tree-sitter-biber

Tree sitter parser for .bib files using biber syntax. This is subtly different to the original bibtex syntax, but largely similar.

The main differences are how comments are handled. For bibtex, anything outside an @ entry or command is a proper comment. In biber, this is also a comment but is considered "junk" (and warns you appropriately in the log file). Instead, biber requires % to consider a line a proper comment.

Similarly, the comment command is also treated differently. In bibtex, the parsing of the command stops immediately after @comment. So @comment{this is a comment} is equivalent to @comment this is a comment, because neither sentences are considered inside a command or entry. In contrast, biber will grab the entire body, and throw an error if it doesn't exist. I.e., @comment this is an error will throw because biber is expecting an opening delimiter (brace or parenthesis).

Here's a scenario where this may be relevant in real life:

@comment{email me at [email protected]}

biber will be fine, but bibtex will throw an error (because the entry domain has no valid opening delimiter).

And, to be fair to bibtex, this will break biber but not bibtex:

@comment{ { }

More can be found here https://tex.stackexchange.com/questions/261261/are-comments-discouraged-in-a-bibtex-file/262282

Importantly, % is always a comment outside of a string in biber. So another real life difference is

@ARTICLE{auchunbekannt,
  title = {Beispielaufsatz},
  journal = {Zeitschrift},
  year = {2001},
  %volume = {7},
  pages = {1--35, 99--291},
  annotation = {lorem},
}

where bibtex will throw an error, but biber just considers the volume line to be a comment.

Similarly,

@BOOK(Car%ey,
,
  AUTHOR="G. V. Carey",
  TITLE="Mind the Stop: A Brief Guide to Punctuation", PUBLISHER="Penguin",
  YEAR="1958"
)

The extra comma is necessary for biber, and the key name is Car. For bibtex, the extra comma throws a "missing a field" error, but the key is completely valid (though good luck referencing it in the document).

Note that special TeX characters inside of entries, besides ones biber (and bibtex) actually looks for, are effectively ignored. So it is not really possible to accurately parse the contents of a string, nor desirable for a faithful parser. If you have an unescaped % in the string, it will break when the citation is used in a document (but is completely valid biber syntax).

This answer also does a good job of explaining what the actual terms bibtex, biber, natbib, and biblatex actually refer to. https://tex.stackexchange.com/questions/25701/bibtex-vs-biber-and-biblatex-vs-natbib