#' Global wrapper function for generating, saving and scraping info
#'
#' @description Function wraps the `buildGSlinks()`, `save_htmls()`, and scrape functions.
#' @param and_terms Vector of alphanumeric terms searched using the AND Boolean operator,
#' specified by Google Scholar as 'with all of the words'.
#' @param exact_phrase Vector of alphanumeric terms enclosed in inverted commas and searched
#' as phrases (e.g. "large cat"), specified by Google Scholar as 'with the exact phrase'.
#' @param or_terms Vector of alphanumeric terms searched using the OR Boolean operator,
#' specified by Google Scholar as 'with at least one of the words'.
#' @param not_terms Vector of alphanumeric terms searched using the NOT Boolean operator,
#' specified by Google Scholar as 'without the words'.
#' @param language Two-letter language code for search language. The default is 'en' (English).
#' @param year_from Integer full numeric year (e.g. 2000) from which searching will be performed
#' (inclusive). If no value provided, all years are searched.
#' @param year_to Integer full numeric year (e.g. 2020) to which searching will be performed
#' (inclusive). If no value provided, all years are searched.
#' @param start_page Integer specifying which page(s) of search results should be displayed.
#' If multiple pages are selected, multiple URLs are returned, one for each page of ten
#' search results. The default is set to generate a list of 100 URLs (maximum set of Google
#' Scholar results visible).
#' @param pages Integer for the number of pages of search results to be returned (one link per
#' page). A maximum of 100 pages can be displayed in Google Scholar. The default value is 1.
#' @param incl_cit Logical argument (TRUE or FALSE) specifying whether citations should be
#' included in the search
#' @param incl_pat Logical argument (TRUE or FALSE) specifying whether patents should be
#' included in the search
#' @param titlesearch Logical argument (TRUE or FALSE) specifying whether the search should
#' be performed on article titles only or anywhere in the record. The default is FALSE.
#' @param authors The names of authors searched for.
#' @param source The name of the source of the articles (e.g. academic journal).
#' @param pause Integer specifying the number of seconds to wait between download attempts.
#' The default value is 4 seconds.
#' @param backoff A logical argument (TRUE or FALSE) specifying whether responsive backing-off
#' should be used. If set to TRUE, the time between calls is varied depending on how long the
#' server takes to respond to the original request. The responsive back-off time is set to
#' multiple the response time by the `pause` time: i.e. if the system takes 1.02 seconds to
#' respond and `pause` time is set to 4 seconds, a 4.10 second delay will be employed before
#' the next call. The default for back-off is `FALSE`.
#' @return A list containing: 1) (data) a data frame containing all information that can be
#' extracted from all html files in the working directory; and, 2) (report) a report of the
#' links generated and the input variables used.
#' @examples
#' \dontrun{
#' and_terms <- c('river', 'aquatic')
#' exact_phrase <- c('water chemistry')
#' or_terms <- c('crayfish', 'fish')
#' not_terms <- c('lobster', 'coral')
#' year_from <- 1900
#' year_to <- 2020
#' info <- save_and_scrapeGS(and_terms, exact_phrase, or_terms, not_terms, pages = 3)
#' info
#' }
#'@export
save_and_scrapeGS <- (and_terms = '',
exact_phrase = '',
or_terms = '',
not_terms = '',
= 'en',
year_from = '',
year_to = '',
start_page = 1,
pages = 1,
incl_cit = ,
incl_pat = ,
titlesearch = ,
authors = '',
= '',
pause = 4,
backoff = ){
links <- (and_terms = and_terms,
exact_phrase = exact_phrase,
or_terms = or_terms,
not_terms = not_terms,
= ,
year_from = year_from,
year_to = year_to,
start_page = start_page,
pages = pages,
incl_cit = incl_cit,
incl_pat = incl_pat,
titlesearch = titlesearch,
authors = authors,
= )
htmls <- (urls = links,
pause = pause,
backoff = backoff)
<- ()
(i 1:(htmls)){
info <- ((htmls[i]))
<- info$
<- dplyr::bind_rows(, )
}
<-
report <- ('File generated: ',
('Search date, time, timezone: ',
(),
' (',
(),
')',
sep = ''),
'\n',
'Search parameters:',
('All these words: ',
(and_terms,
collapse = '; '),
sep = ''),
('This exact word or phrase: ',
('"',
(exact_phrase),
'"',
sep = ''),
sep = ''),
('Any these words: ',
(or_terms,
collapse = '; '),
sep = ''),
('None of these words: ',
(not_terms,
collapse = '; '),
sep = ''),
( == ''){
'Language: Any'
} {
('Language: ',
,
sep = '')
},
('Between and including these years:',
(year_from == ''){
'all'
} {
year_from
},
'and',
(year_to == ''){
'now'
} {
year_to
},
sep = ' '),
(pages == ''){
'Number of pages exported: 1'
} {
('Number of pages exported: ',
pages,
sep = '')
},
(start_page == ''){
'Starting from page: 1'
} {
('Starting from page: ',
start_page,
sep = '')
},
(incl_cit == ''){
'Citations included: TRUE'
} {
('Citations included: ',
incl_cit,
sep = '')
},
(incl_pat == ''){
'Patents included: TRUE'
} {
('Citations included: ',
incl_pat,
sep = '')
},
(titlesearch == ''){
'Search only in the title: FALSE'
} {
('Search only in the title: ',
titlesearch,
sep = '')
},
('Authors:',
authors,
sep = ' '),
('Source:',
,
sep = ' '),
(pause == ''){
'Paused 4 seconds between calls.'
} {
('Paused ',
pause,
' seconds between calls',
sep = '')
},
(backoff == ''){
'Sensitive backing off not employed (see help files for more information).'
} {
'Sensitive backing off employed (see help files for more information).'
},
('Search date, time, timezone: ',
(),
' (',
(),
')',
sep = ''),
'Google Scholar search pages exported:',
(links,
collapse = '\n'),
'\n',
sep = '\n')
output <- ( = ,
report = report)
(output)
}
