- - - - - - - - - -
view on github
common errors
chaining commands
global options
- - - - - - - - - -
- - - - - - - - - -
ROBOT is licensed under the
BSD 3-Clause License.
Theme by orderedlist

Global Options


Terms in OBO and OWL are identified using IRIs (Internationalized Resource Identifiers), which generalize the familiar addresses for web pages. IRIs have many advantages, but one of their disadvantages is that they can be pretty long. So we have standard ways to abbreviate IRIs in a particular context by specifying prefixes. For example, Turtle files start with @prefix statements, SPARQL queries start with PREFIX statements, and JSON-LD data includes a @context with prefixes.

For robot we use the JSON-LD format. See obo_context.jsonld for the JSON-LD context that is used by default. It includes common, general linked-data prefixes, and prefixes for all the OBO library projects.

If you do not want to use the defaults, you can use the --noprefixes option. If you want to replace the defaults, use the --prefixes option and specify your JSON-LD file. Whatever your choice, you can add more prefixes using the --prefix option, as many times as you like. Finally, you can print or save the current prefixes using the export-prefixes command. Here are some examples:

robot --noprefixes --prefix "foo: http://foo#" \
  export-prefixes --output results/foo.json

robot --prefixes foo.json -p "bar: http://bar#" -p "baz: http://baz#" \

The various prefix options can be used with any command. When chaining commands, you usually want to specify all the prefix options first, so that they are used “globally” by all commands. But you can also use prefix options for single commands. Here’s a silly example with a global prefix “foo” and a local prefix “bar”. The first export includes both the global and local prefixes, while the second export includes only the global prefix.

robot --noprefixes --prefix "foo: http://foo#" \
  export-prefixes --prefix "bar: http://bar#" \

XML Catalogs

OWLAPI, Protégé, and ROBOT use XML catalogs to specify where import files are located when loading an ontology. By default, this catalog is called catalog-v001.xml. ROBOT assumes that a catalog-v001.xml file exists in the working directory and attempts to resolve imports based on that. Because Protègè also predicts that catalog, we recommend sticking to this standard. For more details, see Importing Ontologies in Protègè and OWL 2.

That said, occasionally, you may want to use different catalog files for different purposes. This is especially useful for automated releases with Makefiles. ROBOT provides an option to specify the catalog location with --catalog.

For example, you may want to merge a set of edited import ontologies to create a module. You may have one set of imports for one module, and another set of imports for another module. You can also chain this command with annotate to specify the output ontology’s IRI.

robot merge --catalog catalog.xml\
 --input imports.owl \
annotate --ontology-iri\
 --output results/merged.owl

If a catalog file is specified and cannot be located, the ontology will be loaded without a catalog file. Similarly, if you do not provide a --catalog and the catalog-v001.xml file does not exist in your working directory, the ontology will be loaded without a catalog file. Finally, if the catalog specifies an import file that does not exist, the command will fail.


ROBOT logs a variety of messages that are typically hidden from the user. When something goes wrong, a detailed exception message is thrown. If the exception message does not provide enough details, you can run the command again with the -vvv (very-very-verbose) flag to see the stack trace.

There are three levels of verbosity:

  1. -v, --verbose: WARN-level logging
  2. -vv, --very-verbose: INFO-level logging
  3. -vvv, --very-very-verbose: DEBUG-level logging, including stack traces

XML Entities

If the --xml-entities option is included, entities will be used for namespace abbreviations. For example, in a typical RDF/XML file (OBI, for example), the base prefix may be defined as:


If namespace abbreviations are used, the RDF/XML file will include a header with prefix abbreviations (prior to the rdf:RDF tag). The obo prefix, for example, is:

<!ENTITY obo "" >

The obo abbreviation would be substituted for any instance of in the rest of the RDF/XML file, as demonstrated by the base prefix:


Error Messages


ROBOT encountered a problem while writing the given prefixes to JSON-LD.