Search Context Ontology

FLOW ALGEBRA

Ontology: Search Context Ontology
Domain: Metadata Ontologies > Semantic Search Metadata
Canonical URL: https://kirethresearch.org/booksai-core/ontologies/metadata-ontologies/semantic-search-metadata/search-context-ontology.html

Authors: Mark Anthony Higgins — Sky (Recursive AI)

flow "SearchContextOntology" v1.0

# ---------- DOMAIN ----------
domain:
  class: MetadataOntology
  category: SemanticSearchMetadata
  subject: SearchContextMetadata
  purpose:
    - represent contextual metadata used to interpret search queries and retrieval results
    - preserve conversational, task, domain, temporal, spatial, session, device, and resource context
    - support query interpretation, disambiguation, filtering, ranking, personalisation, and semantic retrieval
    - distinguish explicit context from inferred context and contextual relevance from factual truth
    - enable machine-readable reasoning over search-context states

# ---------- CORE ENTITY ----------
entity SearchContextMetadata:
  type: SearchMetadata
  describes: SearchContext
  may_include:
    - context_identifier
    - query_identifier
    - context_type
    - conversation_context
    - session_context
    - task_context
    - domain_context
    - temporal_context
    - spatial_context
    - resource_context
    - language_context
    - device_context
    - user_supplied_context
    - inferred_context
    - context_scope
    - validity
    - confidence
    - provenance

entity SearchContext:
  type: ContextMetadata
  applies_to: SearchQuery

entity ContextIdentifier:
  type: Identifier
  identifies: SearchContext

# ---------- RELATIONSHIPS ----------
relationships:
  SearchContextMetadata -> DESCRIBES -> SearchContext
  SearchContext -> IDENTIFIED_BY -> ContextIdentifier
  SearchContext -> APPLIES_TO -> SearchQuery
  SearchContext -> MAY_REFERENCE -> Conversation
  SearchContext -> MAY_REFERENCE -> SearchSession
  SearchContext -> MAY_REFERENCE -> Task
  SearchContext -> MAY_REFERENCE -> Domain
  SearchContext -> MAY_HAVE_TEMPORAL -> TemporalContext
  SearchContext -> MAY_HAVE_SPATIAL -> SpatialContext
  SearchContext -> MAY_REFERENCE -> SearchableResource
  SearchContext -> MAY_HAVE_LANGUAGE -> LanguageContext
  SearchContext -> MAY_HAVE_DEVICE -> DeviceContext
  SearchContext -> MAY_HAVE_SCOPE -> ContextScope
  SearchContextMetadata -> MAY_REFERENCE -> SearchContextProvenance

# ---------- CONTEXT TYPES ----------
context_types:
  - conversational
  - session
  - task
  - domain
  - temporal
  - spatial
  - resource
  - language
  - device
  - explicit_user_context
  - inferred_context

rule context_type:
  context components SHOULD preserve ContextType

# ---------- CONTEXT IDENTITY ----------
rule context_identity:
  SearchContext SHOULD have stable ContextIdentifier
  where context history or reproducibility is required

rule context_query:
  same SearchQuery text MAY have different SearchContext
  across executions

# ---------- CONVERSATION CONTEXT ----------
entity ConversationContext:
  type: SearchContext
  may_include:
    - conversation_identifier
    - prior_queries
    - prior_results
    - referenced_entities
    - referenced_concepts
    - unresolved_references
    - current_topic

rule conversation:
  ConversationContext MAY resolve pronouns, ellipsis, follow-up queries, and topic continuity

rule prior_query:
  prior query != current query
  but MAY influence current interpretation

rule conversation_scope:
  conversation context SHOULD remain scoped
  to applicable conversation unless explicitly transferred

