uri switch, backend refactor, and 3.0 #120
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Result of running `ontload uri-switch NIF-Ontology` (in ttl/) from pyontutils at commit 53f5e7a41b8500fe4639d860f1d2419fe5048492. All the changes are documented in ttl/generated/NIF-NIFSTD-mapping.ttl using owl:sameAs to map from the new uri to the old iri. Almost all changes are the result of placing the fragment portion of the old http://ontology.neuinfo.org/NIF/ identifiers in the new resolver namespace http://uri.neuinfo.org/nif/nifstd/. Changes not following this patter are as follows. 1. All sao- fragments have been converted to sao fragments, unifying the SAO namespace. 2. All 'readable' identifiers in the ontology have been placed in the new NIFRID -> http://uri.neuinfo.org/nif/nifstd/readable/ namespace. Almost all of these were in the backend, and/or were predicates that were used in a select few files. 3. All 'stolen' identifiers where a fragment from another ontology was placed in some nifstd namespace have been placed direcly under NIFSTD. 4. The following identifiers (aside from sao- mentioned above) had their fragments modified. NIFGA:FMAID_7191 -> NIFSTD:FMA_7191 NIFRES:Class_2 -> NLX:293 NIFSUB:nlx_sub_1008001 -> NLXSUB:1008001 NIFNCBISLIM:nif_organ_110519 -> NLXORG:110519 BIRNANN:_birnlex_limbo_class -> NIFRID:birnlexLimboClass BIRNANN:_birnlex_retired_class -> NIFRID:birnlexRetiredClass http://ontology.neuinfo.org/NIF/Backend/DC_Term -> NIFRID:dctermsClass http://ontology.neuinfo.org/NIF/Backend/SKOS_Entity -> NIFRID:skosClass http://ontology.neuinfo.org/NIF/Backend/_backend_class -> NIFRID:BackendClass A superset of these can be obtained using the following python commands run from inside the ttl/generated directory. ``` import rdflib g = rdflib.Graph().parse('NIF-NIFSTD-mapping.ttl', format='turtle') [(g.namespace_manager.qname(old), g.namespace_manager.qname(new)) for new, old in g.subject_objects(rdflib.OWL.sameAs) if 'sao-' not in old and new.rsplit('/', 1)[-1] != old.rsplit('#', 1)[-1]] ``` For the exact changes see uri_replacements and uri_reps_nonstandard in ontload.py or have a peek at ttl/generated/NIF-NIFSTD-mapping.ttl Please see ontload.py starting at line 386 through 648 for the details of the implementation. https://github.com/tgbugs/pyontutils/blob/53f5e7a41b8500fe4639d860f1d2419fe5048492/pyontutils/ontload.py#L386
Result of the execution of qnamefix.py (in ttl/) from pyontutils commit 53f5e7a41b8500fe4639d860f1d2419fe5048492. This commit completes the uri switch to NIFSTD. The only remaining iris that *could* be prefixed using the old prefixes are in annotation properties and have special cases. Below are the two oneliners I used to verify that the switch has gone according to plan. review of all prefixes `grep -r "@Prefix" | grep -v NIF-NIFSTD | grep -v Binary | \ cut -d':' -f2- | sort -u` review of remaining unprefixed urls `grep "<http.\+>" *.ttl */*.ttl */*/*.ttl | grep -v 'NIF-NIFSTD' | \ grep -v '\.ttl>' | grep -v '\.owl>' | grep -v "@Prefix" | \ cut -d'<' -f2 | cut -d'>' -f1 | sort -u` This commit also completes the ontology side of #59. However we still need to make sure that all the new uris in NIF-NIFSTD-mapping.ttl work with the resolver and ultimately resolve to an interlex record (which will come down the road as soon as we can switch off the old neurolex).
This commit removes the last (known) modifications to the external ontologies. ero.owl and pr.owl have had all cases of _birnlex_retired_class removed. Made owl:DeprecatedClass sco NIFRID:birnlexRetiredClass in BIRNANN instead of ero-bridge. This does suggest that we should refactor to use owl:DeprecatedClass and maybe just add a local label.
This commit generally completes the managment of deprecated classes coming from external ontologies. uberon-dead was added to the import chain, but so-dead was not since I need to rework the chemical/molecule import chain first. The last change we need to make for deads is to transition NIFRID:birnlexRetiredClass to owl:DeprecatedClass.
