Nothing
R/get_refs.R
In citationchaser: Perform Forward and Backwards Chasing in Evidence Syntheses
Defines functions
get_citation
getLENSData
get_refs
Documented in
get_citation
getLENSData
get_refs
#' Automated citation chasing in systematic reviews
#'
#' @description This function takes a list of articles in the form of established
#' identifiers (e.g. digital object identifiers) and sends a request to the
#' lens.org API to firstly identify all cited references in all articles (in the
#' form of lists of lens IDs), and then query these lens IDs to bring back full
#' citation information for all listed records. Deduplicates references to the
#' same records across articles, resulting in an RIS file and a summary report
#' in the console.
#' @param article_list List of article identifiers for which the reference
#' lists will be returned. Must be a list/vector of identifiers, e.g.
#' '"10.1186/s13750-018-0126-2" "10.1002/jrsm.1378"'.
#' @param type Specification of the type of input provided. The default is
#' 'doi' (digital object identifier), but any of the following are accepted:
#' "pmid" (PubMed ID), "pmcid" (PubMed Central ID), "magid" (Microsoft
#' Academic ID), "coreid" (CORE identifier), lens_id" (The Lens.org ID),
#' "title" (article title; much lower specificity).
#' @param get_records Specification of whether to look for records referenced
#' within the input articles ('references'), records citing the input articles
#' ('citations'), or both ('both').
#' @param save_object Option to save the resultant ris file as an object in
#' the Global Environment. The default is FALSE.
#' @param token An access key for the lens.org API. Tokens can be obtained by
#' applying for scholarly API access and creating a token once approved. See
#' 'https://www.lens.org/lens/user/subscriptions#scholar' for further details.
#' @return An RIS file is saved to the working directory. A report is printed
#' to the console. If 'save_object=TRUE', the RIS file is returned as an
#' object
#' @importFrom maditr vlookup
#' @importFrom httr content
#' @importFrom jsonlite fromJSON
#' @importFrom utils write.table
#' @importFrom tibble tibble
#' @importFrom dplyr mutate group_split bind_rows
#' @importFrom MESS cumsumbinning
#' @export
#' @examples
#' \dontrun{
#' article_list <- c("10.1007/978-3-642-37048-9_13",
#' "10.1111/sum.12030",
#' "10.5194/bg-13-3619-2016",
#' "10.1016/j.agee.2012.09.006")
#' token <- 'token'
#' refs <- get_refs(article_list, get_records = 'references', token = token)
#' refs
#' }
get_refs <- function(article_list,
type = 'doi',
get_records,
save_object = FALSE,
token) {
# set the maximum number of results returned for each query
max_results <- 500
# if the number of input articles is 1, then split the input list where there are commas (and strip out whitespace)
if(length(article_list) == 1){
article_list <- trimws(unlist(strsplit(article_list, '[,]')))
}
print('Input record list:')
print(article_list)
## input article search
# build query for input article search
request1 <- paste0('{\n\t"query": {\n\t\t"terms": {\n\t\t\t"',type,'": ["', paste0('', paste(article_list, collapse = '", "'), '"'),']\n\t\t}\n\t},\n\t"size":500\n}')
print('Input article request:')
print(request1)
# perform article search and extract text results
data <- getLENSData(token, request1)
print('Input request executed.')
# report requests remaining within the limit (currently 50/min)
requests_remaining <- data[["headers"]][["x-rate-limit-remaining-request-per-minute"]]
print(paste0('Remaining requests = ', requests_remaining))
# extract the JSON content of the response
record_json <- httr::content(data, "text")
# convert json output from article search to list
record_list <- jsonlite::fromJSON(record_json)
print('Results converted to df from JSON output:')
print(record_list$results)
# error messages
if (data$status_code == 404){
print('Error 404: no matched records')
return('Warning: Your search returned no matched results. Please double check your input article identifiers and try again.')
}
if (data$status_code == 429){
print('Error 429: system overloaded')
return('Warning: Right now there are too many people using citationchaser. This has been logged and we will endeavour to increase the bandwith as soon as possible. Please try again later.')
}
if (record_list$total == 0){
print('Error 1: no matched records')
return('Warning: Your search returned no matched results. Please double check your input article identifiers and try again.')
}
# report number of input articles returned
input_number <- record_list[["total"]]
# list input article lens IDs (for later use)
articles_id <- record_list[["data"]][["lens_id"]]
print('Returned records from input request:')
print(articles_id)
### search for citations of input articles
if (get_records == 'citations') {
print('Seeking citations...')
# citations per article
citation_count <- record_list[["data"]][["scholarly_citations_count"]]
citation_count[is.na(citation_count)] <- 0
print('Citations per record sought:')
print(citation_count)
# sum of all citations
all_citations <- sum(citation_count, na.rm = TRUE)
print('Total citations sought:')
print(all_citations)
# return error message if input article(s) have no citations
#if (record_list[["data"]][["scholarly_citations_count"]] == 0 || is.null(record_list[["data"]][["scholarly_citations_count"]]) == TRUE){
if (is.null(all_citations) == TRUE || identical(all_citations, 0) == TRUE){
print('Error 2: no matched citations')
return('Warning: Your input articles have no recorded citations in the Lens.org database')
}
## group articles into chunks based on max number of results per search
# list citing articles per input article
cit_by_art <- record_list[["data"]][["scholarly_citations"]]
citations <- unlist(record_list[["data"]][["scholarly_citations"]])
print('Citations record list: ')
print(citations)
# remove duplicates across articles
citations_unique <- unique(citations)
print('Deduplicated citations record list:')
print(citations_unique)
# set up dataframe for groups of articles to search
cit_counts_df <- tibble::tibble(articles_id,
citation_count,
cit_by_art)
# remove records with no citations
cit_counts_df[is.na(cit_counts_df)] <- 0
cit_counts_df <- cit_counts_df[cit_counts_df$citation_count!=0,]
# subset of records that indidivually have more citations than the max number of results per query
single <- subset(cit_counts_df, citation_count >= max_results)
single$export_group <- 0
# subset of records that have fewer citations than the max number of results per query
rest <- subset(cit_counts_df, citation_count < max_results)
#the following code mistakenly bins around 500 (>500 records in some groups)
#rest <- mutate(rest,
# # divide by max citations allowed
# export_group = ceiling(cumsum(citation_count) / max_results))
#fix:
rest <- dplyr::mutate(rest, export_group = MESS::cumsumbinning(citation_count, 500))
# bind both dataframes back together
cit_counts_df <- rbind(single, rest)
print(cit_counts_df)
# get a list of vectors of the ids for searches within the export limits
citgroups <- dplyr::group_split(cit_counts_df, export_group)
citgroups_single <- dplyr::group_split(single, export_group)
citgroups_rest <- dplyr::group_split(rest, export_group)
print(citgroups_rest)
# define query function
run_request <- function(input){
# build query for article citations - this will pull back a maximum of 500 hits from lens.org
request <- paste0('{
"query": {
"terms": {
"lens_id": [', paste0('"', paste(input, collapse = '", "'), '"'), ']
}
},
"size":500,
"scroll": "1m",
"include": ["lens_id", "authors", "publication_type", "title", "external_ids", "start_page", "end_page", "volume", "issue", "references", "scholarly_citations", "source_urls", "abstract", "date_published", "year_published", "references_count", "scholarly_citations_count", "source"]
}')
# perform search and extract results
results <- getLENSData(token, request)
return(results)
}
cit_results <- data.frame()
# run the query function for each cluster of records, looping through and recording the timing
if(length(citgroups_rest) > 0){
tStart_cit <- Sys.time()
for (i in 1:length(citgroups_rest)){
print(paste0('Running group ', i, ' request...'))
print(unlist(citgroups_rest[[i]]$cit_by_art))
data_cit <- run_request(unlist(citgroups_rest[[i]]$cit_by_art))
requests_remaining <- data_cit[["headers"]][["x-rate-limit-remaining-request-per-minute"]]
# print the requests remaining to the Shinyapps log
print(paste0('Remaining requests = ', requests_remaining))
# extract and convert the results to a JSON
record_json_cit <- httr::content(data_cit, "text")
record_list_cit <- jsonlite::fromJSON(record_json_cit)
# error messages
if (data_cit$status_code == 404){
return('Warning: Your search returned no matched results. Please double check your input article identifiers and try again.')
