This is yet-another-bibtex-to-RDF converter, designed to perform a very low-level transformation of BibTex to RDF. In particular, tex macros and syntax in the field values are preserved, rather than being converted to some other representation (e.g. Unicode).
The advantage of this is that the resulting RDF data maintains enough of the original file to reconstruct a version that will produce identical output when run through bibtex. This is done by expanding all the string definitions and cross-references in the source file as completely as possible. Thus, these aspects of original file, as well as syntactic variations like capitalization of fields, are lost in the transformation.
The disadvantage is that some applications might prefer a higher-level transformation that removed the tex syntax. The low-level translation was chosen because it is fairly easy to remove tex syntax, but essentially impossible to restore it outside of a limited set of transformations.
How do I get the source code?
You need a Subversion client. Type svn in your shell and see if you have one already installed. If not, go to the Subversion web site and get one.
Then type
svn co http://simile.mit.edu/repository/RDFizers/bibtex2rdf/
the source code will be fetched and downloaded to the ./bibtex2rdf directory.
How do I build it?
This RDFizer requires two things for you to build it:
- a Java Virtual Machine installed on your machine (version 1.4 or greater). Type 'java -version' at your shell prompt to know what version you have. If you don't have it, go to http://www.java.com and download it.
- Apache Ant installed. Type 'ant -version' at your shell prompt to know what version you have. If ant is not installed, go to http://ant.apache.org/ and download it. Don't panic, the installation is really fast and simple.
Once you're set (and you have the ant command 'ant' in your path), go to your command shell and type:
ant
How do I run it?
Now you are ready to launch it, and you can do it by typing
(unix) ./bib2rdf
at the command line and follow the instructions there.
What ontology does this RDFizer use?
The destination ontology is a slight modification of bibtex owl ontology written by Nick Knouf:
http://purl.oclc.org/NET/nknouf/ns/bibtex
Of the three bibtex ontologies I could find, this was the most "modern" and well documented, and also the most transparent mapping of bibtex to RDF. In this ontology, a bibtex file consists of a set of entries, which each have a set of properties, including the type of entry and the database key that is used to refer to the entry from within latex.
How about URIs?
One other noteworthy feature of this parser is the way that it assigns URIs to entries. Entries are assigned a URI that is an MD5 hash of the type, key, and all the fields and values. Thus, identical entries that appear in different files (or the same file at a different URL) will be given identical URIs. Moreover, this hash is invariant to permutations in the fields and capitalization variations in the field keys. Adding or removing a field or changing a field value will change the hash.
The URIs are created from a hash of the record data, so multiple imports the same record will be given the same URI, even if it appears in a different file. This hash is based on the type, key, and all fields *except* the crossref field. If a crossref field is present, then the values from the crossref are included in the hash. The hash is robust to permutations and capitalization of the field names, but not changes to the values.
What is left to do?
- Preserve Bibtex preamble in RDF
Credits
This software was created and is maintained by the Simile Project and in particular:
- Nick Matsakis
