For most of the project I use base R and the stringr library. The .txt file to parse over is hosted on GitHub and accessed using the readLines function.
library(stringr)
library(RCurl)## Loading required package: bitopschess_file <- readLines('https://raw.githubusercontent.com/Logan213/DATA607_Project1/master/tournamentinfo.txt', warn = FALSE)Looking over the file I thought it would be good to break up the file into the rows that contain the names, opponents, and total points, and the rows that contain the other required information (state, pre-rating).
For both sections, I used a simple “for loop” to parse over the file and store information from the “nth” row into new objects for the name and state rows, called chess_names and state_rows, respectively.
After looping through the text rows, the resulting characters are split up with strsplit and a simple regular expression as the information we want is separated by a bar (|).
# name rows
chess_names <- c()
for (i in 0:length(chess_file))
{
if (i %% 3 == 2)
{
chess_names <- append(chess_names, chess_file[i], after = length(chess_names))
}
}
chess_names <- unlist(strsplit(chess_names, "\\|"))
head(chess_names)## [1] " Pair " " Player Name "
## [3] "Total" "Round"
## [5] "Round" "Round"# state rows
state_rows <- c()
for (i in 0:length(chess_file))
{
if (i %% 3 == 0)
{
state_rows <- append(state_rows, chess_file[i], after = length(state_rows))
}
}
state_rows <- unlist(strsplit(state_rows, "\\|"))
head(state_rows[12:20])## [1] " ON " " 15445895 / R: 1794 ->1817 "
## [3] "N:2 " "W "
## [5] "B " "W "Since all we need are a few pieces of information and the .txt file has some sort of stucture, we can use R’s seq function to select characters from the chess_names and state_rows objects without using a loop.
Player_Name <- str_trim(chess_names[seq(13, 650, 10)])
State <- str_trim(state_rows[seq(12, length(state_rows), 10)])
Total_Points <- as.numeric(str_trim(chess_names[seq(14, length(chess_names), 10)]))The script above will store the information in a vector for each of the three basic pieces that we need (Player Name, State, and Total Points). Here is a preview of the Player_Name object:
head(Player_Name)## [1] "GARY HUA" "DAKSHESH DARURI" "ADITYA BAJAJ"
## [4] "PATRICK H SCHILLING" "HANSHI ZUO" "HANSEN SONG"Things get a little trickier now, as the Pre-Tournament rating is nested with some other information, and in some cases the score has some additional notation. To get what we want, we’ll just use the same technique as above, but going over the text twice, using a combination of selecting data with seq, regular expressions and the str_extract function from the stringr package.
Pre_Rating <- str_trim(state_rows[seq(13, length(state_rows), 10)])
Pre_Rating <- unlist(strsplit(Pre_Rating, "\\/"))
Pre_Rating <- str_trim(Pre_Rating[seq(2, length(Pre_Rating), 2)])
Pre_Rating <- str_extract(Pre_Rating, "\\d{3,4}")
Pre_Rating[1:10]## [1] "1794" "1553" "1384" "1716" "1655" "1686" "1649" "1641" "1411" "1365"The opponent scores are not actually given directly in the .txt file. Each opponent number is given in the round, but the pre-tournament rating has to be looked up underneath the name of the opponent. To get the average scores, we’ll first gather the opponent numbers and create a vector from each round. Then, we can look up the opponent rating in the Pre_Rating object, which is conveniently in the same order as the player listing in the .txt file.
round_1 <- str_trim(chess_names[seq(15, length(chess_names), 10)])
round_1 <- as.numeric(str_extract(round_1, "\\d{1,2}"))
round_2 <- str_trim(chess_names[seq(16, length(chess_names), 10)])
round_2 <- as.numeric(str_extract(round_2, "\\d{1,2}"))
round_3 <- str_trim(chess_names[seq(17, length(chess_names), 10)])
round_3 <- as.numeric(str_extract(round_3, "\\d{1,2}"))
round_4 <- str_trim(chess_names[seq(18, length(chess_names), 10)])
round_4 <- as.numeric(str_extract(round_4, "\\d{1,2}"))
round_5 <- str_trim(chess_names[seq(19, length(chess_names), 10)])
round_5 <- as.numeric(str_extract(round_5, "\\d{1,2}"))
round_6 <- str_trim(chess_names[seq(20, length(chess_names), 10)])
round_6 <- as.numeric(str_extract(round_6, "\\d{1,2}"))
round_7 <- str_trim(chess_names[seq(21, length(chess_names), 10)])
round_7 <- as.numeric(str_extract(round_7, "\\d{1,2}"))
#Preview of round_x vector
tail(round_7)## [1] 44 NA 37 NA NA 54I couldn’t get a simpler nested loop to work for this, so below is the process to match up each round’s opponent number to the corresponding Pre-Tournament Ranking for that player’s opponent. A loop parses over the corresponding round_x object, and then selects the Pre_Rating item. This works because the opponent number is the number we would use to slice or select the correct Pre-tournament rating.