}
if (data_cit$status_code == 429){
return('Warning: Right now there are too many people using citationchaser. This has been logged and we will endeavour to increase the bandwith as soon as possible. Please try again later.')
}
if (record_list_cit$total == 0){
return('Warning: Your search returned no matched results. Please double check your input article identifiers and try again.')
}
# convert the results to a dataframe
record_list_cit_df <- as.data.frame(record_list_cit)
cit_results <- dplyr::bind_rows(cit_results, record_list_cit_df)
tEnd_cit <- Sys.time()
# back off if the requests per minute is close to the limit (currently 50/min)
if (data_cit[["headers"]][["x-rate-limit-remaining-request-per-minute"]] < 1){
t_cit <- tEnd_cit - tStart_cit
Sys.sleep(60 - t)
}
}
}
## cursor-based pagination for any single-record query with more than 500 citations
# only run if there are single-record queries
if (length(citgroups_single) != 0){
runrequest <- function(input){
request <- paste0('{
"query": {
"terms": {
"lens_id": [', paste0('"', paste(input, collapse = '", "'), '"'), ']
}
},
"size": ',max_results,',
"scroll": "1m",
"include": ["lens_id", "authors", "publication_type", "title", "external_ids", "start_page", "end_page", "volume", "issue", "references", "scholarly_citations", "source_urls", "abstract", "date_published", "year_published", "references_count", "scholarly_citations_count", "source"]
}')
# perform article search and extract text results
data <- getLENSData(token, request)
requests_remaining <- data[["headers"]][["x-rate-limit-remaining-request-per-minute"]]
# print the requests remaining to the Shinyapps log
print(paste0('Remaining requests = ', requests_remaining))
# extract and convert the results to a JSON
record_json <- httr::content(data, "text")
record_list <- jsonlite::fromJSON(record_json)
# error messages
if (data$status_code == 404){
return('Warning: Your search returned no matched results. Please double check your input article identifiers and try again.')
}
if (data$status_code == 429){
return('Warning: Right now there are too many people using citationchaser. This has been logged and we will endeavour to increase the bandwith as soon as possible. Please try again later.')
}
if (record_list$total == 0){
return('Warning: Your search returned no matched results. Please double check your input article identifiers and try again.')
}
# convert to a dataframe
record_df <- data.frame(record_list)
# if a result contains more than the max number of records per request, use cursor-based pagination
if(record_list[["total"]] > max_results) {
sets <- ceiling(record_list[["total"]] / max_results) # calculate the number of queries needed for those with more than the max number of results
scroll_id <- record_list[["scroll_id"]] # extract the scroll id from the query to go back to the same search
for (i in 2:sets){ # loop through the sets of results needed to bring back all records into a dataframe
scroll_id <- record_list[["scroll_id"]] #extract the latest scroll_id from the last query
request <- paste0('{"scroll_id": "', # new query based on scroll_id and including 'include' for efficiency
scroll_id,
'","include": ["lens_id", "authors", "publication_type", "title", "external_ids", "start_page", "end_page", "volume", "issue", "references", "scholarly_citations", "source_urls", "abstract", "date_published", "year_published", "references_count", "scholarly_citations_count", "source"]}')
# perform article search and extract text results
data <- getLENSData(token, request)
requests_remaining <- data[["headers"]][["x-rate-limit-remaining-request-per-minute"]]
print(paste0('Remaining requests = ', requests_remaining))
record_json <- httr::content(data, "text")
record_list <- jsonlite::fromJSON(record_json) # convert json output from article search to list
# error messages
if (data$status_code == 404){
return('Warning: Your search returned no matched results. Please double check your input article identifiers and try again.')
}
if (data$status_code == 429){
return('Warning: Right now there are too many people using citationchaser. This has been logged and we will endeavour to increase the bandwith as soon as possible. Please try again later.')
}
if (record_list$total == 0){
return('Warning: Your search returned no matched results. Please double check your input article identifiers and try again.')
}
new_df <- data.frame(record_list)
# output
record_df <- dplyr::bind_rows(record_df,new_df) # bind the latest search dataframe to the previous dataframe
}
}
return(record_df)
}
# loop through single-record queries that are less than the maximum allowed results per query
for (i in 1:length(citgroups_single)){
data_cit <- runrequest(unlist(citgroups_single[[i]]$cit_by_art))
requests_remaining <- data_cit[["headers"]][["x-rate-limit-remaining-request-per-minute"]]
print(paste0('Remaining requests = ', requests_remaining))
# error messages
if (data$status_code == 404){
return('Warning: Your search returned no matched results. Please double check your input article identifiers and try again.')
}
if (data$status_code == 429){
return('Warning: Right now there are too many people using citationchaser. This has been logged and we will endeavour to increase the bandwith as soon as possible. Please try again later.')
}
cit_results <- dplyr::bind_rows(cit_results, data_cit)
#tEnd_cit <- Sys.time()
#if (data_cit[["headers"]][["x-rate-limit-remaining-request-per-minute"]] < 1){
# t_cit <- tEnd_cit - tStart_cit
# Sys.sleep(60 - t) # pause to limit requests below 50 requests/min
#}
}
}
# remove duplicate records
cit_results <- cit_results[!duplicated(cit_results$data.lens_id),]
all_results_cit <- cit_results$data.lens_id
# convert json to ris style
type_list <- data.frame(type = c("ABST", "BOOK", "CHAP", "COMP", "CONF", "DATA", "JOUR"),
description = c("abstract reference", "whole book reference", "book chapter reference", "computer program", "conference proceeding", "data file", "journal/periodical reference"),
publication_type = c("reference entry", "book", "book chapter", "component", "conference proceedings", "dataset", "journal article"))
publication_type_cit <- cit_results$data.publication_type
authors_cit <- list()
for (i in 1:length(cit_results$data.authors)) {
authors_cit <- unlist(c(authors_cit, paste0(cit_results$data.authors[[i]]$last_name, ', ',
cit_results$data.authors[[i]]$first_name, collapse = '; ')))
}
title_cit <- cit_results$data.title
year_cit <- cit_results$data.year_published
abstract_cit <- cit_results$data.abstract
start_page_cit <- cit_results$data.start_page
end_page_cit <- cit_results$data.end_page
source_title_cit <- list()
for (i in 1:length(cit_results[,1])) {
source_title_cit <- unlist(c(source_title_cit, cit_results$data.source[[1]][i]))
}
volume_cit <- cit_results$data.volume
issue_cit <- cit_results$data.issue
publisher_cit <- cit_results$data.source.publisher
issn_cit <- cit_results$data.source.issn
doi_cit <- unlist(lapply(cit_results$data.external_ids, function(ch) maditr::vlookup('doi', ch, result_column = 'value', lookup_column = 'type')))
# generate data table for Shiny UI
level1_table_cit <- data.table::data.table(authors = authors_cit,
year = year_cit,
title = title_cit,
source_title = source_title_cit,
publisher = publisher_cit,
volume = volume_cit,
issue = issue_cit,
start_page = start_page_cit,
end_page = end_page_cit,
doi = doi_cit)
# generate RIS file
level1_ris_cit <- paste(paste0('\n',
'TY - ', maditr::vlookup(publication_type_cit, type_list, result_column = 'type', lookup_column = 'publication_type'), '\n',
'AU - ', authors_cit, '\n',
'TI - ', title_cit, '\n',
'PY - ', year_cit, '\n',
'AB - ', abstract_cit, '\n',
'SP - ', start_page_cit, '\n',
'EP - ', end_page_cit, '\n',
'JF - ', source_title_cit, '\n',
'VL - ', volume_cit, '\n',
'IS - ', issue_cit, '\n',
'PB - ', publisher_cit, '\n',
# 'SN - ', issn_cit, '\n',
'DO - ', doi_cit, '\n',
'ER - '),
collapse = '\n')
# generate ris build report
ris_records_cit <- lengths(regmatches(level1_ris_cit, gregexpr("TY - ", level1_ris_cit)))
stage1_report_cit <- paste0('Your ', scales::comma(input_number), ' articles were cited a total of ', scales::comma(all_citations), ' times. This corresponds to ',
scales::comma(length(citations_unique)), ' unique article IDs. Your RIS file is ready for download and contains ',
scales::comma(ris_records_cit), ' records exported from Lens.org.')
