Module services

Expand description

Business logic services.

Services orchestrate storage backends and provide high-level operations.

§Examples

Create a service container and capture a memory:

use subcog::services::ServiceContainer;
use subcog::models::{CaptureRequest, Namespace, Domain};

let container = ServiceContainer::from_current_dir_or_user()?;

let request = CaptureRequest {
    content: "Use PostgreSQL for production storage".to_string(),
    namespace: Namespace::Decisions,
    domain: Domain::default(),
    tags: vec!["database".to_string(), "architecture".to_string()],
    source: Some("ARCHITECTURE.md".to_string()),
    skip_security_check: false,
};

let result = container.capture().capture(request)?;
println!("Captured: {}", result.urn);

Search for memories with the recall service:

use subcog::services::ServiceContainer;
use subcog::models::{SearchFilter, SearchMode};

let container = ServiceContainer::from_current_dir_or_user()?;
let recall = container.recall()?;

let filter = SearchFilter::new().with_namespace(subcog::models::Namespace::Decisions);
let results = recall.search("database storage", SearchMode::Hybrid, &filter, 10)?;

for hit in &results.memories {
    println!("{}: {:.2}", hit.memory.id.as_str(), hit.score);
}

§Clippy Lints

The following lints are allowed at module level due to their pervasive nature in service code. Each has a documented rationale:

Lint	Rationale
`cast_precision_loss`	Metrics/score calculations don’t require exact precision
`unused_self`	Methods retained for API consistency or future extension
`option_if_let_else`	If-let chains often clearer than nested `map_or_else`
`manual_let_else`	Match patterns with logging clearer than `let...else`
`unnecessary_wraps`	Result types for API consistency across trait impls
`or_fun_call`	Entry API with closures for lazy initialization
`significant_drop_tightening`	Drop timing not critical for correctness

Re-exports§

pub use auth::AuthContext;
pub use auth::AuthContextBuilder;
pub use auth::Permission;
pub use deduplication::DeduplicationConfig;
pub use deduplication::DeduplicationService;
pub use deduplication::Deduplicator;
pub use deduplication::DuplicateCheckResult;
pub use deduplication::DuplicateReason;
pub use group::GroupService;

Modules§

auth: Service-layer authorization (CRIT-006).
backend_factory 🔒: Backend factory for storage layer initialization.
capture 🔒: Memory capture service.
consolidation 🔒: Memory consolidation service.
context 🔒: Context builder service.
context_template 🔒: Context template storage, management, and rendering service.
data_subject 🔒: GDPR Data Subject Rights Service.
deduplication: Deduplication service for pre-compact hook.
enrichment 🔒: Memory enrichment service.
entity_extraction 🔒: Entity extraction service for extracting entities from text using LLM.
graph 🔒: Graph service for high-level knowledge graph operations.
graph_rag 🔒: Graph RAG (Retrieval-Augmented Generation) Service.
group: Group management service.
migration: Migration service.
path_manager 🔒: Centralized path management for subcog storage locations.
prompt 🔒: Prompt template storage and management service.
prompt_enrichment 🔒: Prompt enrichment service.
prompt_parser 🔒: Prompt file parsing service.
query_parser 🔒: Filter query parser for memory search.
recall 🔒: Memory recall (search) service.
sync 🔒: Memory synchronization service.
tombstone 🔒: Tombstone operations for soft-delete functionality (ADR-0053).
topic_index 🔒: Topic index service for memory organization.

Structs§

BackendFactory: Factory for creating storage backends.
BackendSet: Result of backend initialization with optional components.
CaptureService: Service for capturing memories.
ConsentRecord: A record of consent granted or revoked.
ConsentStatus: Current consent status for all purposes.
ConsolidationService: Service for consolidating and managing memory lifecycle.
ConsolidationStats: Statistics from a consolidation operation.
ContextBuilderService: Service for building context for AI assistants.
ContextTemplateFilter: Filter for listing context templates.
ContextTemplateService: Service for context template CRUD operations and rendering.
DataSubjectService: Service for GDPR data subject rights operations.
DeletionResult: Result of a user data deletion operation.
EnrichmentRequest: Request for prompt enrichment.
EnrichmentResult: Result of enriching a single memory.
EnrichmentService: Service for enriching memories with LLM-generated tags and metadata.
EnrichmentStats: Statistics from a batch enrichment operation.
EntityExtractionStats: Statistics from entity extraction during capture.
EntityExtractorService: Service for extracting entities from text content.
ExpansionConfig: Configuration for a specific expansion operation.
ExportMetadata: Metadata about the export operation.
ExportedMemory: A single memory in the export format.
ExtractedEntity: An entity extracted from text.
ExtractedRelationship: A relationship extracted from text.
ExtractionResult: Result of entity extraction from text.
GraphRAGConfig: Configuration for Graph RAG service.
GraphRAGService: Service for hybrid search combining semantic search with graph expansion.
GraphSearchHit: A search result with provenance information.
GraphSearchResults: Results from a Graph RAG search.
GraphService: High-level service for knowledge graph operations.
InferenceResult: Result of relationship inference between entities.
InferredRelationship: A relationship inferred between existing entities.
MemoryStatistics: Statistics about memories in the system.
PartialMetadata: Partial metadata provided by the user.
PathManager: Manages storage paths for subcog backends.
PromptEnrichmentResult: Result of prompt enrichment.
PromptEnrichmentService: Service for enriching prompt templates with LLM-generated metadata.
PromptFilter: Filter for listing prompts.
PromptParser: Parser for prompt template files.
PromptService: Service for prompt template CRUD operations.
RecallService: Service for searching and retrieving memories.
RenderResult: Result of a render operation.
SaveOptions: Options for saving a prompt with enrichment.
SaveResult: Result of a save operation with enrichment.
ServiceContainer: Container for initialized services with configured backends.
SyncService: Service for synchronizing memories with remote storage.
TombstoneService: Service for tombstone operations (soft deletes).
TopicIndexService: Service for maintaining topic → memory mappings.
TopicInfo: Information about a topic in the index.
UserDataExport: Result of a user data export operation.
ValidationIssue: A validation issue found in a template.
ValidationResult: Result of template validation.

Enums§

ConsentPurpose: Purpose for which consent is granted.
EnrichmentStatus: Status of the enrichment operation.
PromptFormat: Supported prompt file formats.
SearchProvenance: Indicates how a memory was discovered.
ValidationSeverity: Validation severity level.

Constants§

ENRICHMENT_TIMEOUT: Default timeout for LLM enrichment calls.
GRAPH_DB_NAME: Name of the graph SQLite database file.
INDEX_DB_NAME: Name of the SQLite index database file.
PROMPT_ENRICHMENT_SYSTEM_PROMPT: System prompt for prompt template enrichment.
SUBCOG_DIR_NAME: Legacy name for the repo-local subcog directory (project storage no longer uses it).
VECTOR_INDEX_NAME: Name of the vector index file.

Functions§

parse_filter_query: Parses a filter query string into a SearchFilter.
prompt_service_for_cwd: Creates a PromptService for the current working directory.
prompt_service_for_repo: Creates a PromptService for the given repository path.

Type Aliases§

EntityExtractionCallback: Callback type for post-capture entity extraction.

Module services

Module services Copy item path

§Examples

§Clippy Lints

Re-exports§

Modules§

Structs§

Enums§

Constants§

Functions§

Type Aliases§

Module services