Documentation
¶
Overview ¶
Package ingest declares the core port used to turn a source corpus into a validated knowledge graph. Implementations live in driven adapters (for example internal/graphify); the tenant domain depends only on this contract.
Index ¶
Constants ¶
This section is empty.
Variables ¶
This section is empty.
Functions ¶
This section is empty.
Types ¶
type AddOptions ¶
type AddOptions struct {
URL string
TargetDir string
Author string
Contributor string
WorkingDir string
Timeout time.Duration
}
AddOptions describes one public source to materialize in a corpus.
type Engine ¶
type Engine interface {
Preflight(context.Context) (string, error)
Add(context.Context, AddOptions) (AddResult, error)
Extract(context.Context, ExtractOptions) (GraphStats, Result, error)
}
Engine is the port driven by tenant bootstrap.
type ExtractOptions ¶
type ExtractOptions struct {
CorpusDir string
OutputDir string
WorkingDir string
Backend string
Model string
Mode string
MaxWorkers int
TokenBudget int
MaxConcurrency int
APITimeout time.Duration
Timeout time.Duration
Force bool
NoCluster bool
// Env contains per-attempt credentials and routing overrides. Adapters must
// pass it only to the child engine process and must not persist or log it.
Env []string
}
ExtractOptions controls a headless semantic extraction.
type GraphStats ¶
type GraphStats struct {
Path string `json:"path"`
Nodes int `json:"nodes"`
Edges int `json:"edges"`
}
GraphStats is the minimal machine-readable health view of a graph.
Click to show internal directories.
Click to hide internal directories.