report_cit <- stage1_report_cit
return(list(display = level1_table_cit, ris = level1_ris_cit, report = report_cit, df = cit_results))
### search for references articles
} else if (get_records == 'references') {
if (is.null(record_list[["data"]][["references_count"]]) == TRUE || identical(record_list[["data"]][["references_count"]], 0) == TRUE){
return('Warning: Your input articles contained no references in the Lens.org database')
}
# obtain reference lists from article search
reference_count <- record_list[["data"]][["references_count"]]
reference_count[is.na(reference_count)] <- 0
all_references <- sum(reference_count, na.rm = TRUE)
ref_by_art <- record_list[["data"]][["references"]]
references <- unlist(record_list[["data"]][["references"]])
references_unique <- unique(references)
deduped_references <- length(references_unique)
ref_counts_df <- tibble::tibble(articles_id,
reference_count,
ref_by_art)
ref_counts_df[is.na(ref_counts_df)] <- 0
#below code not working because it was tagging excess records from one group onto the next (of 500)
#ref_counts_df <- dplyr::mutate(ref_counts_df,
# # cumulatively add up citation count
# cumulative_n = cumsum(reference_count),
# # divide by max citations allowed
# export_group = floor(cumsum(reference_count) / 500))
ref_counts_df <- dplyr::mutate(ref_counts_df, export_group = MESS::cumsumbinning(reference_count, 500))
# get a list of vectors of the ids
refgroups <- dplyr::group_split(ref_counts_df, export_group)
# define query function
run_request <- function(input){
# build query for article citations - this will pull back a maximum of 1,000 hits from lens.org, so not great if lots of refs
request <- paste0('{
"query": {
"terms": {
"lens_id": [', paste0('"', paste(input, collapse = '", "'), '"'), ']
}
},
"size":500,
"include": ["lens_id", "authors", "publication_type", "title", "external_ids", "start_page", "end_page", "volume", "issue", "references", "scholarly_citations", "source_urls", "abstract", "date_published", "year_published", "references_count", "scholarly_citations_count", "source"]
}')
#perform references search and extract text results
results <- getLENSData(token, request)
return(results)
}
# run the query function for each tibble, adding to a final dataset
ref_results <- data.frame()
tStart_ref <- Sys.time()
for (i in 1:length(refgroups)){
data_ref <- run_request(unique(unlist(refgroups[[i]]$ref_by_art)))
requests_remaining <- data_ref[["headers"]][["x-rate-limit-remaining-request-per-minute"]]
print(paste0('Remaining requests = ', requests_remaining))
record_json_ref <- httr::content(data_ref, "text")
record_list_ref <- jsonlite::fromJSON(record_json_ref)
# error messages
if (data_ref$status_code == 404){
return('Warning: Your search returned no matched results. Please double check your input article identifiers and try again.')
}
if (data_ref$status_code == 429){
return('Warning: Right now there are too many people using citationchaser. This has been logged and we will endeavour to increase the bandwith as soon as possible. Please try again later.')
}
if (record_list_ref$total == 0){
return('Warning: Your search returned no matched results. Please double check your input article identifiers and try again.')
}
record_list_ref_df <- as.data.frame(record_list_ref)
ref_results <- dplyr::bind_rows(ref_results, record_list_ref_df)
tEnd_ref <- Sys.time()
if (data_ref[["headers"]][["x-rate-limit-remaining-request-per-minute"]] < 1){
t_ref <- tEnd_ref - tStart_ref
Sys.sleep(60 - t) # pause to limit requests below 50 requests/min
}
}
ref_results <- ref_results[!duplicated(ref_results$data.lens_id),]
all_results_ref <- ref_results$data.lens_id
# convert json to ris style
type_list <- data.frame(type = c("ABST", "BOOK", "CHAP", "COMP", "CONF", "DATA", "JOUR"),
description = c("abstract reference", "whole book reference", "book chapter reference", "computer program", "conference proceeding", "data file", "journal/periodical reference"),
publication_type = c("reference entry", "book", "book chapter", "component", "conference proceedings", "dataset", "journal article"))
publication_type_ref <- ref_results$data.publication_type
authors_ref <- list()
for (i in 1:length(ref_results$data.authors)) {
authors_ref <- unlist(c(authors_ref, paste0(ref_results$data.authors[[i]]$last_name, ', ',
ref_results$data.authors[[i]]$first_name, collapse = '; ')))
}
title_ref <- ref_results$data.title
year_ref <- ref_results$data.year_published
abstract_ref <- ref_results$data.abstract
start_page_ref <- ref_results$data.start_page
end_page_ref <- ref_results$data.end_page
source_title_ref <- list()
for (i in 1:length(ref_results[,1])) {
source_title_ref <- unlist(c(source_title_ref, ref_results$data.source[[1]][i]))
}
volume_ref <- ref_results$data.volume
issue_ref <- ref_results$data.issue
publisher_ref <- ref_results$data.source.publisher
issn_ref <- ref_results$data.source.issn
doi_ref <- unlist(lapply(ref_results$data.external_ids, function(ch) maditr::vlookup('doi', ch, result_column = 'value', lookup_column = 'type')))
level1_table_ref <- data.table::data.table(authors = authors_ref,
year = year_ref,
title = title_ref,
source_title = source_title_ref,
publisher = publisher_ref,
volume = volume_ref,
issue = issue_ref,
start_page = start_page_ref,
end_page = end_page_ref,
doi = doi_ref)
level1_ris_ref <- paste(paste0('\n',
'TY - ', maditr::vlookup(publication_type_ref, type_list, result_column = 'type', lookup_column = 'publication_type'), '\n',
'AU - ', authors_ref, '\n',
'TI - ', title_ref, '\n',
'PY - ', year_ref, '\n',
'AB - ', abstract_ref, '\n',
'SP - ', start_page_ref, '\n',
'EP - ', end_page_ref, '\n',
'JF - ', source_title_ref, '\n',
'VL - ', volume_ref, '\n',
'IS - ', issue_ref, '\n',
'PB - ', publisher_ref, '\n',
# 'SN - ', issn_ref, '\n',
'DO - ', doi_ref, '\n',
'ER - '),
collapse = '\n')
# ris build report
ris_records_ref <- lengths(regmatches(level1_ris_ref, gregexpr("TY - ", level1_ris_ref)))
stage1_report_ref <- paste0('Your ', scales::comma(input_number), ' articles contained a total of ', scales::comma(all_references), ' references. This corresponds to ',
scales::comma(deduped_references), ' unique IDs. Your RIS file is ready for download and contains ', scales::comma(ris_records_ref), ' records exported from Lens.org.')
