Schema clean-up: chemdcat-ap inheritance, quality audit, catalyst/reaction classification, real-world example datasets - #120
Merged
HendrikBorgelt merged 7 commits intoJul 14, 2026
Conversation
…esolve it CatalyticReaction now specializes chemdcat-ap's ChemicalReaction instead of re-declaring EvaluatedActivity in parallel, and ChemicalReactor specializes Reactor instead of Device -- both previously asserted the same class_uri as their chemdcat-ap counterpart without an is_a relationship. The custom Excel/docs tooling (schema_to_excel.py, inbox_to_schema.py, generate_schema_docs.py) hand-rolls its own schema merge and previously only understood mixins, not is_a, so it was blind to inherited chemdcat-ap slots. Teaches it to climb is_a chains for slot resolution, mark inherited rows as read-only reference material in the Excel workbook, and skip non-owned slots in the inbox workflow's editability and deletion-detection checks so inherited fields don't trigger false "will be removed" warnings.
Fixes cross-contaminated ontology mappings (Catalyst/ChemicalProduct exact_mappings pointing at each other's terms, an overly-specific Duration class_uri, a wrong lense_mode slot_uri, HeatingRate's class_uri), sweeps 295 slots for correct is_a relation-typing, fixes multivalued anti-patterns on booleans/counts/min-max pairs, and resolves the reactant/used_reactant overlap left over from the chemdcat-ap inheritance change. Also fixes a self-referential recursion bug in the sunburst chart generator and adds the new precipitating_concentration/precursor_quantity CURIE fixes. Test fixtures updated to match the corrected slot names/ranges.
…tion Adds operating-condition slots for each of the 8 ChemicalReactor subclasses (ElectrochemicalReactor, CSTR, PlugFlowReactor, Autoclave, SlurryReactor, Microreactor, FixedBedReactor, FluidizedBedReactor), and introduces catalyst_type (CatalysisResearchFieldEnum) + catalyst_form (CatalystFormEnum) as two independent slots on CatalyticReaction, replacing an earlier single overloaded field. Reaction type classification stays on rdf_type via the existing DCAT-AP-PLUS Pattern 3 convention (already used by Synthesis, Characterization, Simulation, and CatalysisDataset), rather than adding a dedicated slot. catalyst_type/catalyst_form and reaction rdf_type are both kept `recommended` rather than `required`: catalyst/reaction classification can be genuinely disputed or not yet covered by the controlled vocabulary for a novel case, and a mandatory field would force a premature or contested classification. See the rationale recorded directly on these slots below for the fuller argument, including why this two-slot design is kept instead of a CatalystType/ReactionType class hierarchy also explored elsewhere.
Adds a dedicated catalysis-dataset.md page documenting the CatalysisDataset model and its rdf_type classification pattern. Rewrites the stale design-patterns.md content and adds a section on the new chemdcat-ap inheritance chain. Fixes the contributing.md gap around the Excel-inbox edit-and-PR workflow (previously undocumented, pushing contributors to GitHub issues only) and aligns inbox/README.md with it.
Adds 3 fully-worked CatalysisDataset records built from real experimental data (mixed CO/CO2 methanation, citral hydrogenation, carbonylation chemistry), validated end-to-end against the schema and published as docs/assets/examples for reference. Adds generate_example_outputs.py to publish them during doc generation, and updates the CI workflow to run `just test` (schema + pytest + linkml-run-examples) and upload the example validation output as a build artifact.
Fixes CURIEs for precursor_quantity and adds the new precipitating_concentration slot (nfdi4cat#109, nfdi4cat#59, nfdi4cat#34). Renames "CatCore" to "CoreMeta4Cat" and aligns the vocabulary workbook's sheet-description table with its actual current sheets (nfdi4cat#111, nfdi4cat#53). Generalizes "Voc4Cat term" wording to "CURIE" since the schema also uses CHMO/QUDT/OBI/SIO/NCIT/AFE terms (nfdi4cat#110). Documents the Excel workbook's structure/colour-coding/columns and the inbox PR workflow in detail, with real workbook screenshots and the automated check's actual validation messages (nfdi4cat#112). Regenerates all derived artifacts (Python dataclasses/pydantic models, generated schema docs, Excel workbook, sunburst charts) from the final schema state.
…rtifacts Rebasing onto upstream/main surfaced additional duplication beyond the already-resolved coremeta4cat_reaction_ap.yaml conflict: PR nfdi4cat#118 also added has_cathode/has_anode/has_cell_operating_mode/has_active_area/ has_faradaic_current and several catalyst-bed/stirrer geometry slots to coremeta4cat_common.yaml that duplicate slots already defined locally in coremeta4cat_reaction_ap.yaml (same concepts, same VOC4CAT terms in most cases, different slot names), plus has_conversion/has_space_time_yield/ has_selectivity referencing nfdi4cat#118's Conversion/SpaceTimeYield/Selectivity classes -- which aren't included in this PR (see the PR description's note on keeping those for separate discussion). Removes the now-orphaned duplicates so the schema resolves cleanly, and regenerates all derived artifacts against the final merged state.
