Files
feedkit/normalize/registry.go
Eric Rakestraw a6c133319a feat(feedkit): add optional normalization hook and document external API
Introduce an optional normalization stage for feedkit pipelines via the new
normalize package. This adds:

- normalize.Normalizer interface with flexible Match() semantics
- normalize.Registry for ordered normalizer selection (first match wins)
- normalize.Processor adapter implementing pipeline.Processor
- Pass-through behavior when no normalizer matches (normalization is optional)
- Func helper for ergonomic normalizer definitions

Update root doc.go to fully document the normalization model, its role in the
pipeline, recommended conventions (Schema-based matching, raw vs normalized
events), and concrete wiring examples. The documentation now serves as a
complete external-facing API specification for downstream daemons such as
weatherfeeder.

This change preserves feedkit’s non-framework philosophy while enabling a
clean separation between data collection and domain normalization.
2026-01-13 18:23:43 -06:00

141 lines
3.5 KiB
Go

package normalize
import (
"context"
"fmt"
"sync"
"gitea.maximumdirect.net/ejr/feedkit/event"
)
// Registry holds a set of Normalizers and selects one for a given event.
//
// Selection rule (simple + predictable):
// - iterate in registration order
// - the FIRST Normalizer whose Match(e) returns true is used
//
// If none match, the event passes through unchanged.
//
// Why "first match wins"?
// Normalization is usually a single mapping step from a raw schema/version into
// a normalized schema/version. If you want multiple transformation steps,
// model them as multiple pipeline processors (which feedkit already supports).
type Registry struct {
mu sync.RWMutex
ns []Normalizer
}
// Register adds a normalizer to the registry.
//
// Register panics if n is nil; this is a programmer error and should fail fast.
func (r *Registry) Register(n Normalizer) {
if n == nil {
panic("normalize.Registry.Register: normalizer cannot be nil")
}
r.mu.Lock()
defer r.mu.Unlock()
r.ns = append(r.ns, n)
}
// Normalize finds the first matching Normalizer and applies it.
//
// If no normalizer matches, it returns the input event unchanged.
//
// If a normalizer returns (nil, nil), the event is dropped.
func (r *Registry) Normalize(ctx context.Context, in event.Event) (*event.Event, error) {
if r == nil {
// Nil registry is a valid "feature off" state.
out := in
return &out, nil
}
r.mu.RLock()
ns := append([]Normalizer(nil), r.ns...) // copy for safe iteration outside lock
r.mu.RUnlock()
for _, n := range ns {
if n == nil {
// Shouldn't happen (Register panics), but guard anyway.
continue
}
if !n.Match(in) {
continue
}
out, err := n.Normalize(ctx, in)
if err != nil {
return nil, fmt.Errorf("normalize: normalizer failed: %w", err)
}
// out may be nil to signal "drop".
return out, nil
}
// No match: pass through unchanged.
out := in
return &out, nil
}
// Processor adapts a Registry into a pipeline Processor.
//
// It implements:
//
// Process(ctx context.Context, in event.Event) (*event.Event, error)
//
// which matches feedkit/pipeline.Processor.
//
// Optionality:
// - If Registry is nil, Processor becomes a no-op pass-through.
// - If Registry has no matching normalizer for an event, that event passes through unchanged.
type Processor struct {
Registry *Registry
// If true, events that do not match any normalizer cause an error.
// Default is false (pass-through), which is the behavior you asked for.
RequireMatch bool
}
// Process implements the pipeline.Processor interface.
func (p Processor) Process(ctx context.Context, in event.Event) (*event.Event, error) {
// "Feature off": no registry means no normalization.
if p.Registry == nil {
out := in
return &out, nil
}
out, err := p.Registry.Normalize(ctx, in)
if err != nil {
return nil, err
}
if out == nil {
// Dropped by normalization policy.
return nil, nil
}
if p.RequireMatch {
// Detect "no-op pass-through due to no match" by checking whether a match existed.
// We do this with a cheap second pass to avoid changing Normalize()'s signature.
// (This is rare to enable; correctness/clarity > micro-optimization.)
if !p.Registry.hasMatch(in) {
return nil, fmt.Errorf("normalize: no normalizer matched event (id=%s kind=%s source=%s schema=%q)",
in.ID, in.Kind, in.Source, in.Schema)
}
}
return out, nil
}
func (r *Registry) hasMatch(in event.Event) bool {
if r == nil {
return false
}
r.mu.RLock()
defer r.mu.RUnlock()
for _, n := range r.ns {
if n != nil && n.Match(in) {
return true
}
}
return false
}