Skip to content

Improve/Impl (graph) property based vectorization #228

@imbajin

Description

@imbajin

Currently, our current vectorized logic is as follows:

  1. text vector (raw)
  2. graph vector
    • vid(vertexID) based vector
    • property based vector
graph TD
    A["Text (raw)"] --> B("Text Vector - chunk");
    A --> C("Graph - Vertices/Edges extract from Text");
    C --> D("Graph Vector");
    D --> E("VID based Vector - Relatively complete");
    D --> F("Property based Vector - Primitive, needs improvement");

    %% Style to highlight the primitive part
    classDef primitive fill:#f9f,stroke:#333,stroke-width:2px;
    class F primitive;
Loading

The vectorization of vid is relatively complete, but the vectorization based on vertex/edge property is still relatively primitive and needs to be improved in its implementation and application(Could search in the code for details)

Image

Metadata

Metadata

Assignees

Labels

enhancementNew feature or request

Type

Projects

No projects

Milestone

No milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions