150 likes | 272 Views
eagle- i and ISF: a Brief History. 2009: eagle- i ontology (ERO) development begins 2011: the ISF is seeded with the complete eagle- i ontology and extended 2011-present : ERO and ISF ontologies diverge as they are independently developed
E N D
eagle-i and ISF: a Brief History • 2009: eagle-i ontology (ERO) development begins • 2011: the ISF is seeded with the complete eagle-i ontology and extended • 2011-present: ERO and ISF ontologies diverge as they are independently developed • original ERO content removed/changed in ISF, new content added to ERO, etc • Fall 2013: refactoring to harmonize diffs and allow automated generation of ERO from ISF
Refactoring Goals • Harmonize diffs between eagle-icontent covered in ERO and ISF • Refactor the file modularization of ISF (from topic-based to source based approach, for easier updating) • Implement automated pipeline for generation of ERO core ontology from ISF content • Future development of core content in ISF
Refactoring Workstream 1: Core Content • Identify Diffs: the master-diff script lists all axioms from ERO import chain that are different/missing in ISF • Trim MIREOTed axioms: leave only coreterm axioms (ero/isf/vivo) this curation workstream • Classifycore diffs and Resolve: via manual or automated approaches • ERO-Deprecated Terms: ERO contains deprecated terms it originally donated to ISF • ISF-Deprecated Axioms: ISF has deprecated terms/axioms it originally imported from ERO • New ERO Axioms:ERO has created new terms/axioms not in the ISF • Inconsistent IRIs: the same concepts are represented by different terms/modeling in ERO vs ISF
Refactoring Workstream 1: Core Content • Identify Diffs: the master-diff script lists all axioms from ERO import chain that are different/missing in ISF • Trim MIREOTed axioms: leave only coreterm axioms (ero/isf/vivo) this curation workstream • Classify core diffs and Resolve: via manual or automated approaches • ERO-Deprecated Terms: maintain terms in a separate file outside of ISF import chain (ei-deprecated.owl) • ISF-Deprecated Axioms: move axioms into a separate file outside of the ISF import chain (ei_local.owl) • New ERO Axioms: add axioms to ISF in separate file in the ISF import chain (isf_temp.owl), or (ei_local.owl) • Inconsistent IRIs: IRI mapping and merging is required to resolve IRI differences
Refactoring Files for Core Content(Workstream 1) • ei_deprecated.owl- will hold terms that were deprecated in ERO. Will not be in ISF import chain. • isf_temp.owl– will hold axioms new to ERO that can be added to ISF. Will be read into the ISF import chain. • ei_local.owl- will hold axioms that belong only in ERO. Will not be part of the ISF import chain. • ero.owl– will be updated to align with ISF, and be deprecated for further development once first auto-generated version is made from ISF content. * no ISF files will be harmed in the curating of this ontology
Refactoring Workstream2: MIREOTed Content • Goals: • Resolve any diffs in axioms related to MIREOTed terms • Move MIREOTed content into source-specific files to facilitate automated updates • MIREOTed content in ISF currently mixed in same owl files as core (ero/isf/vivo) content. The segregation of content in these files based on topic/domain (source folder) • Preferred modularization strategy is to segregate MIREOTed content by source ontology (as is currently done in eagle-I, e.g. obi_imports.owl, go_imports.owl)
ei ISF
Workstream 2: MIREOTed Content • ERO imports files will be copied/renamed into ‘mireots’ folder of ISF SVN (e.g. go_mireot.owl) • The mireot-refactorscript will read axioms in go_mireot.owl and look for these axioms in ISF files • axioms found in ISF files will remain in go_mireot.owl • axioms absent in ISF files will be triaged in go_mireot-new.owl • Curators manually review mireot-new axioms and move to go_mireot.owlas approved. • Any axioms remaining in source ISF files that declare GO terms will be reviewed and moved to go_mireot.owlif appropriate.
Workstream 2 Outcomes • The ISF ‘source’ files should not contain any declarations of MIREOTed terms (but these classes can be referenced in axioms on other classes). • The ‘mireot’ folder files will contain all MIREOTed terms separated into source-specific files (go_mireot.owl, obi_mireot.owl)
Workstream 3: ERO view from ISF • Once all diffs resolved as outlined above, all content needed to generate ero.owl should exist in the ISF import chain and accessory local ei files: • ei_deprecated.owl • isf_temp.owl • ei_local.owl • Tools/scripts we will use here not yet determined (Ontodog, OWL-tools, custom script) • Will also need to define approach for evaluating new ERO ontology for completeness, consistency, and its support of eagle-i applications
Documentation • Issues and progress tracked in ERO google code, tagged with ‘ISFmigration’: • http://code.google.com/p/eagle-i/issues • Specific changes are to be tracked in an xls spreadsheet shared by curators
Refactoring Workflow • Identify Diffs: the ‘master-diff’ script lists all axioms from ERO import chain that are different/missing in ISF • Separate mireoted term diffs from core term diffs (ero/isf/vivo) for different curation workstreams • Workstream 1: Classify core diffs and resolve via manual or automated approaches • ERO-Deprecated Terms: maintain terms in a separate file outside of ISF import chain (ei-deprecated.owl) • New ERO Axioms:add axioms to ISF in separate file in the ISF import chain (isf_temp.owl) • ISF-Deprecated Axioms:move axioms into a separate file outside of the ISF import chain (ei_local.owl) • Different IRIs for same concepts: IRI mapping and merging is required to resolve IRI differences
Refactoring Workflow • Identify Diffs: the master-diff script lists all axioms from ERO import chain that are different/missing in ISF • Separate mireoted term axioms from core term axioms (ero/isf/vivo) for different curation workstreams • Workstream 1: Classify core diffs and resolve via manual or automated approaches • ERO-Deprecated Terms: ERO contains deprecated terms it originally donated to ISF • New ERO Axioms:ERO has created new terms/axioms not in the ISF • ISF-Deprecated Axioms: ISF has deprecated terms it originally imported from ERO • Different IRIs for same concepts: the same concepts are represented by different terms/modeling in ERO vs ISF
Refactoring Workflow • Identify Diffs: the ‘master-diff’ script lists all axioms from ERO import chain that are different/missing in ISF • Separate mireoted term diffs from core term diffs (ero/isf/vivo) for different curation workstreams • Workstream 1: Classify core diffs and resolve via manual or automated approaches • ERO-Deprecated Terms: maintain terms in a separate file outside of ISF import chain (ei-deprecated.owl) • New ERO Axioms:add axioms to ISF in separate file in the ISF import chain (isf_temp.owl) • ISF-Deprecated Axioms:move axioms into a separate file outside of the ISF import chain (ei_local.owl) • Different IRIs for same concepts: IRI mapping and merging is required to resolve IRI differences