BiblioJSON

A JSON-based bibliography format with importers and exporters for BibLaTeX, BibTeX, CSL-JSON, RIS, ENW, EndNote XML, Citavi, DOCX citations, and ODT citations.

Note: This package was previously published to npm as biblatex-csl-converter. The project has been renamed to bibliojson; please update your dependencies.

Try the demo here.

Supported Conversions

Main formats

From	To
BibLaTeX	BiblioJSON
CSL	BiblioJSON
BiblioJSON	CSL
BiblioJSON	BibLaTeX

Additional import formats

From	To
EndNote	BiblioJSON
RIS	BiblioJSON
Citavi	BiblioJSON
DOCX native format	BiblioJSON
ODT native format	BiblioJSON

We can even read the citation information from various citation managers inside ODT/DOCX files!

The BiblioJSON format

BiblioJSON is the JSON-based bibliography format at the heart of this library. It is designed to preserve as much information as possible from source formats — especially BibLaTeX — so that data can be exported again to either BibLaTeX or CSL-JSON without loss.

Why BiblioJSON instead of CSL-JSON?

CSL-JSON cannot represent everything that BibLaTeX can express. If we used CSL-JSON as the internal representation, details such as BibLaTeX-specific fields, name particles, or date granularity would be lost on round-trip. BiblioJSON keeps that information structured and accessible, so converting BibLaTeX → BiblioJSON → CSL, or CSL → BiblioJSON → BibLaTeX, stays as faithful as possible.

Top-level structure: BibDB

A parsed bibliography is returned as a BibDB — a plain object whose keys are numeric entry IDs and whose values are EntryObjects:

{
  "1": { /* EntryObject */ },
  "2": { /* EntryObject */ }
}

EntryObject

Each entry has the following shape:

Key	Type	Description
entry_key	string	The citation key (e.g. "doe2023").
bib_type	string	The BiblioJSON entry type (e.g. "article-journal", "book").
fields	object	The main, normalized fields for the entry.
unexpected_fields	object	Fields that are valid BibLaTeX fields but not expected for this entry type (only present when processUnexpected is enabled).
unknown_fields	object	Fields that are not part of the BibLaTeX spec (only present when processUnknown is enabled).
raw_fields	object	The almost-raw, pre-processed field contents from the source file.
location	object	Start/end offsets of the entry in the source text (only when includeLocation is enabled).
raw_text	string	The raw source text of the entry (only when includeRawText is enabled).
incomplete	boolean	Set when the parser could not fully parse the entry.

Field value types

Values inside fields are stored according to their semantic type so that exporters can render them correctly:

BiblioJSON type	Used for	Example JSON value
f_literal / f_long_literal	Plain text fields such as abstract or note.	[{"type":"text","text":"..."}]
f_title	Title fields (title, booktitle, journaltitle, …). Titles keep capitalization and markup.	[{"type":"text","text":"The Title"}]
f_name	Names (author, editor, translator, …).	[{"family":[{"type":"text","text":"Doe"}],"given":[{"type":"text","text":"Jane"}]}]
f_date	Dates (date, origdate, urldate, …) as EDTF strings.	"2023-05"
f_key	Controlled vocabulary (langid, pagination, type, …).	"english"
f_integer	Integer values (edition, volume, …).	3
f_verbatim	Values that should not be interpreted (doi, isbn, …).	"10.1000/abc"
f_uri	URI values (url, eprint, …).	"https://example.org"

Rich text is represented as a NodeArray — an array of nodes. A plain text node looks like {"type":"text","text":"..."}. Nodes can also carry marks for formatting such as bold, italic, small-caps, superscript, or subscript, allowing markup to survive round-trips.

Minimal example

{
  "1": {
    "entry_key": "sample1",
    "bib_type": "article-journal",
    "fields": {
      "title": [{"type": "text", "text": "Sample title"}],
      "author": [
        {
          "family": [{"type": "text", "text": "Doe"}],
          "given": [{"type": "text", "text": "Jane"}]
        }
      ],
      "date": "2023",
      "journaltitle": [{"type": "text", "text": "Journal Name"}]
    }
  }
}

For the complete list of entry types and fields, see the TypeScript definitions in src/const.ts or the API documentation in API.md.