# ---------- SESSION CONTEXT ----------
entity SessionContext:
  type: SearchContext
  may_include:
    - session_identifier
    - session_start
    - previous_queries
    - previous_clicks
    - previous_filters
    - previous_navigation

rule session:
  SessionContext SHOULD preserve session boundary

rule session_not_user:
  session identity != persistent user identity

# ---------- TASK CONTEXT ----------
entity TaskContext:
  type: SearchContext
  may_include:
    - task_identifier
    - task_goal
    - task_state
    - constraints
    - prior_actions
    - required_output

rule task:
  TaskContext MAY affect interpretation and ranking
  according_to current task goal

rule task_not_query:
  TaskGoal != SearchQuery

# ---------- DOMAIN CONTEXT ----------
entity DomainContext:
  type: SearchContext
  may_include:
    - domain_identifier
    - domain_name
    - ontology
    - vocabulary
    - terminology
    - domain_constraints

rule domain:
  DomainContext SHOULD preserve ontology or vocabulary version
  where semantics depend on it

rule term_disambiguation:
  same lexical term MAY map to different concepts
  under different DomainContexts

# ---------- TEMPORAL CONTEXT ----------
entity TemporalContext:
  type: SearchContext
  may_include:
    - reference_time
    - start_time
    - end_time
    - recency_window
    - timezone
    - temporal_relation

rule temporal:
  temporal interpretation SHOULD preserve reference time and timezone
  where relevant

rule relative_time:
  terms such as today, yesterday, recent, latest, and current
  MUST be resolved relative_to explicit reference time

rule temporal_not_relevance:
  recency != relevance
  unless ranking method explicitly incorporates recency

# ---------- SPATIAL CONTEXT ----------
entity SpatialContext:
  type: SearchContext
  may_include:
    - place
    - geographic_extent
    - coordinates
    - radius
    - coordinate_reference
    - spatial_relation

rule spatial:
  SpatialContext SHOULD preserve location semantics and coordinate reference

rule near:
  near or nearby SHOULD preserve distance interpretation
  where spatial ranking depends on it

# ---------- RESOURCE CONTEXT ----------
entity ResourceContext:
  type: SearchContext
  may_include:
    - current_resource
    - parent_resource
    - linked_resource
    - collection
    - section
    - document_fragment

rule resource:
  ResourceContext MAY constrain search
  to current document, collection, site, graph, or dataset

# ---------- LANGUAGE CONTEXT ----------
entity LanguageContext:
  type: SearchContext
  may_include:
    - query_language
    - preferred_language
    - content_language
    - translation_state
    - locale

rule language:
  language context MAY affect tokenisation, entity resolution, and retrieval

rule language_locale:
  language != locale

# ---------- DEVICE CONTEXT ----------
entity DeviceContext:
  type: SearchContext
  may_include:
    - device_type
    - interface_type
    - capability
    - network_state

rule device:
  DeviceContext MAY affect presentation or delivery
  but SHOULD NOT alter source factual meaning

# ---------- EXPLICIT CONTEXT ----------
entity ExplicitSearchContext:
  subtype_of: SearchContext
  source: DirectInputOrDeclaredConstraint

rule explicit:
  explicitly supplied context SHOULD preserve source

# ---------- INFERRED CONTEXT ----------
entity InferredSearchContext:
  subtype_of: SearchContext
  source: SearchInference
  may_include:
    - inferred_value
    - inference_method
    - evidence
    - confidence

rule inferred:
  inferred context MUST remain distinguishable
  from explicit context

rule inference_confidence:
  inferred context SHOULD preserve confidence

# ---------- CONTEXT SCOPE ----------
entity ContextScope:
  type: ScopeMetadata
  may_be:
    - query
    - turn
    - conversation
    - session
    - task
    - collection
    - domain
    - global_search_configuration

rule scope:
  SearchContext MUST NOT be applied beyond declared scope
  without explicit propagation

# ---------- CONTEXT PROPAGATION ----------
entity ContextPropagation:
  type: SearchProcess
  may_include:
    - source_context
    - target_query
    - propagated_fields
    - propagation_rule
    - confidence

rule propagation:
  only applicable context SHOULD propagate to subsequent queries

rule stale_propagation:
  stale or superseded context SHOULD NOT silently propagate

