This vignette provides some guidance to retrieve the information from the EndNote database and comparing it with the files created during the query process. The main issue is one of formatting and platform inter-operability, with EndNote formatting part of the file names and titles to its own liking. Here, retrieve the EndNote database is useful as it holds the location of the .pdf files for each of the articles that have been retrieve. It also holds the various URLs corresponding to the article. Because of text formating issues, EndNote data has to be exported as .xml and as .txt. Both files have the same ordering, but the .xml is encoded as UTF-8 whereas the .txt is encoded as Windows 1252 to allow comparison with the .csv coming from the query process. In the following, the .txt is used to align the EndNote .xml database with query .csv files.

root_dir <- get_rootdir()
out_dir <- "exploitation/out/run79"
out_dir <- file.path(root_dir, out_dir)

Data loading

language_dfs <- get_language_dfs(languages)
meta_df <- get_meta_df(language_dfs)

Aligning EndNote .xml database with query .csv database

for (language in languages){
    xmldf <- get_endnote_xml(language)
    endnote_titles <- get_endnote_titles(language)

    endnote_titles <- make_pretty_str(endnote_titles)
    query_titles <- make_pretty_str(language_dfs[[language]]$Title)

    stopifnot(all(query_titles %in% endnote_titles) & all(endnote_titles %in% query_titles))

    language_dfs[[language]] <- cbind(language_dfs[[language]][match(endnote_titles, query_titles), ], xmldf)

    stopifnot(names(table(language_dfs[[language]]$`electronic-resource-num` == language_dfs[[language]]$DOI)) == "TRUE")

    collected <- ifelse(is.na(language_dfs[[language]]$pdfs), "absent from corpus", "in corpus")
save(language_dfs, file = file.path(root.dir, out.dir, paste0("language_dfs", ".rda")))