2 tasks
2 tasks
This file contains hidden or 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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
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.
closes #109
closes #59
closes #34
closes #111
closes #53
closes #110
closes #112
Summary
Schema clean-up and quality pass across CoreMeta4Cat: proper LinkML inheritance from chemdcat-ap (
CatalyticReaction is_a: ChemicalReaction,ChemicalReactor is_a: Reactor), a full modelling/mapping audit (mismatched ontology mappings,is_arelation-typing sweep across 295 slots, multivalued anti-patterns), catalysis-specific reactor slots, three fully worked real-world example datasets validated as test data, and expanded documentation (inheritance, catalysis-dataset model, working-with-data guide, contributing/getting-started fixes).This also touches the same area as PR #118 (catalyst/reaction classification) — see "On the catalyst_type / has_catalyst_type conflict" below before reviewing the reaction_ap changes.
On the catalyst_type / has_catalyst_type conflict
This branch and the already-merged #118 independently redesigned catalyst/reaction classification while diverging from the same base. This PR keeps this branch's design —
catalyst_type(CatalysisResearchFieldEnum) +catalyst_form(CatalystFormEnum) as two slots, andrdf_type(DCAT-AP-PLUS Pattern 3) for reaction type — instead of #118'shas_catalyst_type: CatalystType/has_reaction_type: ReactionTypeclass hierarchies. Reasoning:CatalystTypeconflates two independent axes. Its subclassesHeterogeneousCatalyst/HomogeneousCatalyst/BioCatalyst/ElectroCatalyst/PhotoCatalyst(catalytic regime) sit as siblings ofThinFilmCatalyst/BulkCatalyst/PowerderedCatalyst/DepositedSampleCatalyst/SupportedCatalsyt(physical form) in one flat hierarchy. These aren't mutually exclusive — a catalyst is routinely both "heterogeneous" and "supported" at once. The five form-related subclasses use VOC4CAT ids identical to this branch'sCatalystFormEnumpermissible values (0000019/0007015/0000017/0000038/0007034), confirming both sides are describing the same underlying vocabulary, just merged into one axis on Added new classes & slots in coremeta4cat_reaction_ap.yaml #118's side. The practical effect shows up in Added new classes & slots in coremeta4cat_reaction_ap.yaml #118's own test data:has_catalyst_type: [{value: "heterogeneous, supported metal oxide"}]— free text stuffed into a genericvaluefield, which defeats the purpose of having ten typed subclasses.has_reaction_typeduplicates an existing mechanism.CatalyticReaction.rdf_typealready carries reaction-type classification via Pattern 3 (the same convention used bySynthesis,Characterization,Simulation, andCatalysisDatasetthroughout this schema) and is alreadyrecommended. Added new classes & slots in coremeta4cat_reaction_ap.yaml #118'sReactionTypeclass_uri (VOC4CAT:0007010) is the same URI already documented as the example value forrdf_typehere — same concept, two competing mechanisms for asserting it.Not claiming #118 was wrong to attempt — it was explicitly a stopgap while this work was in progress. Happy to walk through the reactor-slot and
ReactorPerformanceMeasures/Conversion/SpaceTimeYield/Selectivityadditions from #118 separately; those don't conflict with this design and may be worth keeping.On #116 / #117 (cardinality of catalyst_type / reaction type)
Deliberately left
catalyst_typeand reactionrdf_typeasrecommendedrather thanrequired, per #117's request. In research contexts, classification can be genuinely disputed (a researcher may argue a catalyst is type A, not type B) or the catalyst may not fit any currentCatalysisResearchFieldEnumvalue yet — forcing a mandatory value would push contributors into a premature or contested classification rather than leaving the field open until vocabulary/consensus catches up. Rationale is also recorded as schema comments next to both slots incoremeta4cat_reaction_ap.yaml. Open to discussion in a follow-up issue if a different tradeoff is preferred — flagging #116 as the place to continue that.Other issues addressed in this PR
precursor_quantity, newprecipitating_concentrationslot)Already resolved by earlier work on this branch, no action taken: #112's original ask (contributing.md inbox description) was already covered before this PR; #53's Excel-tab-naming complaint no longer applies (sheets are
Synthesis/Characterization/Reaction/Simulation/CoreMeta4Catnow, not the old split names). #42 will be answered directly on the issue (not a schema change). #69 (broken download link) and #115 (docs home-page PR) are unrelated to this branch's scope.Test plan
just test(schema generation, pytest,linkml-run-examplesagainsttests/data) — all greenCatalysisDatasetrecords validate end-to-endjust gen-docregenerated cleanly, no orphaned references