Moved triples around without moving the location of their deepest point in the import chain. All higher imports are converted to annotations. Synthetic enzymes and degredation enzymes are NOT disjoint, this is trivial if you consider that someday we will made degredation enzymes from scratch. Probably should deprecate synthetic enzyme entirely since it is a role.
Major rearrangement of the import chain for the chemical and molecular ontology files. All of chebi-bridge has been converted to annotations to reflect the fact that the classes are defined in the imported file in question (uberon-bridge still needs to be converted). Annotations have been lifted to the point in the import chain where the classes they are annotating are all already imported. All chem/mol ontologies are now imported via chemical-bridge and then through molecule-bridge. The design here attempts to minimize the number of duplicate owl:Class redefinitions and attempts to point people in the right direction to find where the original classes are defined. This should help increase the modularity of our ontologies and allow us to swap parts and execute updates with fewer issues. It should also simplify signaling about where annotations should be placed. If you want an annotation on external ontology but it in the bridge file that directly imports it as an annotation.
It appears that intention with which these ObjectProperties were originally used was to convey the idea that an object was the sum of its parts. HIP was used on unions (and one intersection by mistake) to assert that the class of the union was part of the whole. Of course this modelling doesn't make sense because there was not obvious way to infer that the members of the union were themselves part of the parent class. I have converted all these cases to has_proper_part (consistent with usage in the rest of nif) and added the members of the union to an owl:disjointUnion which seems to have been the original intention based on some of the editorial notes about the usage of HIP. @cmungall some of these cases reveal a conflict with Uberon modelling (for example UBERON:0003002). I will file a bug.
Moving the scigraph config to pyontutils to decouple config from data. See pyontutils commit d0320b9d0a3e373e1da1d8c41aa3ef081d9e9f7d. tgbugs/pyontutils@d0320b9
…cular function which clearly states that it should be sco to process, which ro does correctly
…riants and removed those from the mapping
…zed in the version of owlapi used by ttl-convert
…dified dev-reasoner.patch accordingly
…d process documentation, as well as dev-reasoner.patch
…s, and dev-reasoner.patch
updated deads, and dev-reasoner.patch Reviewing the changes from our old local version of pato to new latest version (2012-06-19 -> 2017-07-10) showed that we wanted most of these changes an there is nothing actively preventing the upgrade.
…are now imported via their canonical iri
…es and management
…kend-refactor` at 2ba7543d57ffaa4e8f61a912475921f108d8aeae
… also converted nif_backend classes to pure annotations since they are already defined by the imports
Updated the text of LICENSE by copying from https://github.com/gentoo/gentoo/blob/master/licenses/CC-BY-4.0. Updated the README to contain a link to the license as suggested by https://creativecommons.org/choose/results-one?license_code=by&jurisdiction=&version=4.0&lang=en
Following a review of doid.owl we now import it via its canonical iri. Removed local copy of doid.owl. Added symp-dead and trans-dead that are imported via doid, they are imported via doid-bridge (deads from parts of the import chain we do not control should be imported at the first point we can). Assoc updates to catalog-extras, catalog-v001.xml, and dev-reasoner.path.
Previous was BFO1SNAP:DependentContinuant. Now is direct to owl:Thing since units as they are defined here don't have being.
Now using BFO:0000146 2d continuant fiat boundary. Not including this as a general 1.1->2 translation since there is not a consistent mapping depending on whether the usage of the boundary was to refer to a material entity or to some immaterial part. Given that the subclasses below this are not used anywhere else, we can make this call without too much trouble.
… of uberon-bridge that were missed when updating the import chain
Also added (and testing) the new versionIRI workflow. Not rolling out en mass quite yet, but soon.
|
Managed to sneek a few more things like versionIRIs and the doid import review in for this. This is about is ready as it is going to get. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This pull request includes 3 major changes for the 3.0 release of NIFSTD.
uri.neuinfo.orgresolver. This closes transition iris to http://uri.neuinfo.org/nif/nifstd/ #59.The current state of the ontology produces reasoning errors because of obophenotype/uberon#1370. These will persist until that issue is resolved or I complete our load time patching infrastructure.
The old backend is not completely gone since we still import an old version of pr that uses the old BFO1.1 backend. The new pr is huge and switching requires further review/planning since it cannot be dropped directly into the ontology as is and thus did not make the cut for updating the 3.0 release. ero.owl was removed from the import chain pending a review of its classes.
This pull request also starts the process of moving documentation into the repository from the GitHub wiki to reduce dependence on external systems.
Note that this pull request also moves the
scigraph/folder and associated loading and curie mapping (which defines the qnames used by the ontology) to pyontutils/scigraph.