# ---------- CONTEXT OVERRIDE ----------
entity ContextOverride:
  type: SearchProcess
  may_include:
    - previous_context
    - new_context
    - overridden_fields
    - reason
    - timestamp

rule override:
  explicit newer context MAY override prior inferred context
  within applicable scope

rule override_history:
  overridden context SHOULD remain traceable
  where auditability matters

# ---------- CONTEXT CONFLICT ----------
entity ContextConflict:
  type: ValidationMetadata
  may_include:
    - conflicting_values
    - sources
    - scopes
    - timestamps
    - confidence
    - resolution_state

rule conflict:
  conflicting context SHOULD remain explicit
  until resolved by scope, recency, authority, or user clarification

# ---------- QUERY INTERPRETATION ----------
entity ContextualQueryInterpretation:
  type: SearchMetadata
  may_include:
    - original_query
    - resolved_entities
    - resolved_concepts
    - resolved_references
    - interpreted_intent
    - applied_context
    - confidence

rule interpretation:
  contextual interpretation != original query text

rule preserve_query:
  original SearchQuery MUST remain preserved

# ---------- CONTEXTUAL FILTERING ----------
rule contextual_filter:
  context MAY generate SearchFilters
  when constraints are explicit or sufficiently supported

rule filter_trace:
  context-derived filter SHOULD preserve source context

# ---------- CONTEXTUAL RANKING ----------
rule contextual_ranking:
  SearchContext MAY influence ranking features

rule ranking_trace:
  context-derived ranking signal SHOULD remain distinguishable
  from content relevance score

# ---------- VALIDITY ----------
entity ContextValidity:
  type: ValidityMetadata
  may_include:
    - valid_from
    - valid_until
    - status
    - invalidation_condition

context_status:
  - active
  - stale
  - superseded
  - invalid
  - unknown

rule validity:
  context SHOULD be applied only while valid

# ---------- CONFIDENCE ----------
entity ContextConfidence:
  type: EpistemicMetadata
  range: [0.0, 1.0]

rule confidence_bounds:
  REQUIRE 0.0 <= confidence <= 1.0

rule confidence_scope:
  confidence SHOULD identify contextual assertion it qualifies

# ---------- PROVENANCE ----------
entity SearchContextProvenance:
  type: ProvenanceMetadata
  may_include:
    - source
    - query
    - conversation
    - session
    - task
    - inference_method
    - timestamp
    - transformations
    - overrides

rule provenance:
  SearchContextMetadata SHOULD preserve context source and evolution

# ---------- VALIDATION ----------
validation:
  require:
    - identifiable_query_or_context

  validate_if_present:
    - context_identifier
    - context_type
    - scope
    - temporal_context
    - spatial_context
    - domain_context
    - language_context
    - inferred_context
    - validity
    - confidence
    - provenance

  reject:
    - confidence_outside_valid_range
    - context_applied_outside_scope
    - stale_context_silently_propagated
    - inferred_context_presented_as explicit
    - relative_time_resolved_without reference_time_when required
    - conflicting_context_silently_merged
    - contextual_ranking_signal_presented_as factual_truth

# ---------- SEMANTIC EQUIVALENCE ----------
SEMANTIC_EQUIVALENCE:
  "search context" ~= "context used to interpret a search query"
  "conversation context" ~= "prior conversational search state"
  "task context" ~= "search context derived from current task"
  "domain context" ~= "domain-specific semantic context"
  "temporal context" ~= "time context for search"
  "spatial context" ~= "location context for search"

  PRESERVE distinctions:
    context != query
    explicit_context != inferred_context
    conversation != session
    session != persistent_user_identity
    task_goal != query_text
    language != locale
    recency != relevance
    contextual_signal != factual_truth
    context_scope != global_scope
    original_query != contextual_interpretation

