Changelog

Summary of changes.

Oct, 2026: v 2.5

Namespaces
- Ontospy.namespaces, and the namespaces section of the documentation, list only the prefixes an ontology uses or declares, plus rdf, rdfs, owl and xsd, instead of all of rdflib’s ~30 defaults (brick, csvw, dcat, odrl, sosa…). Eg Pizza goes from 30 prefixes to 5. No qnames change
- add_namespace() now also updates Ontospy.namespaces
- SPARQL endpoints: queries send only the prefixes they use, instead of all of rdflib’s defaults; eg a query on DBpedia went from 29 PREFIX lines to 11. Queries through Ontospy.query() can still use rdflib’s well-known prefixes (foaf:, schema:…) without declaring them. If you call rdflib_graph.query() directly on an endpoint, declare the prefixes your query uses
Documentation generation (gendocs)
- Ontology metadata is listed in alphabetical order, each property once, and nested values such as an ontology’s editors (blank nodes) are shown with their names and details instead of as node ids
Examples: a refreshed selection of 32 ontologies, adding widely used standards (PROV-O, DCAT 3, Dublin Core Terms, SKOS, SOSA/SSN, W3C ORG, ODRL, GeoSPARQL 1.1, BFO 2020, FIBO Business Entities, the COAR Resource Types thesaurus and the DCAT-AP SHACL shapes); the index shows more titles and descriptions, and opens each example in a new tab

Oct, 2026: v 2.4

Large and complex ontologies
- Fix RecursionError on large ontologies, eg FOVT, and on class or property hierarchies with cycles or deep multiple inheritance. FOVT (12k classes) now loads in about 2 minutes, and its documentation builds issue-142
- Faster loading: a SHACL query used to run once per class, it now runs once
- Fix descriptions and labels whose value is a URI rather than a string, which crashed the documentation
Properties
- Properties with no domain or range of their own inherit those of their superproperties; the documentation shows them as “inferred from superproperties” instead of owl:Thing, and class “Usage” tables list them under the right class issue-126
- Properties whose type is a subclass of a property type are recognised, eg in Brick brick:tilt a brick:EntityProperty, where brick:EntityProperty rdfs:subClassOf owl:ObjectProperty issue-118
- A property’s main type no longer depends on which of its types sorts first: eg Pizza’s hasBase showed as owl:FunctionalProperty, now owl:ObjectProperty. Property pages list all of a property’s types issue-53
- Properties whose URIs differ only by case, eg FOAF’s givenname and givenName, are no longer merged, and get a page each in the documentation
- Fix blank-node ranges being added to a property’s domains
Documentation generation (gendocs)
- The output folder is no longer emptied before generating the documentation: files are overwritten when needed, and Ontospy warns about files it didn’t write. The new --emptyfolder option (empty_folder=True in Python) clears the folder first issue-60, issue-62
- Non-redundant trees for reasoned ontologies: with --nonredundant, a class isn’t repeated under a parent when it’s already under one of its subclasses. In Python, ontologyClassTree(non_redundant=True) and the other tree methods. Based on pull-87 by Andrea Mignone
- Metadata values that are URLs are shown as links issue-61, and entities have a “See also” section for rdfs:seeAlso issue-65
- Markdown: individual and concept pages are saved as .md again (they were .ms since 2022, so links to them were broken)
Examples regenerated with this version; the index page now shows each ontology’s title, description and counts

Oct, 2026: v 2.3

Namespace prefixes: the prefixes declared in an RDF file now win over rdflib’s defaults. Since rdflib 7, an ontology declaring schema: for http://schema.org/ was shown with schema1: names, because rdflib binds schema to https://schema.org/ by default issue-135
Load RDF from existing rdflib graphs, with the new graph_obj parameter issue-111, based on the idea in pull-114
Fix loading from open files (file_obj), broken since Python 3
Paths, strings, open files and rdflib graphs can be combined when creating an Ontospy object
Every Ontospy parameter is now documented: help(ontospy.Ontospy)
Documentation generation (gendocs)
- Custom templates: every visualization takes a jinja2env parameter, so you can override any of Ontospy’s templates with your own, see “Customising the templates” in the docs pull-138, answers issue-140 and issue-143
- Markdown: descriptions are no longer wrapped in HTML, so Markdown in them is rendered pull-133, fixes issue-64
- Markdown: the class pages list properties in Markdown tables instead of HTML pull-137; descriptions with line breaks or | no longer break the tables, and multiple ranges are comma-separated
Packaging: the licence is declared as an SPDX expression (License-Expression: MIT), replacing the deprecated licence classifier

