The concept of nanopublications was rst proposed about six years ago, but it lacked openly available implementations. The library presented here is the rst one that has become an o cial implementation of the nanopublication community. Its core features are stable, but it also contains uno cial and experimental extensions: for publishing to a decentralized server network, for de ning sets of nanopublications with indexes, for informal assertions, and for digitally signing nanopublications. Most of the features of the library can also be accessed via an online validator interface.
This technical paper presents nanopub-java, which is a library for nanopublications. Its source code can be found here:
https://github:com/Nanopublication/nanopub-java
Nanopublications3 [
The library presented here can be useful in a number of scenarios: { To represent and share small chunks of scienti c knowledge and metadata in RDF in a provenance-aware manner (as nanopublications) { To make RDF content veri able and immutable (with trusty URIs) 3 http://nanopub:org { To de ne large or small datasets of RDF content where the data entries can be individually addressed and recombined in new datasets (with nanopublication indexes) { To quickly publish RDF snippets in a veri able and permanent manner (relying on an existing server network) { To retrieve existing nanopublications from the network (5 millions and counting) { To digitally sign RDF snippets (though this is still experimental) Below the details of the library and its web interface are explained.
The library is built upon the Sesame library [
Trusty URIs [
The features of the library are made available through the Java API as well as via a command line interface using the command np. The following features are part of the core of the library, which means that they deal with stable and agreed-upon structures as de ned by the community: { check / CheckNanopub reads a nanopublication or several of them and checks whether any of the well-formedness criteria are violated. If a trusty URI or a digital signature is found (see below), these are checked too. { mktrusty / MakeTrustyNanopub takes a nanopublication that does not yet have a trusty URI and transforms it into one that is identi ed by a newly created trusty URI. { fix / FixTrustyNanopub takes a nanopublication with a broken trusty URI and xes it, i.e. assigns it a new trusty URI. This is useful when a nanopublication has to be changed, which invalidates the hash. Running this command creates a new nanopublication with a valid trusty URI. (Nanopublications
are immutable, so changing something necessarily leads to a new nanopublication.)
In addition to these core features, the library also contains a number of
uno cial extensions (which may or may not become o cial at some point). There
is code to validate informal assertions speci ed as AIDA sentences [
Below, some of the most important commands are explained by examples based on the np command line tool. The same functionality is also available via the Java API. For the sake of these examples, let us assume that we have a le called nanopubfile.trig that starts with the following RDF pre xes: @prefix xsd: <http://www.w3.org/2001/XMLSchema#>. @prefix dc: <http://purl.org/dc/terms/>. @prefix pav: <http://purl.org/pav/>. @prefix prov: <http://www.w3.org/ns/prov#>. @prefix np: <http://www.nanopub.org/nschema#>. @prefix ex: <http://example.org/>. @prefix : <http://example.org/np1#>. :assertion {
ex:drugA ex:treats ex:diseaseB. :provenance {
:assertion prov:wasDerivedFrom ex:some_publication. } :pubinfo { : pav:createdBy <http://orcid.org/0000-0002-1267-0234>.
: dc:created "2015-08-18T15:36:22+01:00"^^xsd:dateTime.
} The provenance and publication info graphs provide meta-information about the assertion and the entire nanopublication, respectively: The lines above constitute a very simple but complete nanopublication. To make this example a bit more interesting, let us assume that our le contains two more nanopublications that have di erent assertions but are otherwise identical: The de nition of the rst nanopublication in this le starts with the head graph that de nes the structure of the nanopublication by linking to the other graphs: :Head { : a np:Nanopublication; np:hasAssertion :assertion;
np:hasProvenance :provenance; np:hasPublicationInfo :pubinfo.
The actual claim of the nanopublication is stored in the assertion graph: } } @prefix : <http://example.org/np2#>. ... :assertion {
ex:Gene1 ex:isRelatedTo ex:diseaseB. @prefix : <http://example.org/np3#>. ... :assertion {
ex:Gene2 ex:isRelatedTo ex:diseaseB.
