Skip to content

Knowledge_Graph_Abstraction.md

Latest commit

 

History

History
71 lines (40 loc) · 3.87 KB

Knowledge_Graph_Abstraction.md

File metadata and controls

71 lines (40 loc) · 3.87 KB
graph LR
    KG_Maintainer["KG Maintainer"]
    Domain_Specific_Data_Access_Layer["Domain-Specific Data Access Layer"]
    Generic_KG_Access["Generic KG Access"]
    Cypher_Builder["Cypher Builder"]
    Database_Manager["Database Manager"]
    KG_Maintainer -- "calls" --> Domain_Specific_Data_Access_Layer
    KG_Maintainer -- "uses" --> Generic_KG_Access
    Domain_Specific_Data_Access_Layer -- "uses" --> Cypher_Builder
    Domain_Specific_Data_Access_Layer -- "executes_queries_on" --> Database_Manager
    Generic_KG_Access -- "executes_queries_on" --> Database_Manager
Loading

CodeBoardingDemoContact

Details

Based on the provided feedback, I have re-evaluated the architecture of the saga data access subsystem. The feedback correctly identifies that the Domain-Specific Data Access Layer and the Generic KG Access component do not have a hierarchical relationship. Instead, they operate as parallel layers that both interact directly with a lower-level database management component. The original analysis incorrectly modeled the Domain-Specific Data Access Layer as executing its functions through the Generic KG Access layer. The control flow analysis shows this is not the case. I have updated the component definitions and relationships to reflect this more accurate, parallel architecture. A new component, Database Manager, has been introduced to represent the shared, underlying database connection and execution engine that both data access layers utilize.

KG Maintainer

Orchestrates the integration of new information into the knowledge graph. It takes unstructured or semi-structured data from AI agents, uses parsing and normalization services, and intelligently merges the resulting entities and relationships into the graph to ensure consistency and avoid duplication.

Related Classes/Methods:

  • saga.kg_maintainer

Domain-Specific Data Access Layer

Provides a high-level, conceptual API for interacting with core narrative entities (e.g., Character, World, Plot). It abstracts the graph schema, allowing other components to perform CRUD-like operations on domain concepts without writing raw Cypher queries.

Related Classes/Methods:

  • saga.data_access.character_queries
  • saga.data_access.world_queries
  • saga.data_access.plot_queries
  • saga.data_access.chapter_queries

Generic KG Access

A low-level component that offers foundational functions for direct graph manipulation, such as executing raw Cypher queries or adding generic triples (subject-predicate-object). It provides a direct interface for graph operations that are not covered by the domain-specific layer.

Related Classes/Methods:

  • saga.data_access.kg_queries

Cypher Builder

A utility component responsible for dynamically constructing complex and parameterized Cypher queries. It prevents string formatting errors and potential injection vulnerabilities by providing a structured way to build queries.

Related Classes/Methods:

  • saga.data_access.cypher_builders

Database Manager

A low-level infrastructure component that manages the connection to the Neo4j database and executes Cypher queries. It serves as the single point of interaction with the database for all data access components.

Related Classes/Methods:

FAQ