first commit

.Rbuildignore

@@ -0,0 +1,2 @@
+^.*\.Rproj$
+^\.Rproj\.user$

.gitignore

@@ -0,0 +1,4 @@
+.Rproj.user
+.Rhistory
+.RData
+.Ruserdata

DESCRIPTION

@@ -0,0 +1,31 @@
+Package: worldfootballR
+Type: Package
+Title: An R Package To Extract World Football (Soccer) data from fbref.com
+Version: 0.0.0.9000
+Authors@R: person("Jason Zivkovic", email = "jase.ziv83@gmail.com",
+                  role = c("aut", "cre"))
+URL: https://github.com/JaseZiv/worldfootballR
+BugReports: https://github.com/JaseZiv/worldfootballR/issues
+Description: This R package will enable users to extract a number of different game and player statistics and metrics from fbref.com.
+License: GPL-3
+Encoding: UTF-8
+LazyData: true
+Imports:
+    rlang,
+    dplyr,
+    purrr,
+    tidyr,
+    stringr,
+    lubridate,
+    magrittr,
+    scales,
+    xml2,
+    rvest,
+    curl,
+    rvest,
+    xml2
+RoxygenNote: 7.1.0
+Suggests: knitr,
+    rmarkdown,
+    ggplot2
+VignetteBuilder: knitr

NAMESPACE

@@ -0,0 +1,6 @@
+# Generated by roxygen2: do not edit by hand
+
+export(get_match_results)
+export(get_match_urls)
+importFrom(magrittr,"%>%")
+importFrom(rlang,.data)

R/get_match_results.R

@@ -0,0 +1,74 @@
+#' Get match results
+#'
+#' Returns the game results for a given league season
+#'
+#' @param country the three character country code
+#' @param gender gender of competition, either "M" or "F"
+#' @param season_end_year the year the season concludes, in quotes, ie "2021"
+#'
+#' @return returns a dataframe with the results of the competition, season and gender
+#'
+#' @importFrom magrittr %>%
+#' @importFrom rlang .data
+#'
+#' @export
+#'
+#' @examples
+#' \dontrun{
+#' get_match_results(country = "ITA", gender = "M", season_end_year = "2020")
+#' }
+
+
+get_match_results <- function(country, gender, season_end_year) {
+
+  print(paste0("Scraping ", country, " fist division for the ", season_end_year, " season (gender = ", gender, ")"))
+
+  competitions <- .get_tier1_competitions()
+
+  main_url <- "https://fbref.com"
+
+  country <- country
+  gender <- gender
+  season <- season_end_year
+  select_season <- .get_league_season_url(country, gender, season)
+
+
+  fixtures_url <- xml2::read_html(select_season) %>%
+    rvest::html_nodes(".hoversmooth") %>%
+    rvest::html_nodes(".full") %>%
+    rvest::html_nodes("a") %>%
+    rvest::html_attr("href") %>% .[grepl("Fixtures", .)] %>% paste0(main_url, .)
+
+
+  season_summary <- xml2::read_html(fixtures_url) %>%
+    rvest::html_table() %>% .[1] %>% data.frame() %>%
+    dplyr::filter(.data$Date != "")
+
+
+  suppressWarnings(
+    season_summary <- season_summary %>%
+      tidyr::separate(.data$Score, into = c("HomeGoals", "AwayGoals"), sep = "–") %>%
+      dplyr::mutate(HomeGoals = as.numeric(.data$HomeGoals),
+             AwayGoals = as.numeric(.data$AwayGoals),
+             Attendance = as.numeric(gsub(",", "", .data$Attendance)))
+  )
+
+  season_summary <- cbind(season_end_year, season_summary)
+
+  if(!any(stringr::str_detect(names(season_summary), "Round"))) {
+    Round <- rep(NA, nrow(season_summary))
+    season_summary <- cbind(Round, season_summary)
+  }
+
+  if(any(stringr::str_detect(names(season_summary), "xG"))) {
+    season_summary <- season_summary %>%
+      dplyr::select(Season=season_end_year, Round, .data$Wk, .data$Day, .data$Date, .data$Time, .data$Home, .data$HomeGoals, Home_xG=.data$xG, .data$Away, .data$AwayGoals, Away_xG=.data$xG.1, .data$Attendance, .data$Venue, .data$Referee, .data$Notes)
+  } else {
+    season_summary <- season_summary %>%
+      dplyr::select(Season=season_end_year, Round, .data$Wk, .data$Day, .data$Date, .data$Time, .data$Home, .data$HomeGoals, .data$Away, .data$AwayGoals, .data$Attendance, .data$Venue, .data$Referee, .data$Notes)
+  }
+
+
+  return(season_summary)
+
+}