Oct, 2026: v 2.2

Support Python 3.12+: replace the removed SafeConfigParser with ConfigParser pull-144, fixes issue-129 and issue-141
Packaging: dependency minimum versions are now enforced on install (eg rdflib>=6.0.0), pygments>=2.15, and Python 3.10 or later is required; the source distribution (.tar.gz) can be installed again
Fix gendocs -i (documenting individuals) failing with TypeError: 'method' object is not iterable on every template; the multi page HTML now has pages for individuals again issue-145
Reorganise ontospy/extras: the interactive shell (ontospy shell) moves to ontospy/shell; the old, unmaintained experiments in extras/hacks move out of the package to a top-level archive/ folder
SPARQL endpoints
- load_sparql now uses rdflib’s SPARQLStore; pass credentials as a (user, password) tuple
- Results are requested as JSON, which fixes the Repeat node-elements inside property elements error with Virtuoso endpoints such as DBpedia
- PREFIX declarations in a query are moved to the graph namespaces, so endpoints don’t receive them twice
Ontospy.query takes a new format parameter: list (default, as before), csv, json, txt, xml, or df for a pandas DataFrame (requires pandas)
Add Ontospy.add_namespace(prefix, uri) to bind or override a namespace prefix
Entity triples are extracted with graph.triples() instead of a CONSTRUCT query, which some triplestores returned in unexpected formats
Fix RDFLoader.load_file (missing self argument)
Update the RDF validation services suggested when parsing fails
Tests run with pytest instead of the unmaintained nose; CI runs on Python 3.10 to 3.14 issue-131, incorporating pull-130 and pull-132

Nov, 2022: v 2.1.0

D3 visualizations improvements
- Split pages for different object types
- Refactor code

Oct, 2022: v 2.0.0

Full release of version 2. The main library API is the same, but various internals have changed hence this release may be backward-incompatible for users that have custom extensions using Ontospy’s source.

Remove all Django dependencies, replaced with Jinja2
Rename ontodocs module to gendocs
Drop support for python2
Remove setup.cfg and universal build (only py3)
Refactor code / clean up
Refactor setup.py

May, 2022: v 2.0.0-alpha

An alpha release of 2.0.0. More refactoring and code improvement is planned before the full 2.0.0 release!

Merged additional SHACL support branch pull-107
Fix error loading JSONLD graphs issue-1416
Replace lov.okfn.org issue-110

December, 2021: v 1.9.9.4

Remove deprecated Ontospy.__computeTopLayer
Refactor Ontospy.get_class and related methods so to have less duplicated code
- Added Ontospy.get_shapes
Refactor Ontospy.nextClass and related methods so to have less duplicated code
Refactor Ontospy.ontologyClassTree and related methods so to have less duplicated code
Improve Ontospy.get_shapes so that it enforces URI uniqueness only among shapes URIs. This sets the ground for pull-107.
Refactor tests code and clean up tests rdf folder
- Add uco_monolithic.ttl and scigraph for shapes tests

December, 2021: v 1.9.9.3

Fix bug: https://github.com/lambdamusic/Ontospy/issues/105

November, 2021: v 1.9.9.2

Add ability to extract individuals (instances)
- -i option from command line
- hide_individuals=False parameter in Ontospy
All gendocs visualizations updated so to handle instances
Improve handling of multi-ontology graphs: ontologies list ordered alphabetically by default
Add format parameter to scan command
Fix for --verbose option

September, 2021: v 1.9.9

add pref_title and pref_language options to ontospy
update CLI to accept these parameters for documentation generation
- eg ontospy gendocs -l --preflabel label
improve all viz templates to use user provided titles
add |linebreaks to bestDescription method for better rendering
various code improvements for tests

July, 2019: v1.9.8.3

gendocs option: no browser prevents opening html viz in browser
gendocs -l added as shortcut of --lib

