Skip to content

Diff tool for comparing RDF schemas/ontologies; can be applied to both CASE and UCO.


Notifications You must be signed in to change notification settings


Folders and files

Last commit message
Last commit date

Latest commit



19 Commits

Repository files navigation

Cyber-investigation Analysis Standard Expression (CASE)

Read the CASE Wiki tab to learn everything you need to know about the Cyber-investigation Analysis Standard Expression (CASE) ontology. For learning about the Unified Cyber Ontology, CASE's parent, see UCO.


An RDF and ontology trouble shooter for CASE and UCO.

What it does

RDFDiff takes output of a tool (JSON/JSON-LD/XML/etc...) and attempts to validate it against a RDF based ontology (OWL/N3/ttl). Any entry in the tool's output that is NOT in the Ontology (specified via CLI) will display an error. CLI arugments can be added to cause a debugger to start so you may explore the graph to view where things went wrong.

How it works will read in an RDF vocabulary defined via -g (glossary) to check a custom tool's output against via -i. The Python library rdflib is used to turn the RDF schema into tripples which are then broken into three lists; subject, predicate and object. Finally, each element of the tools ouput within the tool's subject and predicate are checked againast the glossary's subject and predicate to confirm the existence of these RDF elements. If an element is found or not found it is displayed to the user. BNodes are skipped as they have no appropriate label in RDF and should not be used to verify an ontology.

Why not SPARQL?

In order to facilitate a broad range of ontologies and custom tool outputs, SPARQL queries are not used for verification. CASE and UCO allow for robust flexibility and this tool aims to compliment this approach.


sudo pip install -r requirements.txt 

Unit Tests

The ontology validator is heavily reliant on 3rd pary libraries, primarily RDFlib. RDFLib is under heavy development. To ensure compatability with new releases unit tests have been written to check for consistency.

Run unit tests:

cd tests;

CLI Usage

  • -g: Define the RDF schema (aka glossary) in use for your ontology.

  • -gf: Define the format the schema is in. By default will try to auto-guess based on extension. However, if additional plugins are installed it is best practice to manually specify it.

  • -i: The external tool's output you want to verify the ontology against.

  • -if: Define RDF schema for tool data that's being injested.

  • --debug: Break on errors that occur within --verify.

  • -tg: Print subject, predicate, object for each graph within tool schema.

  • -gg: Print subject, predicate, object for each graph within glossary

  • schema.

CLI Example

  • Check for inconsistencies between graphs: -g case.ttl -gf turtle -i output.json-ld -if json-ld --verify=1
  • Check for inconsistencies between graphs with color output: -g case.ttl -gf turtle -i output.json-ld -if json-ld --verify=1
  • Enter debug mode. This will cause a PDB session to open when an inconsistency
  • is met. This can be useful for manually navigating the RDF graph in Python.: -g case.ttl -gf turtle -i output.json-ld -if json-ld --verify=1

  • Print all graphs for tool's schema. -g case.ttl -gf turtle -i output.json-ld -if json-ld -tg=1
  • Print all graphs for ontology's schema. -g case.ttl -gf turtle -i output.json-ld -if json-ld -gg=1

I have a question!

Before you post a Github issue or send an email ensure you've done this checklist:

  1. Determined scope of your task. It is not necessary for most parties to understand all aspects of the ontology, mapping methods, and supporting tools.

  2. Familiarize yourself with the labels and search the Issues tab. Typically, only light-blue and red labels should be used by non-admin Github users while the others should be used by CASE Github admins. All but the red Project labels are found in every casework repository.

  3. If/when you run into an issue with a given RDF schema format or the script, please open an issue with the error and as much technical detail as you can provide.


No releases published


No packages published