report_ref <- stage1_report_ref
return(list(display = level1_table_ref, ris = level1_ris_ref, report = report_ref, df = ref_results))
}
}
#' Function to query Lens.org
#'
#' @description Function written by lens.org for use of their API.
#' @param token An access key for the lens.org API. Tokens can be obtained by
#' applying for scholarly API access and creating a token once approved. See
#' 'https://www.lens.org/lens/user/subscriptions#scholar' for further details.
#' @param query A search string formulated according to the Lens.org API
#' documentation: 'https://docs.api.lens.org/request-scholar.html'.
#' @return A summary response. The results are viewable using
#' 'content(data, "text")'. Other details regarding the request (e.g. repsonse
#' times) can be accessed through the main output.
#' @importFrom httr add_headers POST
getLENSData <- function(token, query){
url <- 'https://api.lens.org/scholarly/search'
headers <- c('Authorization' = token, 'Content-Type' = 'application/json')
httr::POST(url = url, httr::add_headers(.headers=headers), body = query)
}
#' Find citation based on identifier
#'
#' @description This function takes a list of articles in the form of established
#' identifiers (e.g. digital object identifiers) and sends a request to the
#' lens.org API to obtain full citation information from the Lens database for
#' all sought articles.
#' @param article_list List of article identifiers for which the reference
#' lists will be returned. Must be a list/vector of identifiers, e.g.
#' '"10.1186/s13750-018-0126-2" "10.1002/jrsm.1378"'.
#' @param type Specification of the type of input provided. The default is
#' 'doi' (digital object identifier), but any of the following are accepted:
#' "pmid" (PubMed ID), "pmcid" (PubMed Central ID), "magid" (Microsoft
#' Academic ID), "coreid" (CORE identifier), lens_id" (The Lens.org ID),
#' "title" (article title; much lower specificity).
#' @param token An access key for the lens.org API. Tokens can be obtained by
#' applying for scholarly API access and creating a token once approved. See
#' 'https://www.lens.org/lens/user/subscriptions#scholar' for further details.
#' @return A dataframe containing the matching citation from Lens.org.
#' @importFrom maditr vlookup
#' @importFrom httr content
#' @importFrom jsonlite fromJSON
#' @importFrom data.table data.table
#' @examples
#' \dontrun{
#' article_list <- c("10.1007/978-3-642-37048-9_13", "10.1111/sum.12030",
#' "10.5194/bg-13-3619-2016", "10.1016/j.agee.2012.09.006")
#' results <- get_citation(article_list)
#' articles <- results$display
#' }
#' @export
get_citation <- function(article_list,
type = 'doi',
token = 'WCFlpCtuJXYI1sDhZcZ8y7hHpri0SEmTnLNkeU4OEM5JTQRNXB9w'){
if(length(article_list) == 1){
article_list <- trimws(unlist(strsplit(article_list, '[,]')))
}
request <- paste0('{\n\t"query": {\n\t\t"terms": {\n\t\t\t"', type, '": ["', paste0('', paste(article_list, collapse = '", "'), '"'),']\n\t\t}\n\t},\n\t"size":500\n}')
# perform article search and extract text results
data <- getLENSData(token, request)
requests_remaining <- data[["headers"]][["x-rate-limit-remaining-request-per-minute"]]
print(paste0('Remaining requests = ', requests_remaining))
record_json <- httr::content(data, "text")
# convert json output from article search to list
record_list <- jsonlite::fromJSON(record_json)
# error messages
if (data$status_code == 404){
return('Warning: Your search returned no matched results. Please double check your input article identifiers and try again.')
}
if (data$status_code == 429){
return('Warning: Right now there are too many people using citationchaser. This has been logged and we will endeavour to increase the bandwith as soon as possible. Please try again later.')
}
if (record_list$total == 0){
return('Warning: Your search returned no matched results. Please double check your input article identifiers and try again.')
}
inputs_df <- as.data.frame(record_list)
# citations and references
citation_count <- record_list[["data"]][["scholarly_citations_count"]]
reference_count <- record_list[["data"]][["references_count"]]
# convert json to ris style
type_list <- data.frame(type = c("ABST", "BOOK", "CHAP", "COMP", "CONF", "DATA", "JOUR"),
description = c("abstract reference", "whole book reference", "book chapter reference", "computer program", "conference proceeding", "data file", "journal/periodical reference"),
publication_type = c("reference entry", "book", "book chapter", "component", "conference proceedings", "dataset", "journal article"))
publication_type <- record_list[["data"]][["publication_type"]]
authors <- list()
for (i in 1:length(record_list[["data"]][["authors"]])) {
authors <- unlist(c(authors, paste0(record_list[["data"]][["authors"]][[i]]$last_name, ', ',
record_list[["data"]][["authors"]][[i]]$first_name, collapse = '; ')))
}
title <- record_list[["data"]][["title"]]
year <- record_list[["data"]][["year_published"]]
abstract <- record_list[["data"]][["abstract"]]
start_page <- record_list[["data"]][["start_page"]]
end_page <- record_list[["data"]][["end_page"]]
source_title <- record_list[["data"]][["source"]][["title"]]
volume <- record_list[["data"]][["volume"]]
issue <- record_list[["data"]][["issue"]]
publisher <- record_list[["data"]][["source"]][["publisher"]]
issn <- record_list[["data"]][["source"]][["issn"]]
doi <- unlist(lapply(record_list[["data"]][["external_ids"]], function(ch) maditr::vlookup('doi', ch, result_column = 'value', lookup_column = 'type')))
article_table <- data.table::data.table(authors = authors,
year = year,
title = title,
source_title = source_title,
publisher = publisher,
volume = volume,
issue = issue,
start_page = start_page,
end_page = end_page,
doi = doi,
Lens_refs = reference_count,
Lens_cited = citation_count)
article_ris <- paste(paste0('\n',
'TY - ', maditr::vlookup(publication_type, type_list, result_column = 'type', lookup_column = 'publication_type'), '\n',
'AU - ', authors, '\n',
'TI - ', title, '\n',
'PY - ', year, '\n',
'AB - ', abstract, '\n',
'SP - ', start_page, '\n',
'EP - ', end_page, '\n',
'JF - ', source_title, '\n',
'VL - ', volume, '\n',
'IS - ', issue, '\n',
'PB - ', publisher, '\n',
# 'SN - ', issn, '\n',
'DO - ', doi, '\n',
'ER - '),
collapse = '\n')
return(list(display = article_table, ris = article_ris, df = inputs_df))
}
Try the citationchaser package in your browser
Any scripts or data that you put into this service are public.
citationchaser documentation built on March 18, 2022, 7:56 p.m.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Defines functions get_citation getLENSData get_refs
Documented in get_citation getLENSData get_refs
#' Automated citation chasing in systematic reviews
#'
#' @description This function takes a list of articles in the form of established
#' identifiers (e.g. digital object identifiers) and sends a request to the
#' lens.org API to firstly identify all cited references in all articles (in the
#' form of lists of lens IDs), and then query these lens IDs to bring back full
#' citation information for all listed records. Deduplicates references to the
#' same records across articles, resulting in an RIS file and a summary report
#' in the console.