December, 2018: v1.9.8.1

ad hoc ordering of commands
examine becomes scan

December, 2018: v1.9.8

small bug fixes
refactor action visualize / activated ‘alpha’ dataviz
- allow to pass viztype from command line
implicit/explicit classes and predicates switch -x
- this is useful to inspect a small RDF snippet for example too see what predicates are being used
option to show raw RDF data -r
improvements to shell command
namespaces printed out by default

December, 2018: v1.9.7

removed support for Python 2
fixed issue with serialization print
comments all go to sterr by default => eg now easier to to ontospy serialize foaf.rdf json-ld > ~/foaf.json
embedded ontodocs plugin within main source code
- requires install with [HTML] flag
utils and quickdocs commands added to CLI
CLI: all commands when called default to HELP
fix bug for retrieving ontologies from online repos

November, 2018: v1.9.6

refactoring and tested integration with ontodocs
allow calling dataviz (ontodocs) from CLI
bug fixing

November, 2018: v1.9.5.2

bug fix for scan command

October, 2018: v1.9.5.1

bug fix with urllib.parse.quote (python3)

September, 2018: v1.9.5

testjsonld command: util to send a jsonld file to https://json-ld.org/playground/

August, 2018: v1.9.4

refactor cli using subcommands
add utils to load RDF and serialize to another format
improve caching command (to clean up cached items folder)
print out ontospy version info only when VERBOSE = True
improve handling of JSONLD (issue with serializing jsonld https://github.com/RDFLib/rdflib-jsonld/issues/42)
rename OntoSpy to Ontospy
add schema.org domainIncludes and rangeIncludes properties for building domain and ranges (https://github.com/lambdamusic/Ontospy/pull/46)
add shaped properties test (https://github.com/lambdamusic/Ontospy/pull/47/commits/ce9c652bd58d42d2a4be4d2de381402561176e58_)

June, 2018: v1.9.2

improved support for JSONLD * https://github.com/RDFLib/rdflib/issues/436
add html5lib to requirements

May, 2018: v1.9.1

improved ontospy.build*entity_from_uri()
- can pass custom ontospy.RDF*Entity class as argument
- example: see tests/test_methods.test4

May, 2018: v1.9

Refactoring method names:

all ‘extract’ methods renamed as ‘build’ renamed as extract_all
build_entity_from_uri: extract all triples for a specific URI (even if not in model) and instantiate RDF_Entity() so that it can be queried further
x.rdfgraph renamed as x.rdflib_graph
x.serialize renamed as x.rdf_source
tidy up all tests code
ontospy properties returning entities: all renamed using the all_* pattern eg all_classes etc..

March, 2018: v1.8.8

fix https://github.com/lambdamusic/Ontospy/issues/33
instances method for classes on-demand
simplified describe method
cli: no args shows help
fix reference to ontodocs
Merge pull request #37 from satra/patch-1

June, 2017: v1.8.1

sparql support and renamed various methods etc…
added shapes

April, 2017: v 1.8

ontospy-docs created (all documentation moved to another notes file)
review command line interaction
provide simple example from python
complete setup.py and installation scripts requirements etc..
rename to ontospy-viz?
clean up init file! too many things there

April, 2017: v1.7.7

remove VIZ stuff

March, 2017: v1.7.6

wiki online
removed dependency on readline for non-shell usage

February, 2017: v1.7.4

allow multi export to be passed a custom folder
- so that we can create a predefined ‘sample set’ of ontologies for docs
do another multi export with selected set of ontologies
publish new ontospy viz outputs
- take a selection of outputs from ontospy viz (only good ontologies)
- put online
made a release without js interactive viz, but address all problems related to basic viz eg html rendering etc…
add entities-tree top level link (with sublinks?)
need an index.html file for each viz, not dashboard
footer bk color for dark themes must be removed [slate, darkly]
improved domain/range panel
add label panel or somewhere.. (for class)
from class button still looks weird
find a way to shorten long URIs, either preview them or add on the fly prefix namespaces @for later
clicking on ontology name open original url
clicking on ontospy goes to https://github.com/lambdamusic/Ontospy
add tree to properties too
allow to click on owl:Thing too to go to top level, create page for it
add tooltip to family tree diagram
bestDescription should join all comment/desc props ** (test using BFO)
clean up legacy exports
test update scripts for other d3 viz (for testing etc..)
set verbose default = False !
static files copy only things on demand (eg d3 only)

