view abstracts_by_pmids.R @ 0:3f4adc85ba5d draft

"planemo upload for repository https://github.com/dlal-group/simtext commit fd3f5b7b0506fbc460f2a281f694cb57f1c90a3c-dirty"
author dlalgroup
date Thu, 24 Sep 2020 02:01:50 +0000
parents
children
line wrap: on
line source

#!/usr/bin/env Rscript
#TOOL2 abstracts_by_pmids
#
#This tool retrieves for all PMIDs in each row of a table the according abstracts and saves them in additional columns.
#
#Input: Tab-delimited table with columns containing PMIDs. The names of the PMID columns should start with “PMID”, e.g. “PMID_1”, “PMID_2” etc.
#
#Output: Input table with additional columns containing abstracts corresponding to the PMIDs from PubMed. 
#The abstract columns are called "ABSTRACT_1", "ABSTARCT_2" etc. 
#
# Usage: $ T2_abstracts_by_pmid.R [-h] [-i INPUT] [-o OUTPUT]
# 
# optional arguments:
# -h, --help                 show help message
# -i INPUT, --input INPUT    input file name. add path if file is not in working directory
# -o OUTPUT, --output OUTPUT output file name. [default "T2_output"]


if ( '--install_packages' %in% commandArgs()) {
  print('Installing packages')
  if (!require('argparse')) install.packages('argparse', repo="http://cran.rstudio.com/");
  if (!require("reutils")) install.packages("reutils", repo="http://cran.rstudio.com/");
  if (!require('easyPubMed')) install.packages('easyPubMed', repo="http://cran.rstudio.com/" );
  if (!require('textclean')) install.packages('textclean', repo="http://cran.rstudio.com/");
}

suppressPackageStartupMessages(library("argparse"))
library("reutils")
suppressPackageStartupMessages(library("easyPubMed"))
suppressPackageStartupMessages(library("textclean"))

parser <- ArgumentParser()
parser$add_argument("-i", "--input", 
                    help = "input fie name. add path if file is not in workind directory")
parser$add_argument("-o", "--output", default="abstracts_by_pmids_output",
                    help = "output file name. [default \"%(default)s\"]")
parser$add_argument("--install_packages", action="store_true", default=FALSE,
                    help="If you want to auto install missing required packages.")

args <- parser$parse_args()

data = read.delim(args$input, stringsAsFactors=FALSE, header= TRUE, sep='\t')
pmids_cols_index <- grep("PMID", names(data))

fetch_abstracts = function(PMIDs, row){
  
  efetch_result <- NULL
  try_num <- 1
  t_0 <- Sys.time()
  
  while(is.null(efetch_result)) {
    
    # Timing check: kill at 3 min
    if (try_num > 1){
      Sys.sleep(time = 1*try_num)
      cat("Problem to receive PubMed data or error is received. Please wait. Try number: ",try_num,"\n") 
    }
    
    t_1 <- Sys.time()
    
    if(as.numeric(difftime(t_1, t_0, units = "mins")) > 3){
      message("Killing the request! Something is not working. Please, try again later","\n")
      return(data)
    }

    efetch_result <- tryCatch({    
      suppressWarnings(efetch(uid=PMIDs, db="pubmed", retmode = "xml"))
    }, error = function(e) {
      NULL
    })
    
    if(!is.null(as.list(efetch_result$errors)$error)){
      if (as.list(efetch_result$errors)$error == "HTTP error: Status 400; Bad Request") {
        efetch_result <- NULL
      }
    }
      
    try_num <- try_num + 1
    
  } #while loop end

  # articles to list
  xml_data <- strsplit(efetch_result$content, "<PubmedArticle(>|[[:space:]]+?.*>)")[[1]][-1]
  xml_data <- sapply(xml_data, function(x) {
    #trim extra stuff at the end of the record
    if (!grepl("</PubmedArticle>$", x))
      x <- sub("(^.*</PubmedArticle>).*$", "\\1", x) 
    # Rebuid XML structure and proceed
    x <- paste("<PubmedArticle>", x)
    gsub("[[:space:]]{2,}", " ", x)}, 
    USE.NAMES = FALSE, simplify = TRUE)
  
  abstract.text = sapply(xml_data, function(x){
    custom_grep(x, tag="AbstractText", format="char")}, 
    USE.NAMES = FALSE, simplify = TRUE)
  
  abstracts <- sapply(abstract.text, function(x){
    if (length(x) > 1){
      x <- paste(x, collapse = " ", sep = " ")
      x <- gsub("</{0,1}i>", "", x, ignore.case = T)
      x <- gsub("</{0,1}b>", "", x, ignore.case = T)
      x <- gsub("</{0,1}sub>", "", x, ignore.case = T)
      x <- gsub("</{0,1}exp>", "", x, ignore.case = T)
    } else if (length(x) < 1) {
      x <- NA
    } else {
      x <- gsub("</{0,1}i>", "", x, ignore.case = T)
      x <- gsub("</{0,1}b>", "", x, ignore.case = T)
      x <- gsub("</{0,1}sub>", "", x, ignore.case = T)
      x <- gsub("</{0,1}exp>", "", x, ignore.case = T)
    }
    x
  }, 
  USE.NAMES = FALSE, simplify = TRUE)
  
  abstracts = as.character(abstracts)
  
  if(length(abstracts)>0){
    data[row,sapply(1:length(abstracts),function(i){paste0("ABSTRACT_",i)})] <- abstracts
    cat(length(abstracts)," abstracts for PMIDs of row ", row, " are added in the table.","\n")
  }
  
  return(data)
}
    

for(row in 1:nrow(data)){
  PMIDs= as.character(unique(data[row, pmids_cols_index]))
  PMIDs = PMIDs[!PMIDs=="NA"]
  
  if(length(PMIDs) > 0){
    data = tryCatch(fetch_abstracts(PMIDs, row),
                    error=function(e){ 
                      Sys.sleep(3)
                      })
  } else {
    print(paste("No PMIDs in row", row))
  }
}

write.table(data, args$output, sep = '\t', row.names = FALSE, col.names = TRUE)