#' @param article_list List of article identifiers for which the reference
#' lists will be returned. Must be a list/vector of identifiers, e.g.
#' '"10.1186/s13750-018-0126-2" "10.1002/jrsm.1378"'.
#' @param type Specification of the type of input provided. The default is
#' 'doi' (digital object identifier), but any of the following are accepted:
#' "pmid" (PubMed ID), "pmcid" (PubMed Central ID), "magid" (Microsoft
#' Academic ID), "coreid" (CORE identifier), lens_id" (The Lens.org ID),
#' "title" (article title; much lower specificity).
#' @param get_records Specification of whether to look for records referenced
#' within the input articles ('references'), records citing the input articles
#' ('citations'), or both ('both').
#' @param save_object Option to save the resultant ris file as an object in
#' the Global Environment. The default is FALSE.
#' @param token An access key for the lens.org API. Tokens can be obtained by
#' applying for scholarly API access and creating a token once approved. See
#' 'https://www.lens.org/lens/user/subscriptions#scholar' for further details.
#' @return An RIS file is saved to the working directory. A report is printed
#' to the console. If 'save_object=TRUE', the RIS file is returned as an
#' object
#' @importFrom maditr vlookup
#' @importFrom httr content
#' @importFrom jsonlite fromJSON
#' @importFrom utils write.table
#' @importFrom tibble tibble
#' @importFrom dplyr mutate group_split bind_rows
#' @importFrom MESS cumsumbinning
#' @export
#' @examples
#' \dontrun{
#' article_list <- c("10.1007/978-3-642-37048-9_13",
#' "10.1111/sum.12030",
#' "10.5194/bg-13-3619-2016",
#' "10.1016/j.agee.2012.09.006")
#' token <- 'token'
#' refs <- get_refs(article_list, get_records = 'references', token = token)
#' refs
#' }
get_refs <- function(article_list,
type = 'doi',
get_records,
save_object = FALSE,
token) {
# set the maximum number of results returned for each query
max_results <- 500
# if the number of input articles is 1, then split the input list where there are commas (and strip out whitespace)
if(length(article_list) == 1){
article_list <- trimws(unlist(strsplit(article_list, '[,]')))
}
print('Input record list:')
print(article_list)
## input article search
# build query for input article search
request1 <- paste0('{\n\t"query": {\n\t\t"terms": {\n\t\t\t"',type,'": ["', paste0('', paste(article_list, collapse = '", "'), '"'),']\n\t\t}\n\t},\n\t"size":500\n}')
print('Input article request:')
print(request1)
# perform article search and extract text results
data <- getLENSData(token, request1)
print('Input request executed.')
# report requests remaining within the limit (currently 50/min)
requests_remaining <- data[["headers"]][["x-rate-limit-remaining-request-per-minute"]]
print(paste0('Remaining requests = ', requests_remaining))
# extract the JSON content of the response
record_json <- httr::content(data, "text")
# convert json output from article search to list
record_list <- jsonlite::fromJSON(record_json)
print('Results converted to df from JSON output:')
print(record_list$results)
# error messages
if (data$status_code == 404){
print('Error 404: no matched records')
return('Warning: Your search returned no matched results. Please double check your input article identifiers and try again.')
}
if (data$status_code == 429){
print('Error 429: system overloaded')
return('Warning: Right now there are too many people using citationchaser. This has been logged and we will endeavour to increase the bandwith as soon as possible. Please try again later.')
}
if (record_list$total == 0){
print('Error 1: no matched records')
return('Warning: Your search returned no matched results. Please double check your input article identifiers and try again.')
}
# report number of input articles returned
input_number <- record_list[["total"]]
# list input article lens IDs (for later use)
articles_id <- record_list[["data"]][["lens_id"]]
print('Returned records from input request:')
print(articles_id)
### search for citations of input articles
if (get_records == 'citations') {
print('Seeking citations...')
# citations per article
citation_count <- record_list[["data"]][["scholarly_citations_count"]]
citation_count[is.na(citation_count)] <- 0
print('Citations per record sought:')
print(citation_count)
# sum of all citations
all_citations <- sum(citation_count, na.rm = TRUE)
print('Total citations sought:')
print(all_citations)
# return error message if input article(s) have no citations
#if (record_list[["data"]][["scholarly_citations_count"]] == 0 || is.null(record_list[["data"]][["scholarly_citations_count"]]) == TRUE){
if (is.null(all_citations) == TRUE || identical(all_citations, 0) == TRUE){
print('Error 2: no matched citations')
return('Warning: Your input articles have no recorded citations in the Lens.org database')
}
## group articles into chunks based on max number of results per search
# list citing articles per input article
cit_by_art <- record_list[["data"]][["scholarly_citations"]]
citations <- unlist(record_list[["data"]][["scholarly_citations"]])
print('Citations record list: ')
print(citations)
# remove duplicates across articles
citations_unique <- unique(citations)
print('Deduplicated citations record list:')
print(citations_unique)
# set up dataframe for groups of articles to search
cit_counts_df <- tibble::tibble(articles_id,
citation_count,
cit_by_art)
# remove records with no citations
cit_counts_df[is.na(cit_counts_df)] <- 0
cit_counts_df <- cit_counts_df[cit_counts_df$citation_count!=0,]
# subset of records that indidivually have more citations than the max number of results per query
single <- subset(cit_counts_df, citation_count >= max_results)
single$export_group <- 0
# subset of records that have fewer citations than the max number of results per query
rest <- subset(cit_counts_df, citation_count < max_results)
#the following code mistakenly bins around 500 (>500 records in some groups)
#rest <- mutate(rest,
# # divide by max citations allowed
# export_group = ceiling(cumsum(citation_count) / max_results))
#fix:
rest <- dplyr::mutate(rest, export_group = MESS::cumsumbinning(citation_count, 500))
# bind both dataframes back together
cit_counts_df <- rbind(single, rest)
print(cit_counts_df)
# get a list of vectors of the ids for searches within the export limits
citgroups <- dplyr::group_split(cit_counts_df, export_group)
citgroups_single <- dplyr::group_split(single, export_group)
citgroups_rest <- dplyr::group_split(rest, export_group)
print(citgroups_rest)
# define query function
run_request <- function(input){
# build query for article citations - this will pull back a maximum of 500 hits from lens.org
request <- paste0('{
"query": {
"terms": {
"lens_id": [', paste0('"', paste(input, collapse = '", "'), '"'), ']
}
},
"size":500,
"scroll": "1m",
"include": ["lens_id", "authors", "publication_type", "title", "external_ids", "start_page", "end_page", "volume", "issue", "references", "scholarly_citations", "source_urls", "abstract", "date_published", "year_published", "references_count", "scholarly_citations_count", "source"]
}')
# perform search and extract results
results <- getLENSData(token, request)
return(results)
}
cit_results <- data.frame()
# run the query function for each cluster of records, looping through and recording the timing
if(length(citgroups_rest) > 0){
tStart_cit <- Sys.time()
for (i in 1:length(citgroups_rest)){
print(paste0('Running group ', i, ' request...'))
print(unlist(citgroups_rest[[i]]$cit_by_art))
data_cit <- run_request(unlist(citgroups_rest[[i]]$cit_by_art))
requests_remaining <- data_cit[["headers"]][["x-rate-limit-remaining-request-per-minute"]]
# print the requests remaining to the Shinyapps log
print(paste0('Remaining requests = ', requests_remaining))
# extract and convert the results to a JSON
record_json_cit <- httr::content(data_cit, "text")
record_list_cit <- jsonlite::fromJSON(record_json_cit)
# error messages
if (data_cit$status_code == 404){
return('Warning: Your search returned no matched results. Please double check your input article identifiers and try again.')
