Rselenium scraping table

I want to extract the data in the "Completed Games" table located here "https://www.chess.com/member/magnuscarlsen".

The code below gives me a list of size 0. The Selenium side of things seems to be working. A firefox browser opens on my desktop and navigates to the page. Any help would be greatly appreciated. I'm at my wits end!

rD <- rsDriver(browser="firefox", port=4442L, verbose=F)
remDr <- rD[["client"]]

remDr$navigate("https://www.chess.com/member/magnuscarlsen")

Sys.sleep(5) # give the page time to fully load
html <- remDr$getPageSource()[[1]]

html <- read_html(html)

signal <- html %>%
  html_nodes("table.table-component table-hover archived-games-table")

Solution

1 If you don't mind not having the accuracy figures (for which I believe there is no published basis for calculation) have a look at the public APIs from Chess.com. You do get all the moves info included.

In particular, the implementations via BigChess package. I amended examples from there below:

All games:

library(rjson)
library(bigchess)

user <- "magnuscarlsen"
json_file <- paste0("https://api.chess.com/pub/player/", user,"/games/archives")
json_data <- fromJSON(paste(readLines(json_file), collapse = ""))
result <- data.frame()
for(i in json_data$archives)
  result <- rbind(result, read.pgn(paste0(i, "/pgn")))

Single month:

library(bigchess)

df <- read.pgn("https://api.chess.com/pub/player/magnuscarlsen/games/2020/12/pgn")
print(df[df$Date == '2020.12.11'])

Adding in your accuracies as requested. Most of the info on that page is actually available via the APIs:


library(bigchess)
#> Warning: package 'bigchess' was built under R version 4.0.3
library(purrr)
library(jsonlite)
#> Warning: package 'jsonlite' was built under R version 4.0.3
#> 
#> Attaching package: 'jsonlite'
#> The following object is masked from 'package:purrr':
#> 
#>     flatten
library(stringr)

try_again <- function(link) {  #https://blog.r-hub.io/2020/04/07/retry-wheel/
  maxtry <- 5
  try <- 1
  resp <- read_json(link)
  while (try <= maxtry && is.null(resp$data)) {
    resp <- read_json(.)
    try <- try + 1
    Sys.sleep(try * .25)
  }
  return(resp)
}

url <- "https://api.chess.com/pub/player/magnuscarlsen/games/2020/12"
result <- data.frame()
result <- read.pgn(paste0(url, "/pgn"))
#> Warning in readLines(con): incomplete final line found on 'https://
#> api.chess.com/pub/player/magnuscarlsen/games/2020/12/pgn'
#> 2021-02-15 20:29:04, successfully imported 47 games
#> 2021-02-15 20:29:04, N moves computed
#> 2021-02-15 20:29:04, extract moves done
#> 2021-02-15 20:29:04, stat moves computed
result <- filter(result, result$Date == "2020.12.11")

data <- read_json(url)
mask <- map(data$games, ~ !is.na(str_match(.x$pgn, 'UTCDate\\s\\"2020\\.12\\.11')[, 1])) %>% unlist()
games <- data$games[mask]
games <- paste0("https://www.chess.com/callback/analysis/game/live/", map(games, ~ str_match(.x$url, "\\d+")[, 1]), "/all")

df <- map_df(games, ~ {
  json_data <- try_again(.x)
  tryCatch(
    data.frame(
      Url = .x,
      WhiteAccuracy = json_data$data$analysis$CAPS$white$all,
      BlackAccuracy = json_data$data$analysis$CAPS$black$all,
      stringsAsFactors = FALSE
    ),
    error = function(e) {
      data.frame(
        Url = .x,
        WhiteAccuracy = NA_integer_,
        BlackAccuracy = NA_integer_,
        stringsAsFactors = FALSE
      )
    }
  )
})

final <- cbind(result, df)
#> Error in .cbind.ts(list(...), .makeNamesTs(...), dframe = FALSE, union = TRUE): non-time series not of the correct length

^{Created on 2021-02-15 by the reprex package (v0.3.0)}