Convert between bibliography formats — bibConvert • rbibutils

Read a bibliography file in one of the supported formats, convert it to nnother format, and write it to a file.

Usage

bibConvert(infile, outfile, informat, outformat, ..., tex, encoding, 
           options)

Arguments

infile: input file, a character string.
outfile: output file, a character string.
informat: input format, a character string, see sections “Supported formats” and “Details”.
outformat: output format, a character string, see sections “Supported formats” and “Details”.
...: not used.
tex: TeX specific options, see Details, a character vector.
encoding: character(2), a length two vector specifying input and output encodings. Default to both is "utf8", see Details.
options: mainly for debugging: additional options for the converters, see Details.

Details

Arguments informat and outformat can usually be omitted, since bibConvert infers them from the extensions of the names of the input and output files, see section "File extensions" below. However, there is ambiguity for the extension "bib", since it is used for Bibtex and BibLaTeX entries. For this extension, the default for both, informat and outformat, is "bibtex".

Package rbibutils supports format "bibentry", in addition to the formats supported by the bibutils library. A bibentry object contains one or more references. Two formats are supported for "bibentry" for both input and output. A bibentry object previously saved to a file using saveRDS (default extension "rds") or an R source file containing one or more bibentry commands. The "rds" file is just read in and should contain a bibentry object.

When bibconvert outputs to an R source file, two variants are supported: "R" and "Rstyle". When (outformat = "R", there is one bibentry call for each reference, just as in a Bibtex file, each reference is a single entry. outformat = "Rstyle" uses the format of print(be, style = "R"), i.e., the bibentry calls are output as a comma separated sequence wrapped in c(). For input, it is not necessary to specify which variant is used.

Note that when the input format and output formats are identical, the conversion is not necessarilly a null operation (except for xml, and even that may change). For example, depending on the arguments the character encoding may change. Also, input BibTeX files may contain additional instructions, such as journal abbreviations, which are expanded and incorporated in the references but not exported. It should be remembered also that there may be loss of information when converting from one format to another.

For a complete list of supported bibliography formats, see section “Supported formats” below. The documentation of the original bibutils library (Putnam 2020) gives further details.

Argument encoding is a character vector containing 2 elements, specifying the encoding of the input and output files. If the encodings are the same, a length one vector can be supplied. The default encodings are UTF-8 for input and output. A large number of familiar encodings are supported, e.g. "latin1" and "cp1251" (Windows Cyrillic). Some encodings have two or more aliases and they are also accepted. If an unknown encoding is requested, a list of all supported encodings will be printed.

Argument tex is an unnamed character vector containing switches for bibtex input and output (mostly output). Currently, the following are available:

uppercase: write bibtex tags/types in upper case.
no_latex: do not convert latex-style character combinations to letters.
brackets: use brackets, not quotation marks surrounding data.
dash: use one dash "-", not two "--", in page ranges.
fc: add final comma to bibtex output.

By default latex encodings for accented characters are converted to letters. This may be a problem if the output encoding is not UTF-8, since some characters created by this process may be invalid in that encoding. For example, a BibTeX file which otherwise contains only cyriilic and latin characters may have a few entries with authors containing latin accented characters represented using the TeX convention. If those characters are not converted to Unicode letters, they can be exported to "cp1251" (Windows Cyrillic) for example. Specifying the option no_latex should solve the problem in such cases.

Argument options is mostly for debugging and mimics the command line options of the bibutils' binaries. The argument is a named character vector and is supplied as c(tag1= val1, tag2 = val2, ...), where each tag is the name of an option and the value is the corresponding value. The value for options that do not require one is ignored and can be set to "". Some of the available options are:

h: help, show all available options.
nb: do not write Byte Order Mark in UTF8 output.
verbose: print intermediate output.
debug: print even more intermediate output.

Supported formats

If an input or output format is not specified by arguments, it is inferred, if possible, from the file extension.