November, 2016: v1.7.1

published 1.7.1 (see git for details)
updated docs
tested for python 3
removed dependency on github3.py
tested on python3
self.graphuri should be different from the SOURCE of the graph!
more refactoring in RDFLoader class
cli launcher for viz: check that args are passed correctly
cli launcher for shell: check that args are passed correctly
updated setup.py
fixed bugs with bin launchers
improved config.py for ontospy-viz
fixed bugs with viz
renamed VERSION.py
updated main for ontospy so to use click
- complete refactorin of code
- add way to print out more info
moved ‘manager’ actions into main ontospy launcher
ontologies listed by default in 2 columns
improved sparql-cli by upgrading to latest pygments
- works ok with local file being loaded in…

September, 2016: v1.6.9

improved markdown export
applied all changes previously added to branch (eg slug)
- in theory dev branch is behind
added config.json to store visualization info
simplified action_visualize
added way to copy static files too
changed OWL:Thing to owl:Thing
Ontospy graph changes
added hide_base_schemas = True to Graph builder
- selectively decides whether to query for rdfs:, owl: classes/predicates etc.. [updated query too]
- improved verbose options to hide everywhere
Turtlep-prompt.py
bootstrapped a way to load prompt completion dynamically from RDF files
removed number from slug fielnames (markdown export etc..)
ontospy -r : loads an interactive session only with that URI!

August, 2016: v1.6.8

completed markdown export
added missing namespaces to entity-graphs
make export into folder //files..
generate MD files - started: done basic template and ontology one - @todo class, prop etc… - can do a tree? yes use ``` and outsup simple termnial tree structure
added pygments to dependencies - sudo pip install pygments -U [v 2.2 for turtle lexer]
Added turtle lexer to docs
removed domain/range from split view docs
added viz_splitter_multi to viz selection routine - split logic so to handle multi file saving
added inferred properties to viz_splitter_multi - layout still not very good @todo improve
‘stats’ method to ontograph object
.slug property to onto entity (filename-safe string)
moved all splitter templates to /splitter/.. - updated setup.py and django set up too
added all entities to splitter viz - removed all splitter to selection tool
started testing adding namespaces to ontology entity, so to make ontology annotations look better - done and it seem to work ok
started splitting files for entity-based split output
added search boxes
refined domain/range display of properties
updated diagram for entities with multiple parents
improved new splitter view - http://methvin.com/splitter/ [old release] - https://github.com/e1ven/jQuery-Splitter [updated for latest jquery]
added -o option for exporting to a file - eg python -m ontospy.ontospy -v -o ~/Desktop/
added simple graph family tree - http://thecodeplayer.com/walkthrough/css3-family-tree
cleaned up a bit the basic visualization
did basic splitter version of viz

July, 2016: v1.6.7

fixed bug with dendogram json links - various other layout improvements to dendogram
raw_input fix
float number division fix
ConfigParser fix
unicode method fix
dict slicing fix
added ‘viz_d3bubblechart’
updated VISUALIZATIONS_LIST catalog
refactored json tree building functions
added several other viz for experimenting
added standard way to add/test viz - eg >python -m ontospy.viz.viz_html
simplified way VISUALIZATIONS_LIST are managed - just import, give unique name and add
simplified way test_run for viz is accedded
- func = locals()[“run”] # main func dynamically run_test_viz(func)
viz_d3packhierarchy: - basic working model for viz_d3packhierarchy
added new folder [viz]
added new static files dir structure in there - added to manifest file
remove support for Python 2.6 or less
manager: updated commands and dialogs too
ontospy: updated commands and shortcuts
loading an onto via import uri doesnt update the ls command - fixed
fixed bug with printTriples returning warnings
removed -l list from manager command
updated list command so to show also dates etc..
completed info inferred command
improved headings on info command
updated _get_prompt
improved print outs of info colors
improved messages on info parents/descendants etc..
added domain_of_inferred = [] and range_of_inferred = [] to classes
improved _printClassDomain and _printClassRange in shell

