-
-
-
2.3.1
minimalkb 2.3.1 *Released on 29 Apr 2022* - Re-release of 2.3.0 with correct version number. No functional changes.
-
2.3.0
minimalkb 2.3.0 *Released on 12 Apr 2022* - Add ROS support via Revise, Query, Sparql, Manage services
-
2.2.1
minimalkb 2.2.1 - better error handling/reporting - slightly stricter statement parsing - updated LICENSE to 2022 - remove requirement for rdflib >= 6.0.0 (works with RDFlib 4.x.x)
-
2.2.0
minimalkb 2.2.0 - change find() API to make it more consistent - whatever the number of variables, always returns a list of dict with the variables as key - the param 'vars' is now optional. - if no variable is passed, search for all variables by default - anonymous variables (starting by `__`) are now named var1, var2,... in the results - removed unused `minimalkb/services` sub-package -
2.1.0
minimalkb 2.1.0 - add `kb.label` to retrieve (multi-lingual) labels associated to a term - add support for querying direct classes, instances, subclasses, superclasses (via eg `kb.details`) - significantly simplify the event subscription mechanism. Now, you simply pass a list of patterns with an optional `one_shot` parameter (false by default) - better [API documentation](https://github.com/severin-lemaignan/minimalkb/blob/master/doc/api.md) - remove `kb.check` as it was strictly equivalent to `kb.exist` - fix bug where loaded ontologies would not be processed by the reasoner - fix internal methods being mistakenly exposed in the API - a few other minor bug fixes
-
2.0.0
minimalkb 2.0.0 **Attention: minimalkb-2.0 is a major rewrite of minimalkb, and only partially backward-compatible with minimalkb-1.x.x** - complete rewrite of the core, around rdflib-6.x.x - as a consequence, much more standard-compilant parsing of statements - integration with [`reasonable`](https://github.com/gtfierro/reasonable) for fast OWL2 RL reasoning. - a new `sparql` API method has been add, to perform raw SPARQL queries. - old compatibility (`compat`) API methods have been removed - the `find` API method does not take a `constraints` parameter any more - `retract` has been renamed to `remove`
-
1.2.1
minimalkb 1.2.1 see CHANGELOG for details.
-
1.2.0
minimalkb-1.2.0 See CHANGELOG.md for the changelog.
-
1.1.2
minimalkb 1.1.2 see CHANGELOG for details
-
-
1.1.0
minimalkb 1.1.0 Main changes since 1.0.0: - fix issue with SQLite triple store where statements were uniquely identified by an unstable hash algorithm.
-
1.0.0
minimalkb 1.0.0 Main changes since 0.9: - port to python3 - added (some) documentation - api: load() can take option 'models' parameter - provide more exhaustive API documentation with starting with --debug - added support for OWL2 RL rules cls-thing and cls-nothing1
-
0.9
minimalkb 0.9 Main changes since 0.8.1: - fixes + doc in the reasoning engine - clean-up of unit-tests (cf current results below) - a few other minor bugfixes --- Unit-tests results: - `testing/test.py`: 17 tests, 1 expected error (unsupported feature: multiple var in multi-pattern queries) , 0 failure - `testing/test_reasoner.py`: 5 tests, 0 errors, 0 failures - `testing/test_embedded.py`: 17 tests, 4 errors, 3 failures --- Some stats regarding the Dialogs unit-tests (unchanged since 0.8.1): (note that some of the failures are Dialogs bugs, not minimalkb ones) ====================================================================== | suite | nb tests | tests run | failures | errors || OK | |--------------------------------------------------------------------| | statements | 63 | 63 | 12 | 2 || 49 | |--------------------------------------------------------------------| | sentence | 33 | 33 | 0 | 0 || 33 | |--------------------------------------------------------------------| | discrimination | 13 | 13 | 6 | 5 || 2 | |--------------------------------------------------------------------| | parser | 72 | 72 | 7 | 0 || 65 | |--------------------------------------------------------------------| | verbalization | 76 | 76 | 4 | 0 || 72 | |--------------------------------------------------------------------| | questions | 42 | 42 | 13 | 0 || 29 | |--------------------------------------------------------------------| | TOTAL | 299 | 299 | 42 | 7 || 250| ====================================================================== Total time: 14.257743sec
-
0.8.1
minimalkb 0.8.1 Main changes since 0.8: - version correctly bumped ;-)
-
0.8
minimalkb 0.8 Main changes since 0.7: - improved a bit behaviour of memoization: 'undecided' reasoning results are not memoized anymore; 'clear' also clear memoized results - use only the default model (previously all) when no model is specified - 'lookup' request has been significantly improved, in particular by taking into account labels. - 'check' has a first partial implementation (by just checking if checked statement are already asserted) - fix a bad bug leading to models being ignored in an 'about' request - fix a race with the reasoner when clearing the database --- Some stats regarding the Dialogs unit-tests: (note that some of the failures are Dialogs bugs, not minimalkb ones) ====================================================================== | suite | nb tests | tests run | failures | errors || OK | |--------------------------------------------------------------------| | statements | 63 | 63 | 12 | 2 || 49 | |--------------------------------------------------------------------| | sentence | 33 | 33 | 0 | 0 || 33 | |--------------------------------------------------------------------| | discrimination | 13 | 13 | 5 | 6 || 2 | |--------------------------------------------------------------------| | parser | 72 | 72 | 7 | 0 || 65 | |--------------------------------------------------------------------| | verbalization | 76 | 76 | 4 | 0 || 72 | |--------------------------------------------------------------------| | questions | 42 | 42 | 13 | 0 || 29 | |--------------------------------------------------------------------| | TOTAL | 299 | 299 | 41 | 8 || 250 | ====================================================================== Total time: 23.148309sec
-
0.7
minimalkb 0.7 Main changes since 0.6: - fix a serious issue with events (that were not triggered more than once) - support the NEW_CLASS_EVENT properly - slightly improved the reasoning capabilities: every models are now correctly classified, owl:equivalentClass and a few common owl symmetric predicates (owl:sameAs, owl:differentFrom) are handled to some extend. Also, compatibility with the C++ liboro bindings (hence, oro-view) has been extended to events. --- Some stats regarding the unit-tests: ====================================================================== | suite | nb tests | tests run | failures | errors || OK | |--------------------------------------------------------------------| | statements | 63 | 63 | 14 | 11 || 38 | |--------------------------------------------------------------------| | sentence | 33 | 33 | 0 | 0 || 33 | |--------------------------------------------------------------------| | discrimination | 13 | 13 | 4 | 7 || 2 | |--------------------------------------------------------------------| | parser | 72 | 72 | 7 | 0 || 65 | |--------------------------------------------------------------------| | verbalization | 76 | 76 | 4 | 0 || 72 | |--------------------------------------------------------------------| | questions | 42 | 42 | 11 | 4 || 27 | |--------------------------------------------------------------------| | TOTAL | 299 | 299 | 40 | 22 || 237 | ====================================================================== Total time: 34.212394sec
-
0.6.0
minimalKB 0.6.0 Main changes: - many improvements to the way minimalKB deals with ending/closing - support for 'embedding' minimalKB in a Python application (via 'pykb.KB' constructor option) - support for RPC call with keyword arguments - a couple of other bugs smashed out.
-
0.5.0
minimalKB 0.5 Main changes: - support for functional properties (new statements replace existing values instead of adding new facts) - fixed a bug in event id generation that broke the event system - a handful of other minor fixes