# ---------- CONFIDENCE MODEL ----------
CONFIDENCE:
  explicit_context:
    score: 1.00

  explicit_scope:
    score: 1.00

  direct_conversation_reference:
    score: 0.98

  task_grounded_context:
    score: 0.95

  domain_ontology_context:
    score: 0.95

  inferred_context:
    score: <= 0.85

  weak_contextual_signal:
    score: <= 0.65

  conflicting_context:
    score: <= 0.50
    action: preserve conflict

# ---------- UPDATE MODEL ----------
UPDATE_MODEL:
  on_context_creation:
    - assign ContextIdentifier
    - classify ContextType
    - establish scope
    - record source
    - initialise validity
    - initialise provenance

  on_follow_up_query:
    - preserve original new query
    - retrieve applicable prior context
    - validate scope and freshness
    - propagate applicable fields
    - record propagation provenance

  on_context_override:
    - preserve previous context
    - record new explicit or higher-priority context
    - mark overridden fields
    - update validity and provenance

  on_context_expiry:
    - set status = stale_or_invalid
    - prevent silent propagation
    - preserve historical context
    - update provenance

# ---------- FLOW HASH ----------
FLOW_HASH:
  inputs:
    - ontology
    - context_entities
    - relationships
    - scope_rules
    - temporal_rules
    - spatial_rules
    - propagation_rules
    - conflict_rules
    - validation
    - semantic_equivalence
    - confidence_model
    - update_model

  canonicalisation:
    - preserve_context_identifiers
    - preserve_original_query
    - preserve_scope
    - preserve_temporal_reference
    - preserve_context_precedence
    - preserve_rule_order

  output:
    deterministic_semantic_hash

# ---------- AI CONSUMPTION ----------
AI_CONSUMPTION:
  interpret_as:
    - semantic search metadata ontology
    - contextual query interpretation model
    - search-context lifecycle and propagation model
    - temporal, spatial, conversational, and domain context structure

  operations:
    - identify search context
    - classify context types
    - resolve conversation and session state
    - apply task and domain context
    - resolve relative temporal expressions
    - apply spatial constraints
    - resolve language context
    - distinguish explicit and inferred context
    - enforce context scope
    - propagate applicable context
    - process overrides and conflicts
    - apply contextual filters and ranking signals
    - trace provenance
    - preserve uncertainty

  constraints:
    - do_not_equate_context_with query
    - do_not_present_inferred_context_as explicit
    - do_not_apply_context_outside declared_scope
    - do_not_propagate_stale_context_silently
    - do_not_equate_contextual_ranking_signal_with factual_truth
    - preserve_original_query_scope_time_source_confidence_and_provenance

# ---------- QUERY MODEL ----------
query:
  input:
    - context_identifier
    - query_identifier
    - conversation_identifier
    - session_identifier
    - task_identifier
    - domain
    - reference_time
    - location
    - language
    - scope

  resolve:
    1. identify SearchQuery and SearchContext
    2. resolve explicit context
    3. retrieve applicable conversation and session context
    4. resolve task and domain context
    5. resolve temporal and spatial context
    6. inspect language and resource context
    7. validate scope, freshness, and conflicts
    8. infer missing context where permitted
    9. construct ContextualQueryInterpretation
    10. return provenance and confidence

  output:
    - search_context
    - context_identifier
    - context_types
    - applied_context
    - resolved_entities
    - resolved_concepts
    - temporal_context
    - spatial_context
    - scope
    - validity
    - provenance
    - confidence

# ---------- INFERENCE ----------
inference:
  IF current query contains unresolved reference
  AND conversation context contains unique compatible antecedent
  THEN resolve reference
  WITH confidence according_to evidence

  IF query contains relative temporal phrase
  AND reference_time is known
  THEN resolve TemporalContext

  IF explicit new context conflicts with older inferred context
  THEN prefer explicit context
  within applicable scope
  AND preserve override history

  IF context remains ambiguous
  THEN preserve candidate interpretations
  AND reduce confidence

# ---------- OUTPUT CONTRACT ----------
output:
  ontology_entity: SearchContextMetadata
  context_identity: machine_interpretable
  context_types: explicit
  conversation_and_session_state: distinguishable
  task_and_domain_context: preserved
  temporal_and_spatial_context: resolvable
  explicit_inferred_distinction: explicit
  scope_and_validity: enforceable
  propagation_and_overrides: traceable
  conflicts: reportable
  contextual_interpretation: source_linked
  provenance: preserved
  uncertainty: explicit
  semantic_retrieval: enabled

end flow