ROBOT is a command-line tool for working with ontologies, especially Open Biomedical Ontologies. It builds on OWLAPI and is designed to eventually replace Oort and many functions of OWLTools. Currently implemented commands include: reporting on differences between ontologies, merging ontologies, extracting ontology modules, filtering ObjectProperties, and reasoning. Commands can be chained together to form powerful, repeatable workflows. ROBOT is in early development but is available for use under an open source (BSD) license.
ROBOT is a new command-line tool for working with ontologies,
with special emphasis on Open Biomedical Ontologies (OBO)
OWLTools and Oort are the predecessors to ROBOT. OWLTools
provides various functionality to support the Gene Ontology (GO)
and many other ontology projects. In particular, it has supported
the transition of GO and other projects from OBO format toward
OWL format
OWLTools and Oort provide more core functionality than ROBOT currently does. However, they have “evolved” over the years without an overarching design. The resulting tools are powerful, but somewhat difficult for new users to learn, and the code base was designed around certain assumptions that held true for OBO format ontologies but do not apply to OWL format ontologies.
With ROBOT we aim to create a more modular and extensible code base for developers, and provide a friendlier and more consistent interface for users. We plan to incrementally replace OWLTools and Oort, and promote the use of ROBOT as a standard tool for use by OBO projects. In this paper we describe the design of ROBOT, its usage, and some of our future plans.
robot-command implements the command-line interface
robot-core is designed to be used as a library from any language
that runs on the Java Virtual Machine. Operations are classes that
provide static methods for working with ontologies, usually building
on functionality from OWLAPI
DiffOperation: find the differences between two ontologies
FilterOperation: filter axioms by ObjectProperty MergeOperation: merge axioms from one or more ontologies and their imports into a single ontology ReasonOperation: use a reasoner to add axioms to an ontology
We plan to provide operations for MIREOT
The robot-command module implements ROBOT’s commandline interface using the Apache Commons CLI library.3. Each command implements the Command interface, with main and execute methods, as well as other common methods for getting the command name and usage information. Commands do not necessarily correspond to operations, but often do. The currently implemented commands are:
Overton et al
AnnotateCommand: add annotations to an ontology ConvertCommand: save an ontology to another format: RDFXML, RDFOWL, Turtle, Manchester Syntax, OWL Functional Syntax, or OBO.
DiffCommand: show the differences between two ontologies ExportPrefixesCommand: show the current prefixes (see below)
Commands can be called individually or by CommandManager instances. When used with a manager, commands can be chained together into powerful workflows. When chained, commands communicate using a CommandState object that contains the current OWLOntology instance. Internally, command accepts a state, updates it, and returns it for the next command to use. At the command-line, each command is named by a verb and followed by zero or more options and their arguments. Figure 1 shows an example of a chain of commands for releasing an ontology. robot \ merge --input edit.owl \ reason --reasoner ELK \ annotate \ --annotation-file annotations.ttl \ --output results/example.owl \ convert --output results/example.obo
One source of problems with OWLTools has been the use of prefixed identifiers in OBO and OWL. With ROBOT we have made use of the JSON-LD standard4 and the JSONLD-Java5 library to provide a concise and effective method for specifying and resolving prefixes. A JSON-LD context is a JavaScript object with keys and values that specify how names and prefixes should be expanded, optionally with additional type information. The JSON-LD specification provides an algorithm for resolving names and prefixes to full IRIs.
ROBOT supports the use of JSON-LD for specifying prefixes to be used when loading and saving ontologies and sets of ontology terms. We also support loading ontology and annotation data from JSON-LD files, and from YAML files6 when they have certain structure corresponding to JSON-LD. 5
With ROBOT we aim to provide a user-friendly tool for ontology users and ontology developers. The project repository includes a README with installation instructions, a tutorial on commandline usage with example data files, and full JavaDoc documentation for the code itself. The command-line tool provides interactive help on commands. We plan to include ROBOT in the online OBO Tutorial,7 replacing custom Java code with standard ROBOT commands.
In addition to a suite of unit tests, the tutorial itself serves as an integration test suite. A test harness extracts the example commands from the tutorial document, runs them, and compares the resulting files to the known-good example data files. Continuous Integration systems run all tests every time code is committed to version control. 6
At ICBO we plan to demonstrate the installation and use of ROBOT for a range of ontology operations, based on our tutorial document. ROBOT will also be used in the OBO Tutorial session at the conference. 7
ROBOT is a user-friendly tool for working with ontologies at the command-line and for scripting ontology workflows. It is designed to replace Oort and many of the functions of OWLTools, and to address a wider audience of ontology developers. While still in early development, ROBOT can be used today, and we appreciate feedback via our GitHub repository.
JAO, HD, DOS, and CJM were supported by the National Human Genome Research Institute (NHGRI) P41 grant 5P41HG00227309 to the Gene Ontology Consortium. In addition, HD and CJMs contribution was also supported by the Director, Office of Science, Office of Basic Energy Sciences, of the U.S. Department of Energy under Contract No. DE- AC02-05CH11231. JAO’s contribution was also supported by the Immune Epitope Database and Analysis Project, funded by the National Institutes of Health [HHSN272201200010C].