In f0nzie/petro.One: Statistics and Text Mining for Oil and Gas Papers from OnePetro Metadata

Use this algorithm for a function that labels the discipline of each paper by looking at the keywords.

# read folder ./inst/extdata
# detect files ending .txt and put them in a vector or list
# read text file
    # read first line and put it as a discipline
    # read next lines (subjects) as rows and put them in a dataframe
    # variable1: subject; variable2: discipline
    # remove blank rows
    # keep the dataframe
# move to next file
# save dataframe as RDA file in ./data

# data files (.txt) are under `extdata`
extdata <- system.file("extdata", "disciplines", package = "petro.One")
extdata

# list text files in extdata
pat <- ".txt"
label_files <- list.files(path = extdata, pattern = pat, full.names = TRUE)
label_files

# read first line of each file to get the name of the discipline
# first line will always contain the discipline
discipline <- lapply(label_files, function(x) readLines(x, n = 1))
discipline

create dataframe of subjects, discipline

# create a dataframe of subjects and corresponding disciplines
cumdf <- data.frame()  # empty dataframe
for (i in seq_len(length(discipline))) {
    # read the discipline text file with its subjects
    df <- read.table(label_files[i], header=F, sep = "\n", stringsAsFactors = FALSE)
    df$V2 <- discipline[[i]]  # add the discipline to the subject
    cumdf <- rbind(cumdf, df) # accumulate rows
}
# set the name of the columns
names(cumdf) <- c("subject", "discipline")
cumdf

# save to .rda
labels <- cumdf
save(labels, file = "../data/disciplines.rda")

# read label files, one row at a time
datalist = lapply(label_files, function(x) read.table(x, header=F, sep = "\n")) 
datafr = do.call("rbind", datalist) 
datafr

utils::read.table(data_file, header = TRUE)

f0nzie/petro.One documentation built on May 29, 2019, 12:05 a.m.