API feels overwhelming, RSelenium lets you control a real browser to mimic your manual click—no need to dig into backend requests.

Step 1: Set Up RSelenium #

First, install the package and start a browser session (we'll use Chrome here):

install.packages("RSelenium")
library(RSelenium)

# Start a Chrome browser session (adjust the port if needed)
driver <- rsDriver(browser = "chrome", port = 4567L)
remDr <- driver[["client"]]

Step 2: Navigate to the Page and Click the Like Button #

# Open the comment page in the browser
remDr$navigate(url1)

# Find all username elements and locate the one for "eto"
username_elements <- remDr$findElements(using = "css selector", value = ".comment-author-name")

for (elem in username_elements) {
  if (elem$getElementText()[[1]] == "eto") {
    # Find the parent comment block, then the green like icon
    parent_comment <- elem$findElement(using = "xpath", value = "./ancestor::div[contains(@class, 'comment-item')]")
    like_button <- parent_comment$findElement(using = "css selector", value = ".comment-likes-icon")
    
    # Click the like button
    like_button$clickElement()
    cat("Clicked the like button for 'eto'!\n")
    break
  }
}

# Clean up: close the browser and stop the driver
remDr$close()
driver$server$stop()

Important Tips #

• Anti-Scraping Checks: Many sites block automated requests. If you hit errors, check if you need to log in first (RSelenium makes this easy—just enter your credentials manually in the browser window).
• Selector Updates: CSS/XPath selectors (like .comment-item or .comment-author-name) depend on the site's HTML structure. Use your browser's Inspect tool to verify them if they stop working.
• Rate Limiting: Don't send too many requests too quickly—you might get temporarily blocked from the site.

内容的提问来源于stack exchange，提问作者Newbie_R