In the table below column Abbreviation shows the abbreviation for arguments informat and outformat, column FileExt gives the default file extension for that format, column Input (Output) contains TRUE if the format is supported for input (output) and FALSE otherwise. Column Description gives basic description of the format.

Abbreviation	FileExt	Input	Output	Description
ads	ads	FALSE	TRUE	ADS reference format
bib	bib	TRUE	TRUE	BibTeX
bibtex	bibtex	TRUE	TRUE	BibTeX
biblatex	biblatex	TRUE	TRUE	BibLaTeX
copac	copac	TRUE	FALSE	COPAC format references
ebi	ebi	TRUE	FALSE	EBI XML
end	end	TRUE	TRUE	EndNote (Refer format)
endx	endx	TRUE	FALSE	EndNote XML
isi	isi	TRUE	TRUE	ISI web of science
med	med	TRUE	FALSE	Pubmed XML references
nbib	nbib	TRUE	TRUE	Pubmed/National Library of Medicine nbib format
ris	ris	TRUE	TRUE	RIS format
R, r, Rstyle	R	TRUE	TRUE	R source file containing bibentry commands
rds	rds	TRUE	TRUE	bibentry object in a binary file created by saveRDS()
xml	xml	TRUE	TRUE	MODS XML intermediate
wordbib	wordbib	TRUE	TRUE	Word 2007 bibliography format

The file "easyPubMedvig.xml" used in the examples for Pubmed XML ("med") was obtained using code from the vignette in package easyPubMed (Fantini 2019).

Value

The function is used for the side effect of creating a file in the requested format. It returns a list, currently containing the following components:

infile: name of the input file,
outfile: name of the output file,
nref_in: number of references read from the input file,
nref_out: number of references written to the output file.

Normally, nref_in and nref_out are the same. If some references were imported successfully but failed on export,

nref_out may be smaller than nref_in. In such cases informative messages are printed during processing. (If this happens silently, it is probably a bug and please create an issue on Github.)

Author

Georgi N. Boshnakov

References

Damiano Fantini (2019). “easyPubMed: Search and Retrieve Scientific Publication Records from PubMed.” R package version 2.13, https://CRAN.R-project.org/package=easyPubMed.

Chris Putnam (2020). “Library bibutils, version 6.10.” https://sourceforge.net/projects/bibutils/.

Examples

fn_biblatex <- system.file("bib", "ex0.biblatex",  package = "rbibutils")
fn_biblatex
#> [1] "/home/runner/work/_temp/Library/rbibutils/bib/ex0.biblatex"
## file.show(fn_biblatex)

## convert a biblatex file to xml
modl <- tempfile(fileext = ".xml")
bibConvert(infile = fn_biblatex, outfile = modl, informat = "biblatex", outformat = "xml")
#> $infile
#> [1] "/home/runner/work/_temp/Library/rbibutils/bib/ex0.biblatex"
#> 
#> $outfile
#> [1] "/tmp/RtmpvaOur9/file1b13577ea6e.xml"
#> 
#> $nref_in
#> [1] 3
#> 
#> $nref_out
#> [1] 3
#> 
## file.show(modl)

## convert a biblatex file to bibtex
bib <- tempfile(fileext = ".bib")
bib2 <- tempfile(fileext = ".bib")
bibConvert(infile = fn_biblatex, outfile = bib, informat = "biblatex", outformat = "bib")
#> $infile
#> [1] "/home/runner/work/_temp/Library/rbibutils/bib/ex0.biblatex"
#> 
#> $outfile
#> [1] "/tmp/RtmpvaOur9/file1b138c6cbdd.bib"
#> 
#> $nref_in
#> [1] 3
#> 
#> $nref_out
#> [1] 3
#> 
## file.show(bib)

