Documentation
¶
Index ¶
Constants ¶
This section is empty.
Variables ¶
This section is empty.
Functions ¶
Types ¶
type Config ¶
type Config struct {
EmbeddingProvider string `json:"embedding_provider"` // embedding provider name (e.g., "openai")
EmbeddingModel string `json:"embedding_model"` // embedding model name (e.g., "text-embedding-3-small")
Dimensions int `json:"dimensions"` // vector dimensions (default 1536)
DefaultCollection string `json:"default_collection"` // default collection to create on init (optional)
}
Config controls pgvector plugin behavior.
type Plugin ¶
type Plugin struct {
// contains filtered or unexported fields
}
Plugin implements vector similarity search using the pgvector PostgreSQL extension. Workflows can store and query embedding vectors with metadata.
func (*Plugin) Info ¶
func (p *Plugin) Info() plugin.PluginInfo
Info returns plugin metadata for discovery and documentation.
func (*Plugin) Init ¶
Init initializes the plugin with the given environment. It parses optional configuration, ensures the pgvector extension is installed, and optionally creates a default collection.
func (*Plugin) Migrations ¶
Migrations returns the database schema for pgvector collections and embeddings. Tables are idempotent (IF NOT EXISTS) and safe to run multiple times.
func (*Plugin) RegisterHostFunctions ¶
func (p *Plugin) RegisterHostFunctions(scope plugin.FuncRegistry) error
RegisterHostFunctions registers workflow-callable functions on the scoped function registry. The plugin name is implicit -- each plugin gets its own scope, so function names need not be globally unique.
Source Files