After the information is taken from Pre_Rating, we then store it in a new object called rnd_x_score, and then convert it from character to numeric, since we will be calculating an average in the next step.
#round 1 opponents
rnd_1_score <- c()
for (i in 1:length(round_1))
{
rnd_1_score <- append(rnd_1_score, Pre_Rating[round_1[i]], after = length(rnd_1_score))
}
rnd_1_score <- as.numeric(rnd_1_score)
#round 2 opponents
rnd_2_score <- c()
for (i in 1:length(round_2))
{
rnd_2_score <- append(rnd_2_score, Pre_Rating[round_2[i]], after = length(rnd_2_score))
}
rnd_2_score <- as.numeric(rnd_2_score)
#round 3 opponents
rnd_3_score <- c()
for (i in 1:length(round_3))
{
rnd_3_score <- append(rnd_3_score, Pre_Rating[round_3[i]], after = length(rnd_3_score))
}
rnd_3_score <- as.numeric(rnd_3_score)
#round 4 opponents
rnd_4_score <- c()
for (i in 1:length(round_4))
{
rnd_4_score <- append(rnd_4_score, Pre_Rating[round_4[i]], after = length(rnd_4_score))
}
rnd_4_score <- as.numeric(rnd_4_score)
#round 5 opponents
rnd_5_score <- c()
for (i in 1:length(round_5))
{
rnd_5_score <- append(rnd_5_score, Pre_Rating[round_5[i]], after = length(rnd_5_score))
}
rnd_5_score <- as.numeric(rnd_5_score)
#round 6 opponents
rnd_6_score <- c()
for (i in 1:length(round_6))
{
rnd_6_score <- append(rnd_6_score, Pre_Rating[round_6[i]], after = length(rnd_6_score))
}
rnd_6_score <- as.numeric(rnd_6_score)
#round 7 opponents
rnd_7_score <- c()
for (i in 1:length(round_7))
{
rnd_7_score <- append(rnd_7_score, Pre_Rating[round_7[i]], after = length(rnd_7_score))
}
rnd_7_score <- as.numeric(rnd_7_score)
#Preview of rnd_7_score
head(rnd_7_score)## [1] 1716 1649 1663 1794 1629 1563In order to calculate the average, we combine the rnd_x_score objects into a matrix called round_scores, and then turn it into a data frame to easly calculate the average. R allows us to simultaneously create the new column and then enter the calculated means of each row.
Since we only need this one column for our final .csv file, and have already created vectors for the other information, we’ll separate it out from the data frame and store it in a new object.
round_scores <- cbind(rnd_1_score, rnd_2_score, rnd_3_score, rnd_4_score, rnd_5_score, rnd_6_score, rnd_7_score)
round_scores <- data.frame(round_scores)
round_scores$Average <- round(rowMeans(round_scores, na.rm = TRUE))
Avg_Opp_Pre_Rating <- round_scores$Average
head(Avg_Opp_Pre_Rating)## [1] 1605 1469 1564 1574 1501 1519To put the information we were looking for in a nice package, we use cbind and combine the columns for the Tournament Player Name, State, Total Points, Pre-Tournament Rating, and the Average of the Pre-Tournament Ratings for each player’s opponent into a matrix stored in chess_summary. We’ll then convert this into an R data frame, before using write.csv to store the data in a new .csv file.
#glue together the pieces
chess_summary <- cbind(Player_Name, State, Total_Points, Pre_Rating, Avg_Opp_Pre_Rating)
chess_summary <- data.frame(chess_summary)
head(chess_summary)## Player_Name State Total_Points Pre_Rating Avg_Opp_Pre_Rating
## 1 GARY HUA ON 6 1794 1605
## 2 DAKSHESH DARURI MI 6 1553 1469
## 3 ADITYA BAJAJ MI 6 1384 1564
## 4 PATRICK H SCHILLING MI 5.5 1716 1574
## 5 HANSHI ZUO MI 5.5 1655 1501
## 6 HANSEN SONG OH 5 1686 1519#write the .csv file
write.csv(chess_summary, file = "tournament_summary.csv")