## convert a biblatex file to bibentry
rds <- tempfile(fileext = ".rds")
fn_biblatex
#> [1] "/home/runner/work/_temp/Library/rbibutils/bib/ex0.biblatex"
rds
#> [1] "/tmp/RtmpvaOur9/file1b13206d4ac4.rds"
be <- bibConvert(fn_biblatex, rds, "biblatex", "bibentry")
bea <- bibConvert(fn_biblatex, rds, "biblatex") # same
readRDS(rds)
#> Beach CM, Davidson R (1983). “Distribution-free statistical inference
#> with Lorenz curves and income shares.” _Review of Economic Studies_,
#> *50*, 723-735.
#> 
#> Beach CM, Kaliski SF (1986). “Lorenz curve inference with sample
#> weights: an application to the distribution of unemployment
#> experience.” _Appl. Statist._, *35*, 38-45.
#> 
#> Lambert PJ (1993). _The distribution and redistribution of income. A
#> mathematical analysis_, 2 edition. Manchester University Press,
#> Manchester.

## convert to R source file
r <- tempfile(fileext = ".R")
bibConvert(fn_biblatex, r, "biblatex")
#> $infile
#> [1] "/home/runner/work/_temp/Library/rbibutils/bib/ex0.biblatex"
#> 
#> $outfile
#> [1] "/tmp/RtmpvaOur9/file1b1355663d7b.R"
#> 
#> $nref_in
#> [1] 3
#> 
#> $nref_out
#> [1] 3
#> 
#> $bib
#> Beach CM, Davidson R (1983). “Distribution-free statistical inference
#> with Lorenz curves and income shares.” _Review of Economic Studies_,
#> *50*, 723-735.
#> 
#> Beach CM, Kaliski SF (1986). “Lorenz curve inference with sample
#> weights: an application to the distribution of unemployment
#> experience.” _Appl. Statist._, *35*, 38-45.
#> 
#> Lambert PJ (1993). _The distribution and redistribution of income. A
#> mathematical analysis_, 2 edition. Manchester University Press,
#> Manchester.
#> 
## file.show(r)
cat(readLines(r), sep = "\n")
#> bibentry(bibtype = "Article",
#>          key = "beachanddavidson1983",
#>          author = c(person(given = c("Charles", "M"),
#>                            family = "Beach"),
#>                     person(given = "Russell",
#>                            family = "Davidson")),
#>          title = "Distribution-free statistical inference with Lorenz curves and income shares",
#>          journal = "Review of Economic Studies",
#>          year = "1983",
#>          month = "Feb",
#>          volume = "50",
#>          pages = "723--735")
#> 
#> bibentry(bibtype = "Article",
#>          key = "beachandkaliski1986a",
#>          author = c(person(given = c("C", "M"),
#>                            family = "Beach"),
#>                     person(given = c("S", "F"),
#>                            family = "Kaliski")),
#>          title = "Lorenz curve inference with sample weights: an application to the distribution of unemployment experience",
#>          journal = "Appl. Statist.",
#>          year = "1986",
#>          volume = "35",
#>          pages = "38--45")
#> 
#> bibentry(bibtype = "Book",
#>          key = "lambert1993",
#>          author = person(given = c("Peter", "J"),
#>                          family = "Lambert"),
#>          title = "The distribution and redistribution of income. A mathematical analysis",
#>          year = "1993",
#>          edition = "2",
#>          publisher = "Manchester University Press",
#>          address = "Manchester")
#> 

fn_cyr_utf8 <- system.file("bib", "cyr_utf8.bib",  package = "rbibutils")

## Can't have files with different encodings in the package, so below
## first convert a UTF-8 file to something else.
##
## input here contains cyrillic (UTF-8) output to Windows Cyrillic,
## notice the "no_latex" option
a <- bibConvert(fn_cyr_utf8, bib, encoding = c("utf8", "cp1251"), tex = "no_latex")

## now take the bib file and convert it to UTF-8
bibConvert(bib, bib2, encoding = c("cp1251", "utf8"))
#> $infile
#> [1] "/tmp/RtmpvaOur9/file1b138c6cbdd.bib"
#> 
#> $outfile
#> [1] "/tmp/RtmpvaOur9/file1b1325899553.bib"
#> 
#> $nref_in
#> [1] 2
#> 
#> $nref_out
#> [1] 2
#> 