To check and validate these three nanopublications, we can now use the following
command:
$ np check nanopubfile.trig
Summary: 3 valid (not trusty);
$ np mktrusty nanopubfile.trig
These nanopublications can now be transformed into ones with trusty URIs using
the following command (resulting in a new le trusty.nanopubfile.trig):
Using the same command in verbose mode with the argument -v shows us the
newly generated trusty URIs for the three nanopublications:
$ np mktrusty -v nanopubfile.trig
Nanopub URI: http://example.org/np1#RAHGB0WzgQijR88g_rIwtPCmzYgyO4wRMT7M91ouhojsQ
Nanopub URI: http://example.org/np2#RA4xTdhe2gPctqvAwdgTU4eRiR1aTQlTYJcF3Sohe5Cus
Nanopub URI: http://example.org/np3#RAEjvXP0xTkeIa2mKmYT66i_PAJ-u-k0uRBd6_sMe9qG0
As they are tiny snippets of data, nanopublications are most useful when they
grouped and combined in small or large collections. We therefore need a simple
method to refer to collections or sets of nanopublications, which is achieved by
the experimental proposal of nanopublication indexes [
Index URI: http://np.inn.ac/RAFa_x4h0ng_NXtof35Ie9pQVsAY69Ab3ZQMir2NP8vGc The nanopublications of the new index are saved in a le called index.trig unless speci ed otherwise with the argument -o.
Moving to the part that involves the server network, nanopublications that have trusty URIs (which includes nanopublication indexes) can be published to the network with the following command: $ np publish trusty.nanopubfile.trig 3 nanopubs published at http://np.inn.ac/ The publication status of a given nanopublication can be checked like this: $ np status -a http://example.org/np1#RAHGB0WzgQijR88g_rIwtPCmzYgyO4wRMT7M91ouhojsQ URL: http://np.inn.ac/RAHGB0WzgQijR88g_rIwtPCmzYgyO4wRMT7M91ouhojsQ URL: http://ristretto.med.yale.edu:8080/nanopub-server/RAHGB0WzgQijR88g_rIwtPCmzYgyO... URL: http://nanopub-server.ops.labs.vu.nl/RAHGB0WzgQijR88g_rIwtPCmzYgyO4wRMT7M91ouhojsQ URL: http://nanopubs.stanford.edu/nanopub-server/RAHGB0WzgQijR88g_rIwtPCmzYgyO4wRMT7... URL: http://nanopubs.semanticscience.org/RAHGB0WzgQijR88g_rIwtPCmzYgyO4wRMT7M91ouhojsQ Found on 5 nanopub servers.
A given nanopublication that is published on the server network can be retrieved via its URI:
$ np get http://www.tkuhn.ch/bel2nanopub/RAhV9IpiUEjbentzGivp1Lbx0BVegp5sgE3BwS0S2RAYM All the servers in the network are checked until the nanopublication is found and successfully veri ed. This command is therefore reliable even if one or several servers are down. Instead of the complete URI, it is also possible to just specify the trusty URI artifact code:
$ np get RAhV9IpiUEjbentzGivp1Lbx0BVegp5sgE3BwS0S2RAYM To get the content of a nanopublication index (and not just the top-most index nanopublication), argument -c can be used:
$ np get -c -o content.trig RAtF0ivB9B8cb-u3K_zElgmRBxiDwfym1yVBRY6VAyWvE Argument -o speci es again the name of the output le. The remaining commands as introduced above are equally intuitive to use. Just entering the command without any arguments will output a list of all argument options. 4
Many of the features described above are made available through the nanopublication validator Web interface,6 an instance of which can be accessed at http://nanopub:inn:ac. Figure 1 shows a screenshot. With this interface, the wellformedness of nanopublications can be checked as well as the adherance to a
number of patterns. They can furthermore be transformed into di erent RDF serializations, and published to the server network. Loading of nanopublications is possible via form input, upload, fetching from a URL, SPARQL endpoint access, and retrieval from the server network. In general, this web interface and the underlying library are supposed to support the development of best practices for the nanopublication community by providing a solid basis for discussion, by allowing for experimental features to be tested and discussed, and by facilitating the implementation of prototypes.
The current server network on which many of the uno cial features depend, can be explored via a monitor interface at http://npmonitor:inn:ac. 5
The nanopub-java library provides a stable implementation of the nanopublication concept, adhering to its speci ed guidelines. It is openly licensed under the terms of the MIT license, is available on The Central Repository,7 and has so far been used in about a dozen open-source codebases.8
In general, we believe that this library can be a valuable resource for tools that use RDF data in the context of provenance recording, reproducibility, data publishing, data reuse, and reliable retrieval of Linked Data. 7 https://search:maven:org/#artifactdetailsjorg:nanopubjnanopubj1:7jjar 8 https://github:com/Nanopublication/nanopub-java#usage-tracking