May, 2016: v1.6.6

added import command for local files too [from shell] - import web - import local - import repo
serialize at ontology level should serialize the whole thing
added GLOBAL_DISABLE_CACHE to init.py for testing cache-less
updated verbose output of ontology load
fixed error with display of domain/ranges not in ontospy entities - eg with schema.org ‘dependencies’
removed ‘display’ command
added info options: [‘toplayer’, ‘parents’, ‘children’, ‘ancestors’, ‘descendants’]
added incremental search by passing a pattern
cleaned up ontospy.py and moved stuff into init.py
fixed ontospy.tests.load_local and ontospy.tests.load_remote
test fails with http://purl.org/dc/terms/AgentClass - fixed
foaf fails with inspect - fixed
moved zen quotes outside
improve inspect for ontologies - testing adding serialization section - other improvements
inspect ==* info
added ‘download’ keyword facility
improved download
removed shell and tree commands
added ls <..> tree
Threading removed..
completed new inspect command
added better printdebug info, especially for bootstrap command
experimented with threads
add JSON LD stuff - https://github.com/RDFLib/rdflib-jsonld
file command (contains rename and delete)
fixed bug with missing ‘/’ in filenames
added bootstrap command for empty repos
shell: add ARG to ‘download’ a specific model
- eg download http://www.rkbexplorer.com/ontologies/acm
added extra help from console run
added changelog on MD file
added a version-based caching mechanism
import from web feature expanded
added ‘download’ command
renamed ‘quit’ into ‘q’
screencast
various other improvements
pyfiglet
removed render module
add commands for delete, rename from within shell
add commands for load from uri
add command for visualize

February, 2016: v1.6.5

updated documentation with new template
- https://textblob.readthedocs.org/en/dev/
small updates to shell
show qname also for list
fixed error with all() method retuning complex objects
rationalized opts lists and help methods
refactored messages with help_.. mechanism
added ‘shell’ command to shell
moved manager up one level…
added a temporary namespace abbreviation bsaed on file name
renamd utils to manager
added domain/range info to classes and properties
completed display of ‘stats’ and ‘overview’
fixed bug with extra ‘/’ in path
moved export stuff into -e (ran from main ontospy)
importer has become ‘ontospy-utils’
added a routine to test different RDF serializations - eg so to address http://pliny.cch.kcl.ac.uk/ontology/base.owl
show namespaces
tree: no need to show IDs anymore - solves also inconsistency with N displayed within ls command
add find command - like ls, but with a pattern
updated print triples (defined locally)
refactored folder structure {core/, extras/}
renamed /importer and /exporter
printGenericTree>showtype
- [prob this is enough not to have to show top level items to trees]
updated API: tree / ls
fixed django loading error https://docs.djangoproject.com/en/dev/releases/1.7/#standalone-scripts
fixed bug with printing 2 columns results in one column with odd etc…
better support of library location updating
added delete operation
added pprint2columns to select ontology function
option to select a viz output
added basic d3 tree viz
fixed bug with ‘next’
make models dir selectable by user
added ontospy-manager
simplified main ontospy a lot

October, 2015: v1.6.4

ontospy-doc
added option to select from local / cache
added domain/range (with some limitations due to not all classes being rendered in ontospy eg owl or rdf ones )
added save on github option
viz: shows skos concepts too
changed ‘repo’ into ‘library’
removed cached column from report
updated extraction of triples so that blank nodes are followed through
- eg python -m ontospy.tools.genviz ~/.ontospy/models/time.owl
in shell, added commands for toplayer, ancestors etc…..
first simple w3c like documentation page like http://xmlns.com/foaf/spec/
added python -m ontospy.tools.callviz ~/.ontospy/models/philosurfical_2010.owl
added a first django based templating mechanism
added hidden folder /.index/cache - updated code for creation
added server app
added simple test to generate a viz from a local file
- python -m ontospy.hacks.callviz
fixed error with http://ecoinformatics.org/oboe/oboe.1.0/oboe-core.owl when doing –import
tested out python webserver - idea: put all exports into folder which is then browsed as a server
make the readthedocs documentation the main one - rst, mkdown.. make it very minimal!
fixed error with ontospy-web circular imports - now ontospy-web is only standalone
updated how to and DOCUMENTATION import ontospy etc..
refactored all the code - cleaned up internal modules loading _ https://docs.python.org/2/tutorial/modules.html _ http://python-notes.curiousefficiency.org/en/latest/python_concepts/import_traps.html
fixed bug with –web urllib request
moved docs to top level
next command
various improvements
fixed error with pickling max recursion exceeded
redirects handling for any ontospy loading