Usage

import {BibLatexParser} from "bibliojson"

// synchronous:
let parser = new BibLatexParser(input, {processUnexpected: true, processUnknown: true})
let bib = parser.parse()

// asynchronous:
let parser = new BibLatexParser(input, {processUnexpected: true, processUnknown: true})
parser.parseAsync().then((bib) => { ... })

API Documentation

For comprehensive API documentation, see API.md.

Extracting Citations from ODT and DOCX Files

See CITATIONS_IN_DOCS.md for documentation on how Zotero, Mendeley, EndNote, Citavi, and JabRef store citations and bibliographies inside ODT and DOCX files.

Note on quality of conversions

While writing this library, we have had a lot of experience converting between CSL and BibLaTeX and have also worked with the exports and imports generated by Zotero. This is open source/free software, making it easy to verify whether conversions are correct. Conversions involving these formats are therefore generally more reliable.

Other software, such as EndNote, Citavi, and Mendeley, is proprietary and closed source, making testing more difficult—especially without licenses for these products. We rely on publicly available documentation and ODT/DOCX test files produced by others to understand how these formats work and to test our conversions. If something is not converted correctly, please report it so we can improve the library; example files are welcome.

FAQ

---

Q: Why do you use BiblioJSON instead of just the JSON format of CSL? Wouldn't that save you one conversion step?

A: Unfortunately, the CSL JSON format cannot hold all the information we import from BibLaTeX. If we used CSL JSON internally, we would lose information that we may want to export to BibLaTeX later on. BiblioJSON is designed to keep that information structured and lossless.

---

Q: Do you import all information from the imported BibTeX/BibLaTeX files?

A: We only keep the information found in any of the required or optional fields defined in the BibLaTeX documentation. Other fields are removed upon import.

---

Q: How do I see if there are errors when parsing the BibTeX/BibLaTeX file?

A: An array of errors encountered while parsing the file can be found at parser.errors after parsing. There is also parser.warnings for less serious issues.

---

Q: I need access to the raw/non-processed contents of certain fields. What do I do?

A: The fields in their almost raw form can be found under entry.raw_fields[FIELD_NAME].

---

Q: What if I need to process fields that don't follow the BibLaTeX definition?

A: You can initialize the parser with a config object like this: new BibLatexParser(inputString, {processUnexpected: true, processUnknown: {collaborator: 'l_name'}}). The processUnexpected setting enables parsing of fields that are known but shouldn't appear in a bibliography entry due to its type. The processUnknown setting allows parsing of entirely unknown fields. You can set it to true, or to an object containing descriptions of how these unknown fields should be processed. If a field is not specified, it will be processed as a literal field (f_literal). These fields will be available under entry.unexpected_fields[FIELD_NAME] and entry.unknown_fields[FIELD_NAME], respectively.

---

Q: I use variables in my BibLaTeX files. Will your converter read them?

A: Yes, but for the converter to create a string, the variables need to be defined. Undefined variables can also be handled by the BibLaTeX importer/exporter, but when exporting to CSL, they simply print out the variable name inside an HTML tag that is not supported by citeproc (and an error is thrown).

---

Q: I want to run the demo locally.

A: http-server is handy. Do a global install of http-server with npm install http-server -g and run http-server docs.

---

Q: I want to include this on my website, and I don't use npm packages, etc. Is there a file I can just add to the header of my webpage?

A: Yes, you can download such a file here.

---

Upgrading

From 2.x to 3.x

Note that the output getter has been removed. Use parse() instead.

This applies to BibLatexExporter, CSLExporter, and BibLatexParser. Note that the output of BibLatexParser is structured differently when using the parse() function if you previously used the output getter.

In the case of BibLatexExporter and CSLExporter, instead of:

const output = parser.output

Do:

const output = parser.parse()

In the case of BibLatexParser, instead of:

const output = parser.output

Do:

const parsed = parser.parse()
const output = parsed.entries

From 1.x to 2.x

Note that the API for the asynchronous parser has changed.

You need to change instances of this:

let parser = new BibLatexParser(input, {processUnexpected: true, processUnknown: true, async: true})
parser.parse().then((bib) => { ... })

to:

let parser = new BibLatexParser(input, {processUnexpected: true, processUnknown: true})
parser.parseAsync().then((bib) => { ... })