}
if (data_cit$status_code == 429){
return('Warning: Right now there are too many people using citationchaser. This has been logged and we will endeavour to increase the bandwith as soon as possible. Please try again later.')
}
if (record_list_cit$total == 0){
return('Warning: Your search returned no matched results. Please double check your input article identifiers and try again.')
}
# convert the results to a dataframe
record_list_cit_df <- as.data.frame(record_list_cit)
cit_results <- dplyr::bind_rows(cit_results, record_list_cit_df)
tEnd_cit <- Sys.time()
# back off if the requests per minute is close to the limit (currently 50/min)
if (data_cit[["headers"]][["x-rate-limit-remaining-request-per-minute"]] < 1){
t_cit <- tEnd_cit - tStart_cit
Sys.sleep(60 - t)
}
}
}
## cursor-based pagination for any single-record query with more than 500 citations
# only run if there are single-record queries
if (length(citgroups_single) != 0){
runrequest <- function(input){
request <- paste0('{
"query": {
"terms": {
"lens_id": [', paste0('"', paste(input, collapse = '", "'), '"'), ']
}
},
"size": ',max_results,',
"scroll": "1m",
"include": ["lens_id", "authors", "publication_type", "title", "external_ids", "start_page", "end_page", "volume", "issue", "references", "scholarly_citations", "source_urls", "abstract", "date_published", "year_published", "references_count", "scholarly_citations_count", "source"]
}')
# perform article search and extract text results
data <- getLENSData(token, request)
requests_remaining <- data[["headers"]][["x-rate-limit-remaining-request-per-minute"]]
# print the requests remaining to the Shinyapps log
print(paste0('Remaining requests = ', requests_remaining))
# extract and convert the results to a JSON
record_json <- httr::content(data, "text")
record_list <- jsonlite::fromJSON(record_json)
# error messages
if (data$status_code == 404){
return('Warning: Your search returned no matched results. Please double check your input article identifiers and try again.')
}
if (data$status_code == 429){
return('Warning: Right now there are too many people using citationchaser. This has been logged and we will endeavour to increase the bandwith as soon as possible. Please try again later.')
}
if (record_list$total == 0){
return('Warning: Your search returned no matched results. Please double check your input article identifiers and try again.')
}
# convert to a dataframe
record_df <- data.frame(record_list)
# if a result contains more than the max number of records per request, use cursor-based pagination
if(record_list[["total"]] > max_results) {
sets <- ceiling(record_list[["total"]] / max_results) # calculate the number of queries needed for those with more than the max number of results
scroll_id <- record_list[["scroll_id"]] # extract the scroll id from the query to go back to the same search
for (i in 2:sets){ # loop through the sets of results needed to bring back all records into a dataframe
scroll_id <- record_list[["scroll_id"]] #extract the latest scroll_id from the last query
request <- paste0('{"scroll_id": "', # new query based on scroll_id and including 'include' for efficiency
scroll_id,
'","include": ["lens_id", "authors", "publication_type", "title", "external_ids", "start_page", "end_page", "volume", "issue", "references", "scholarly_citations", "source_urls", "abstract", "date_published", "year_published", "references_count", "scholarly_citations_count", "source"]}')
# perform article search and extract text results
data <- getLENSData(token, request)
requests_remaining <- data[["headers"]][["x-rate-limit-remaining-request-per-minute"]]
print(paste0('Remaining requests = ', requests_remaining))
record_json <- httr::content(data, "text")
record_list <- jsonlite::fromJSON(record_json) # convert json output from article search to list
# error messages
if (data$status_code == 404){
return('Warning: Your search returned no matched results. Please double check your input article identifiers and try again.')
}
if (data$status_code == 429){
return('Warning: Right now there are too many people using citationchaser. This has been logged and we will endeavour to increase the bandwith as soon as possible. Please try again later.')
}
if (record_list$total == 0){
return('Warning: Your search returned no matched results. Please double check your input article identifiers and try again.')
}
new_df <- data.frame(record_list)
# output
record_df <- dplyr::bind_rows(record_df,new_df) # bind the latest search dataframe to the previous dataframe
}
}
return(record_df)
}
# loop through single-record queries that are less than the maximum allowed results per query
for (i in 1:length(citgroups_single)){
data_cit <- runrequest(unlist(citgroups_single[[i]]$cit_by_art))
requests_remaining <- data_cit[["headers"]][["x-rate-limit-remaining-request-per-minute"]]
print(paste0('Remaining requests = ', requests_remaining))
# error messages
if (data$status_code == 404){
return('Warning: Your search returned no matched results. Please double check your input article identifiers and try again.')
}
if (data$status_code == 429){
return('Warning: Right now there are too many people using citationchaser. This has been logged and we will endeavour to increase the bandwith as soon as possible. Please try again later.')
}
cit_results <- dplyr::bind_rows(cit_results, data_cit)
#tEnd_cit <- Sys.time()
#if (data_cit[["headers"]][["x-rate-limit-remaining-request-per-minute"]] < 1){
# t_cit <- tEnd_cit - tStart_cit
# Sys.sleep(60 - t) # pause to limit requests below 50 requests/min
#}
}
}
# remove duplicate records
cit_results <- cit_results[!duplicated(cit_results$data.lens_id),]
all_results_cit <- cit_results$data.lens_id
# convert json to ris style
type_list <- data.frame(type = c("ABST", "BOOK", "CHAP", "COMP", "CONF", "DATA", "JOUR"),
description = c("abstract reference", "whole book reference", "book chapter reference", "computer program", "conference proceeding", "data file", "journal/periodical reference"),
publication_type = c("reference entry", "book", "book chapter", "component", "conference proceedings", "dataset", "journal article"))
publication_type_cit <- cit_results$data.publication_type
authors_cit <- list()
for (i in 1:length(cit_results$data.authors)) {
authors_cit <- unlist(c(authors_cit, paste0(cit_results$data.authors[[i]]$last_name, ', ',
cit_results$data.authors[[i]]$first_name, collapse = '; ')))
}
title_cit <- cit_results$data.title
year_cit <- cit_results$data.year_published
abstract_cit <- cit_results$data.abstract
start_page_cit <- cit_results$data.start_page
end_page_cit <- cit_results$data.end_page
source_title_cit <- list()
for (i in 1:length(cit_results[,1])) {
source_title_cit <- unlist(c(source_title_cit, cit_results$data.source[[1]][i]))
}
volume_cit <- cit_results$data.volume
issue_cit <- cit_results$data.issue
publisher_cit <- cit_results$data.source.publisher
issn_cit <- cit_results$data.source.issn
doi_cit <- unlist(lapply(cit_results$data.external_ids, function(ch) maditr::vlookup('doi', ch, result_column = 'value', lookup_column = 'type')))
# generate data table for Shiny UI
level1_table_cit <- data.table::data.table(authors = authors_cit,
year = year_cit,
title = title_cit,
source_title = source_title_cit,
publisher = publisher_cit,
volume = volume_cit,
issue = issue_cit,
start_page = start_page_cit,
end_page = end_page_cit,
doi = doi_cit)
# generate RIS file
level1_ris_cit <- paste(paste0('\n',
'TY - ', maditr::vlookup(publication_type_cit, type_list, result_column = 'type', lookup_column = 'publication_type'), '\n',
'AU - ', authors_cit, '\n',
'TI - ', title_cit, '\n',
'PY - ', year_cit, '\n',
'AB - ', abstract_cit, '\n',
'SP - ', start_page_cit, '\n',
'EP - ', end_page_cit, '\n',
'JF - ', source_title_cit, '\n',
'VL - ', volume_cit, '\n',
'IS - ', issue_cit, '\n',
'PB - ', publisher_cit, '\n',
# 'SN - ', issn_cit, '\n',
'DO - ', doi_cit, '\n',
'ER - '),
collapse = '\n')
# generate ris build report
ris_records_cit <- lengths(regmatches(level1_ris_cit, gregexpr("TY - ", level1_ris_cit)))
stage1_report_cit <- paste0('Your ', scales::comma(input_number), ' articles were cited a total of ', scales::comma(all_citations), ' times. This corresponds to ',
scales::comma(length(citations_unique)), ' unique article IDs. Your RIS file is ready for download and contains ',
scales::comma(ris_records_cit), ' records exported from Lens.org.')