# \donttest{
## Latin-1 example: Author and Title fileds contain Latin-1 accented
##   characters, not real names. As above, the file is in UTF-8
fn_latin1_utf8  <- system.file("bib", "latin1accents_utf8.bib", package = "rbibutils")
## convert to Latin-1, by default the accents are converted to TeX combinations:
b <- bibConvert(fn_latin1_utf8, bib , encoding = c("utf8", "latin1"))
cat(readLines(bib), sep = "\n")
#> @Article{test1,
#> author="L{\^e}{\'a}{\'e}, K{\`a}r{\`e}",
#> title="K{\`a}r{\`e} {\"u}{\"a} L{\^e}{\'a}{\'e}",
#> journal="Latin-1 accents",
#> year="2020",
#> note="Having the similar title and author is helpful for debugging since authors are processed differently from other fields. Here is a formula: {\$}{\backslash}tan($\alpha$){\$}, {\$}{\backslash}sin($\delta$)/$\delta$ $\to$ 1{\$}, {\$}{\backslash}sqrt{\{}x{\}}{\$}."
#> }
#> 
## use "no_latex" option to keep them Latin1:
c <- bibConvert(fn_latin1_utf8, bib , encoding = c("utf8", "latin1"), tex = "no_latex")
## this will show properly in Latin-1 locale (or suitable text editor):
##cat(readLines(bib), sep = "\n")

## gb18030 example (Chinese)
##
## prepare some filenames for the examples below:
xeCJK_utf8    <- system.file("bib/xeCJK_utf8.bib", package = "rbibutils")
xeCJK_gb18030 <- system.file("bib/xeCJK_gb18030.bib", package = "rbibutils")
fn_gb18030 <- tempfile(fileext = ".bib")
fn_rds <- tempfile(fileext = ".rds")

## input bib file utf8, output bib file gb18030:
bibConvert(xeCJK_utf8, fn_gb18030, encoding = c("utf8", "gb18030"))
#> $infile
#> [1] "/home/runner/work/_temp/Library/rbibutils/bib/xeCJK_utf8.bib"
#> 
#> $outfile
#> [1] "/tmp/RtmpvaOur9/file1b136186eea6.bib"
#> 
#> $nref_in
#> [1] 1
#> 
#> $nref_out
#> [1] 1
#> 

## input bib file utf8, output file rds (and the rds object is returned
bibConvert(xeCJK_utf8, fn_rds)
#> $infile
#> [1] "/home/runner/work/_temp/Library/rbibutils/bib/xeCJK_utf8.bib"
#> 
#> $outfile
#> [1] "/tmp/RtmpvaOur9/file1b135f6c6d11.rds"
#> 
#> $nref_in
#> [1] 1
#> 
#> $nref_out
#> [1] 1
#> 
#> $bib
#> 陈骁, 黄声华, 万山明, 庞珽 (2012).
#> “基于电无级变速器的内燃机最优控制策略及整车能量管理.” _电工技术学报_,
#> *27*(2), 133-138.
#> 
unlink(fn_gb18030)
unlink(fn_rds)

## a Pubmed file
fn_med <- system.file("bib/easyPubMedvig.xml", package = "rbibutils")
## convert a Pubmed file to bibtex:
bibConvert(fn_med, bib, informat = "med")
#> $infile
#> [1] "/home/runner/work/_temp/Library/rbibutils/bib/easyPubMedvig.xml"
#> 
#> $outfile
#> [1] "/tmp/RtmpvaOur9/file1b138c6cbdd.bib"
#> 
#> $nref_in
#> [1] 6
#> 
#> $nref_out
#> [1] 6
#> 
## convert a Pubmed file to rds and import:
#bibConvert(fn_med, rds, informat = "med")
# }

unlink(c(modl, bib, bib2, r, rds))