August, 2015, v1.6.3

added command to create/recreate all caches eg –rebuild
refined action_listlocal to include dates etc…
added actions for skos concepts
added support for SKOS
updated command line options
began adding localization for properties
@todo maybe serialize and triples should work only from within a currentEntity - so we can pass also args…
setup: added colorama https://pypi.python.org/pypi/colorama
added more shell functionalities
added new local .ontospy folder for data
tried pickling stuff

June, 2015, v1.6

fixed docs errors
complete quick start examples in docs
added docs to https://readthedocs.org/projects/ontospy/
add new package to PyPi
did a first cut of documentation
removed pizza.ttl by default
add option to list local ontologies
add command line command to list prefix.cc stuff
new namespaces catalog script
tested import paths
installed in local and it seems to work!
completed checking the old tests
fixed testing of lode and frbroo
added new test for sparql endpoints
added more methods on Ontology class i.e. what do we want from it? - get all classes it defines - get all properties it defines
test connecting to a sparql endpoint - how to query an endpoint? are the same sparql functions useful? - https://lawlesst.github.io/notebook/rdflib-stardog.html
started basic integration of sparql
matcher: - match not only classes but also properties against each other - better command line arguments - verbose mode too
command line add option for printing label
get label method: - get the english one by default! - so try to the matching with CIDOC
added colors and labels to tree printing
matcher.py : pass two ontologies, match classes
fixed ERROR in foaf etc.. with recursive methods
cleaned up namespaces
removed duplicates in foaf - due to entities being defined more than once with different Class predicate (e.g. RDF or OWL)
added v.2 to github DEV branch - from now on use branches to do development! - master is still the most recent active version..
added all command line support
fixed error loading http://www.cidoc-crm.org/rdfs/cidoc_crm_v5.0.2.rdfs
split up annotation props vs others using rdftype val
added timing info
get domain and ranges info for objprops - attached also to classes
added method to run sparql query
optimise the allsupers/subs methods – they really slow down queries!
added PARENTS/CHILDREN as the immediate relatives - then DESCENDANTS/ANCESTORS which are calculated at run time - e.g. from class / – get parent / –recurse up until top!

Mar, 2015, v1.1

all supers / all subs - still using the SPARQL stuff for now. Easier..
added tree info for properties too
better API = load then scan to get ontologies
progress with ontospy3 - can handle multiple ontologies - added test RDF file in folder
started adding docs using sphynx
Fixed issue with missing templates for webviz - added data folder to manifest file - changed location of viz file to home folder (cross platform)
updated the import path and removed the vocabs library - now the best way to import should be this:
added tests - python test_simple.py
on uri2nicestring method, if the uri matches exactly a namespace it is not shortened [issue deriving from NPG portal]

January, 2015, v1.0

updated uri2niceString : Literals are returned as quoted strings
updated generation of dict representation of classes: both pred and obj are shortened using uri2niceString
began work on ontospy v.2 - done ontology loading bit
removed utilsRDF

October, 2014, v0.8

installed package on Pypi
added executable: ‘sketchonto’ (generated via distutils)
sketch: added shortcuts - (sub == rdfs:subClassOf) (class == owl:Class)
sketch: added commands for a default_sketch
added Sketch utility
added a dot file export mechanism - so that we can open this stuff quickly in omniGraffle - http://paco.to/2004/omnigraffle-and-graphviz - http://en.wikipedia.org/wiki/DOT_(graph_description_language)
refactored ontology loading routine

July, 2014: v0.5

added webViz method that integrates d3 stuff - to be tested more! - added the tree view template too - but data format isn’t finished yet (ps check also http://bl.ocks.org/mbostock/4063570) http://bl.ocks.org/mbostock/4339184

Jan, 2014: v0.1

Ideally, here’s what I want to do:

load an ontology e.g. a model
internalize the ontology, and be able to query a knowledge base (or the same ontology file) using that model - e.g. autocomplete on properties (=dont have to remember all the model details) - e.g. introspection kind of stuff