ontosample 0.2.6

What changed:

• Updated to owlapy 1.3.3 with all the changes that comes with it. There are no changes on sampler behavior, this is just refactoring.
• Fixed a bug with Forest Fire sampler where random.sample would not work on set objects.
• owlready2 version set to strictly equal 0.40 because in newer versions, removing data property assertion axiom during sampling does not work correctly.
• A few other refactoring done for more efficient code.
• Added tests and git action workflow to ensure the code is not broken on changes.
• Paper experiments reproduction steps removed from README.md

Full Changelog: v0.2.5...v0.2.6

We recommend using the latest version and make sure to check on pypi for yanked releases.

Install/update:

pip install -U ontosample

As always , in case of any question or possible bug, please open a issue and it will be taken care of.

ontosample 0.2.5

Compatibility update patch

• Some refactoring changes due to owlapy 1.1.0
  • base module removed from ontolearn_light (not needed anymore)

To get the latest release:

pip install -U ontosample

ontosample 0.2.3

Bug fixing patch

• Updated imports for owlapy 1.0.2
• Added latest changes from ontolearn into ontolearn_light

To get the latest release:

pip install -U ontosample

ontosample 0.2.2

When trying to use ontosample in ontolearn we found that there are some compatibility issues because classes of ontolearn_light are not recognized as classes of ontolearn.

That is why in this release we made ontolearn related imports of ontosample conditional, based on the presence of ontolearn package. This way, when both packages are installed at the same environment, ontosample will use ontolearn module and not ontolearn_light.

In case of any issue please reach us on the Issues tab.

ontosample 0.2.0

Happy to share the new release of Ontosample.

We got some important changes to the base structure this time. The logic of the samplers stays unchanged.
The following changes were made:

• All the generated knowledge bases/ontologies point to different worlds and no longer conflict with each other.
• Because of that a sampler object can now be reused to perform multiple samples on the initial knowledge base.
• The save_sample method now is static and takes 2 arguments: kb the sampled knowledge base that you want to save and filename the name of the file that will store the ontology (the file will be created at runtime if it does not exist).
• ontolearn module renamed to ontolearn_light so it does not conflict with the main ontolearn package.
• Updated ontolearn_light sub modules with recent changes from main ontolearn package where triplestore logic is removed (a triple store knowledge base is not able to be sampled anyway because it stored in a server).
• Removed methods get_sampled_nodes (can use sampled_kb.individuals_count() instead).
• get_removed_nodes is now renamed to _get_removed_nodes indicating that is only for internal usage. Removed individuals can be retrieved as follows:

  removed_individuals = set(kb.individuals()) - set(sampled_kb.individuals())
  

As always you can use pip install ontosample to get the lates version via the Python Package index.

Don't hesitate to open an issue in case you are having a problem or you just want to suggest something.

Full Changelog: v0.1.1...v0.2.0

ontosample 0.1.1

We are happy to announce the first release of ontosample.

You can now use pip install ontosample to get it via the Python Package index.

Changes since the initialization commit:

• We are keeping only the ontolearn files that are necessary to make the sampling work. Everything else is removed from ontolearn module.
• Refactored the code of samplers, now the code is more compact and they are divided in three modules:
  • classic_samplers.py
  • lpc_samplers.py
  • lpf_samplers.py

Note: If you want to try concept learning algorithms of ontolearn, you have to install the whole ontolearn package because ontosample is using a light version of it.