report_cit <- stage1_report_cit
return(list(display = level1_table_cit, ris = level1_ris_cit, report = report_cit, df = cit_results))
### search for references articles
} else if (get_records == 'references') {
if (is.null(record_list[["data"]][["references_count"]]) == TRUE || identical(record_list[["data"]][["references_count"]], 0) == TRUE){
return('Warning: Your input articles contained no references in the Lens.org database')
}
# obtain reference lists from article search
reference_count <- record_list[["data"]][["references_count"]]
reference_count[is.na(reference_count)] <- 0
all_references <- sum(reference_count, na.rm = TRUE)
ref_by_art <- record_list[["data"]][["references"]]
references <- unlist(record_list[["data"]][["references"]])
references_unique <- unique(references)
deduped_references <- length(references_unique)
ref_counts_df <- tibble::tibble(articles_id,
reference_count,
ref_by_art)
ref_counts_df[is.na(ref_counts_df)] <- 0
#below code not working because it was tagging excess records from one group onto the next (of 500)
#ref_counts_df <- dplyr::mutate(ref_counts_df,
# # cumulatively add up citation count
# cumulative_n = cumsum(reference_count),
# # divide by max citations allowed
# export_group = floor(cumsum(reference_count) / 500))
ref_counts_df <- dplyr::mutate(ref_counts_df, export_group = MESS::cumsumbinning(reference_count, 500))
# get a list of vectors of the ids
refgroups <- dplyr::group_split(ref_counts_df, export_group)
# define query function
run_request <- function(input){
# build query for article citations - this will pull back a maximum of 1,000 hits from lens.org, so not great if lots of refs
request <- paste0('{
"query": {
"terms": {
"lens_id": [', paste0('"', paste(input, collapse = '", "'), '"'), ']
}
},
"size":500,
"include": ["lens_id", "authors", "publication_type", "title", "external_ids", "start_page", "end_page", "volume", "issue", "references", "scholarly_citations", "source_urls", "abstract", "date_published", "year_published", "references_count", "scholarly_citations_count", "source"]
}')
#perform references search and extract text results
results <- getLENSData(token, request)
return(results)
}
# run the query function for each tibble, adding to a final dataset
ref_results <- data.frame()
tStart_ref <- Sys.time()
for (i in 1:length(refgroups)){
data_ref <- run_request(unique(unlist(refgroups[[i]]$ref_by_art)))
requests_remaining <- data_ref[["headers"]][["x-rate-limit-remaining-request-per-minute"]]
print(paste0('Remaining requests = ', requests_remaining))
record_json_ref <- httr::content(data_ref, "text")
record_list_ref <- jsonlite::fromJSON(record_json_ref)
# error messages
if (data_ref$status_code == 404){
return('Warning: Your search returned no matched results. Please double check your input article identifiers and try again.')
}
if (data_ref$status_code == 429){
return('Warning: Right now there are too many people using citationchaser. This has been logged and we will endeavour to increase the bandwith as soon as possible. Please try again later.')
}
if (record_list_ref$total == 0){
return('Warning: Your search returned no matched results. Please double check your input article identifiers and try again.')
}
record_list_ref_df <- as.data.frame(record_list_ref)
ref_results <- dplyr::bind_rows(ref_results, record_list_ref_df)
tEnd_ref <- Sys.time()
if (data_ref[["headers"]][["x-rate-limit-remaining-request-per-minute"]] < 1){
t_ref <- tEnd_ref - tStart_ref
Sys.sleep(60 - t) # pause to limit requests below 50 requests/min
}
}
ref_results <- ref_results[!duplicated(ref_results$data.lens_id),]
all_results_ref <- ref_results$data.lens_id
# convert json to ris style
type_list <- data.frame(type = c("ABST", "BOOK", "CHAP", "COMP", "CONF", "DATA", "JOUR"),
description = c("abstract reference", "whole book reference", "book chapter reference", "computer program", "conference proceeding", "data file", "journal/periodical reference"),
publication_type = c("reference entry", "book", "book chapter", "component", "conference proceedings", "dataset", "journal article"))
publication_type_ref <- ref_results$data.publication_type
authors_ref <- list()
for (i in 1:length(ref_results$data.authors)) {
authors_ref <- unlist(c(authors_ref, paste0(ref_results$data.authors[[i]]$last_name, ', ',
ref_results$data.authors[[i]]$first_name, collapse = '; ')))
}
title_ref <- ref_results$data.title
year_ref <- ref_results$data.year_published
abstract_ref <- ref_results$data.abstract
start_page_ref <- ref_results$data.start_page
end_page_ref <- ref_results$data.end_page
source_title_ref <- list()
for (i in 1:length(ref_results[,1])) {
source_title_ref <- unlist(c(source_title_ref, ref_results$data.source[[1]][i]))
}
volume_ref <- ref_results$data.volume
issue_ref <- ref_results$data.issue
publisher_ref <- ref_results$data.source.publisher
issn_ref <- ref_results$data.source.issn
doi_ref <- unlist(lapply(ref_results$data.external_ids, function(ch) maditr::vlookup('doi', ch, result_column = 'value', lookup_column = 'type')))
level1_table_ref <- data.table::data.table(authors = authors_ref,
year = year_ref,
title = title_ref,
source_title = source_title_ref,
publisher = publisher_ref,
volume = volume_ref,
issue = issue_ref,
start_page = start_page_ref,
end_page = end_page_ref,
doi = doi_ref)
level1_ris_ref <- paste(paste0('\n',
'TY - ', maditr::vlookup(publication_type_ref, type_list, result_column = 'type', lookup_column = 'publication_type'), '\n',
'AU - ', authors_ref, '\n',
'TI - ', title_ref, '\n',
'PY - ', year_ref, '\n',
'AB - ', abstract_ref, '\n',
'SP - ', start_page_ref, '\n',
'EP - ', end_page_ref, '\n',
'JF - ', source_title_ref, '\n',
'VL - ', volume_ref, '\n',
'IS - ', issue_ref, '\n',
'PB - ', publisher_ref, '\n',
# 'SN - ', issn_ref, '\n',
'DO - ', doi_ref, '\n',
'ER - '),
collapse = '\n')
# ris build report
ris_records_ref <- lengths(regmatches(level1_ris_ref, gregexpr("TY - ", level1_ris_ref)))
stage1_report_ref <- paste0('Your ', scales::comma(input_number), ' articles contained a total of ', scales::comma(all_references), ' references. This corresponds to ',
scales::comma(deduped_references), ' unique IDs. Your RIS file is ready for download and contains ', scales::comma(ris_records_ref), ' records exported from Lens.org.')
report_ref <- stage1_report_ref
return(list(display = level1_table_ref, ris = level1_ris_ref, report = report_ref, df = ref_results))
}
}
#' Function to query Lens.org
#'
#' @description Function written by lens.org for use of their API.
