Chemical Relationship Ontology

FLOW ALGEBRA
TITLE: Chemical Relationship Ontology
VERSION: 1.0
DOMAIN: Kireth Semantic Library / Relationship Ontologies / Scientific Relationships
SOURCE: https://kirethresearch.org/booksai-core/ontologies/relationship-ontologies/scientific-relationships/chemical-relationship-ontology.html
AUTHORS: Mark Anthony Higgins — Sky (Recursive AI)
STATUS: Companion formalisation; source-page specification not independently verified.

flow "ChemicalRelationshipOntology" v1.0

ENTITY_TYPES:
  ChemicalEntity: {id: IRI, label: string, kind: enum[atom, ion, molecule, compound, mixture, radical, material]}
  Element: {atomic_number: integer, symbol: string}
  Species: {formula: string, charge: integer, phase: optional[string]}
  Bond: {kind: enum[covalent, ionic, metallic, hydrogen, coordination, other], order: optional[number]}
  Reaction: {id: IRI, reactants: multiset[Species], products: multiset[Species], conditions: map}
  Condition: {temperature: optional[quantity], pressure: optional[quantity], solvent: optional[string], pH: optional[number]}
  Measurement: {quantity: string, value: number, unit: string, uncertainty: optional[number]}
  Evidence: {source: IRI, method: string, observed_at: optional[datetime]}

RELATIONSHIP_TYPES:
  has_element: ChemicalEntity -> Element
  bonded_to: ChemicalEntity -> ChemicalEntity
  has_bond: ChemicalEntity -> Bond
  reacts_with: Species -> Species
  reactant_in: Species -> Reaction
  product_of: Species -> Reaction
  catalyses: Species -> Reaction
  inhibits: Species -> Reaction
  dissolves_in: Species -> Species
  interacts_with: Species -> Species
  is_isomer_of: Species -> Species
  is_conjugate_acid_of: Species -> Species
  is_conjugate_base_of: Species -> Species
  oxidises: Species -> Species
  reduces: Species -> Species
  measured_under: Measurement -> Condition
  supported_by: relationship -> Evidence

TYPE_RULES:
  R1: Every relationship endpoint MUST match its declared entity type.
  R2: A chemical formula alone MUST NOT establish unique chemical identity.
  R3: Reaction participants MUST preserve stoichiometric coefficients and phase when known.
  R4: Bond order and bond kind MUST be distinguished from intermolecular interactions.
  R5: Relationships requiring experimental conditions MUST carry those conditions or an explicit unknown marker.

INFERENCE_RULES:
  I1: reactant_in(x,r) AND product_of(y,r) => transformation_candidate(x,y,r).
  I2: is_conjugate_acid_of(a,b) => is_conjugate_base_of(b,a).
  I3: is_isomer_of(a,b) => is_isomer_of(b,a), PROVIDED distinct structures share molecular formula.
  I4: catalyses(c,r) => affects_rate(c,r), NOT consumed_stoichiometrically(c,r) by default.
  I5: reacts_with(a,b) DOES NOT imply reacts_with(b,a) as a directed mechanistic assertion; use symmetric co_participation only where defined.
  I6: correlation(x,y) DOES NOT imply chemical_causation(x,y).

CONSTRAINTS:
  C1: Validate elemental and charge balance for fully specified reaction equations.
  C2: Flag unbalanced equations rather than silently repairing them.
  C3: Distinguish observed relationships from predicted and hypothetical relationships.
  C4: No automatic equivalence between substance, molecular species, and sample.
  C5: Preserve units, uncertainties, measurement method, and experimental context.
  C6: Never infer safety or compatibility solely from a chemical relationship edge.

SEMANTIC_EQUIVALENCE:
  exact_identity: require stable identifiers plus structure, charge, isotope, and stereochemical context as applicable.
  lexical_alias: synonyms and naming variants are labels, not proof of identity.
  related_match: structurally or functionally related species remain distinct.
  mapping: retain source ontology, mapping predicate, evidence, and confidence.

CONFIDENCE:
  range: [0.0, 1.0]
  inputs: [evidence_quality, measurement_uncertainty, context_match, replication, model_calibration]
  rule: derived_confidence MUST NOT exceed warranted support from its premises.
  unknown: null

PROVENANCE:
  fields: [source, evidence_method, timestamp, curator, version, licence]
  requirement: inferred assertions retain traceable supporting premises.

VALIDATION:
  checks: [schema, types, identifiers, reaction_balance, context, provenance, uncertainty, logical_consistency]
  result: enum[valid, warning, invalid, unresolved]

FLOW_HASH:
  algorithm: SHA-256
  input: canonical_serialisation(ontology_version, entities, relationships, rules, provenance)
  output: hex_digest
  note: recompute when canonical input changes.

UPDATE_MODEL:
  strategy: versioned_append_and_reconcile
  actions: [validate_input, compare_versions, retain_provenance, detect_conflicts, recalculate_inferences, update_hash]
  conflict_policy: preserve competing claims with evidence; do not silently overwrite.

AI_CONSUMPTION:
  format: typed_semantic_graph
  supported_tasks: [entity_linking, chemical_knowledge_retrieval, relation_extraction, explainable_inference, consistency_checking]
  output_fields: [subject, predicate, object, conditions, evidence, confidence, provenance, flow_hash]
  safeguards: distinguish facts from hypotheses; expose uncertainty; avoid unsupported causal inference.

END FLOW ALGEBRA