R/get_match_urls.R

@@ -0,0 +1,44 @@
+#' Get match URLs
+#'
+#' Returns the URL for each match played for a given league season
+#'
+#' @param country the three character country code
+#' @param gender gender of competition, either "M" or "F"
+#' @param season the year the season concludes, in quotes, ie "2021"
+#'
+#' @return returns a character vector of all fbref match URLs for a given competition, season and gender
+#'
+#' @importFrom magrittr %>%
+#' @importFrom rlang .data
+#'
+#' @export
+#'
+#' @examples
+#' \dontrun{
+#' get_match_urls(country = "ENG", gender = "M", season_end_year = "2020")
+#' }
+
+get_match_urls <- function(country, gender, season) {
+  main_url <- "https://fbref.com"
+
+  country <- toupper(country)
+  gender <- toupper(gender)
+  season <- season
+
+  selected_season <- .get_league_season_url(country, gender, season)
+
+  fixtures_url <- xml2::read_html(selected_season) %>%
+    rvest::html_nodes(".hoversmooth") %>%
+    rvest::html_nodes(".full") %>%
+    rvest::html_nodes("a") %>%
+    rvest::html_attr("href") %>% .[grepl("Fixtures", .)] %>% paste0(main_url, .)
+
+  match_report_urls <- xml2::read_html(fixtures_url) %>%
+    # html_nodes(".left~ .left+ .left a") %>%
+    rvest::html_nodes("td.left~ .left+ .left a") %>%
+    rvest::html_attr("href") %>%
+    paste0(main_url, .) %>% unique()
+
+  return(match_report_urls)
+
+}

R/globals.R

@@ -0,0 +1,2 @@
+
+utils::globalVariables(c("."))

R/internals.R

@@ -0,0 +1,73 @@
+#' Get Tier 1 competitions
+#'
+#' Returns a df of the top leagues around the world
+#'
+#' @return returns a dataframe with the tier 1 competitions from around the world
+#'
+#' @importFrom magrittr %>%
+#' @importFrom rlang .data
+#'
+
+.get_tier1_competitions <- function() {
+  main_url <- "https://fbref.com"
+  # read page to all competitions
+  all_comps_url <- xml2::read_html("https://fbref.com/en/comps/")
+  # this just gets the Tier 1 club comps - this will need to be modified if more comps are required
+  comps <- all_comps_url %>% rvest::html_nodes("#all_comps_1_fa_club_league_senior")
+  # get the urls for each competition, then paste fbref url
+  competition_urls <- comps %>% rvest::html_node("tbody") %>% rvest::html_nodes("th a") %>% rvest::html_attr("href")
+  competition_urls <- paste0(main_url, competition_urls)
+  # scrape the table that contains the competitons
+  competitions <- comps %>% rvest::html_nodes(".sortable") %>% rvest::html_table() %>% data.frame()
+  # add the competition url column
+  competitions <- cbind(competitions, competition_urls)
+  # remove the two character country code for the flag, and only leave the 3 character code
+  competitions$Country <- gsub(".*? ", "", competitions$Country)
+
+  return(competitions)
+}
+
+
+
+#' Get URL of league season
+#'
+#' Returns a URL for the selected league seasons
+#'
+#' @param country the three character country code
+#' @param gender gender of competition, either "M" or "F"
+#' @param season the year the season concludes, in quotes, ie "2021"
+
+#' @return a URL for the selected league seasons
+#'
+#' @importFrom magrittr %>%
+#' @importFrom rlang .data
+#'
+.get_league_season_url <- function(country, gender, season) {
+  main_url <- "https://fbref.com"
+
+  competitions <- .get_tier1_competitions()
+
+  league_url <- competitions %>%
+    dplyr::filter(toupper(.data$Country) == toupper(country),
+           toupper(.data$Gender) == toupper(gender)) %>%
+    dplyr::pull(.data$competition_urls)
+
+  league_url <- xml2::read_html(league_url)
+
+  seasons <- league_url %>%
+    rvest::html_nodes("th a") %>%
+    rvest::html_text() %>%
+    gsub(".*-", "", .)
+
+  seasons_urls <- league_url %>%
+    rvest::html_nodes("th a") %>%
+    rvest::html_attr("href") %>%
+    paste0(main_url, .)
+
+  select_season <- cbind(seasons, seasons_urls) %>% data.frame() %>%
+    dplyr::filter(seasons == season) %>%
+    dplyr::pull(seasons_urls)
+
+  return(select_season)
+
+}

worldfootballR.Rproj

@@ -0,0 +1,20 @@
+Version: 1.0
+
+RestoreWorkspace: Default
+SaveWorkspace: Default
+AlwaysSaveHistory: Default
+
+EnableCodeIndexing: Yes
+UseSpacesForTab: Yes
+NumSpacesForTab: 2
+Encoding: UTF-8
+
+RnwWeave: Sweave
+LaTeX: pdfLaTeX
+
+AutoAppendNewline: Yes
+StripTrailingWhitespace: Yes
+
+BuildType: Package
+PackageUseDevtools: Yes
+PackageInstallArgs: --no-multiarch --with-keep.source