#' @param token An access key for the lens.org API. Tokens can be obtained by
#' applying for scholarly API access and creating a token once approved. See
#' 'https://www.lens.org/lens/user/subscriptions#scholar' for further details.
#' @param query A search string formulated according to the Lens.org API
#' documentation: 'https://docs.api.lens.org/request-scholar.html'.
#' @return A summary response. The results are viewable using
#' 'content(data, "text")'. Other details regarding the request (e.g. repsonse
#' times) can be accessed through the main output.
#' @importFrom httr add_headers POST
getLENSData <- function(token, query){
url <- 'https://api.lens.org/scholarly/search'
headers <- c('Authorization' = token, 'Content-Type' = 'application/json')
httr::POST(url = url, httr::add_headers(.headers=headers), body = query)
}
#' Find citation based on identifier
#'
#' @description This function takes a list of articles in the form of established
#' identifiers (e.g. digital object identifiers) and sends a request to the
#' lens.org API to obtain full citation information from the Lens database for
#' all sought articles.
#' @param article_list List of article identifiers for which the reference
#' lists will be returned. Must be a list/vector of identifiers, e.g.
#' '"10.1186/s13750-018-0126-2" "10.1002/jrsm.1378"'.
#' @param type Specification of the type of input provided. The default is
#' 'doi' (digital object identifier), but any of the following are accepted:
#' "pmid" (PubMed ID), "pmcid" (PubMed Central ID), "magid" (Microsoft
#' Academic ID), "coreid" (CORE identifier), lens_id" (The Lens.org ID),
#' "title" (article title; much lower specificity).
#' @param token An access key for the lens.org API. Tokens can be obtained by
#' applying for scholarly API access and creating a token once approved. See
#' 'https://www.lens.org/lens/user/subscriptions#scholar' for further details.
#' @return A dataframe containing the matching citation from Lens.org.
#' @importFrom maditr vlookup
#' @importFrom httr content
#' @importFrom jsonlite fromJSON
#' @importFrom data.table data.table
#' @examples
#' \dontrun{
#' article_list <- c("10.1007/978-3-642-37048-9_13", "10.1111/sum.12030",
#' "10.5194/bg-13-3619-2016", "10.1016/j.agee.2012.09.006")
#' results <- get_citation(article_list)
#' articles <- results$display
#' }
#' @export
get_citation <- function(article_list,
type = 'doi',
token = 'WCFlpCtuJXYI1sDhZcZ8y7hHpri0SEmTnLNkeU4OEM5JTQRNXB9w'){
if(length(article_list) == 1){
article_list <- trimws(unlist(strsplit(article_list, '[,]')))
}
request <- paste0('{\n\t"query": {\n\t\t"terms": {\n\t\t\t"', type, '": ["', paste0('', paste(article_list, collapse = '", "'), '"'),']\n\t\t}\n\t},\n\t"size":500\n}')
# perform article search and extract text results
data <- getLENSData(token, request)
requests_remaining <- data[["headers"]][["x-rate-limit-remaining-request-per-minute"]]
print(paste0('Remaining requests = ', requests_remaining))
record_json <- httr::content(data, "text")
# convert json output from article search to list
record_list <- jsonlite::fromJSON(record_json)
# error messages
if (data$status_code == 404){
return('Warning: Your search returned no matched results. Please double check your input article identifiers and try again.')
}
if (data$status_code == 429){
return('Warning: Right now there are too many people using citationchaser. This has been logged and we will endeavour to increase the bandwith as soon as possible. Please try again later.')
}
if (record_list$total == 0){
return('Warning: Your search returned no matched results. Please double check your input article identifiers and try again.')
}
inputs_df <- as.data.frame(record_list)
# citations and references
citation_count <- record_list[["data"]][["scholarly_citations_count"]]
reference_count <- record_list[["data"]][["references_count"]]
# convert json to ris style
type_list <- data.frame(type = c("ABST", "BOOK", "CHAP", "COMP", "CONF", "DATA", "JOUR"),
description = c("abstract reference", "whole book reference", "book chapter reference", "computer program", "conference proceeding", "data file", "journal/periodical reference"),
publication_type = c("reference entry", "book", "book chapter", "component", "conference proceedings", "dataset", "journal article"))
publication_type <- record_list[["data"]][["publication_type"]]
authors <- list()
for (i in 1:length(record_list[["data"]][["authors"]])) {
authors <- unlist(c(authors, paste0(record_list[["data"]][["authors"]][[i]]$last_name, ', ',
record_list[["data"]][["authors"]][[i]]$first_name, collapse = '; ')))
}
title <- record_list[["data"]][["title"]]
year <- record_list[["data"]][["year_published"]]
abstract <- record_list[["data"]][["abstract"]]
start_page <- record_list[["data"]][["start_page"]]
end_page <- record_list[["data"]][["end_page"]]
source_title <- record_list[["data"]][["source"]][["title"]]
volume <- record_list[["data"]][["volume"]]
issue <- record_list[["data"]][["issue"]]
publisher <- record_list[["data"]][["source"]][["publisher"]]
issn <- record_list[["data"]][["source"]][["issn"]]
doi <- unlist(lapply(record_list[["data"]][["external_ids"]], function(ch) maditr::vlookup('doi', ch, result_column = 'value', lookup_column = 'type')))
article_table <- data.table::data.table(authors = authors,
year = year,
title = title,
source_title = source_title,
publisher = publisher,
volume = volume,
issue = issue,
start_page = start_page,
end_page = end_page,
doi = doi,
Lens_refs = reference_count,
Lens_cited = citation_count)
article_ris <- paste(paste0('\n',
'TY - ', maditr::vlookup(publication_type, type_list, result_column = 'type', lookup_column = 'publication_type'), '\n',
'AU - ', authors, '\n',
'TI - ', title, '\n',
'PY - ', year, '\n',
'AB - ', abstract, '\n',
'SP - ', start_page, '\n',
'EP - ', end_page, '\n',
'JF - ', source_title, '\n',
'VL - ', volume, '\n',
'IS - ', issue, '\n',
'PB - ', publisher, '\n',
# 'SN - ', issn, '\n',
'DO - ', doi, '\n',
'ER - '),
collapse = '\n')
return(list(display = article_table, ris = article_ris, df = inputs_df))
}
Try the citationchaser package in your browser
Any scripts or data that you put into this service are public.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Embedding an R snippet on your website
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.