knowledge-governance/internal/workflows/sourcescan/handler.go

package sourcescan

import (
	"context"
	"encoding/json"
	"errors"
	"sort"
	"time"

	ppconnector "knowledge-governance/internal/connectors/pp"
	"knowledge-governance/internal/governance"
	"knowledge-governance/internal/jobs"
)

type PPScanner interface {
	Scan(context.Context) (ppconnector.Snapshot, string, error)
}

type PPScanStore interface {
	StartSourceScan(context.Context, string, time.Time) (governance.SourceScan, error)
	LoadPPCurrentItems(context.Context, string) (map[string]ppconnector.CurrentItem, error)
	CompleteSourceScan(context.Context, string, governance.SourceScanResult, time.Time) error
	FailSourceScan(context.Context, string, string, time.Time) error
}

type PPHandlerConfig struct {
	Store   PPScanStore
	Scanner PPScanner
	Now     func() time.Time
}

type PPHandler struct {
	store   PPScanStore
	scanner PPScanner
	now     func() time.Time
}

func NewPPHandler(config PPHandlerConfig) (*PPHandler, error) {
	if config.Store == nil || config.Scanner == nil {
		return nil, errors.New("PP scan store and scanner are required")
	}
	if config.Now == nil {
		config.Now = time.Now
	}
	return &PPHandler{store: config.Store, scanner: config.Scanner, now: config.Now}, nil
}

func (h *PPHandler) Handle(ctx context.Context, job jobs.Job) error {
	var payload struct {
		ScanID string `json:"scan_id"`
	}
	if err := json.Unmarshal(job.Payload, &payload); err != nil || payload.ScanID == "" {
		return errors.New("invalid source scan payload")
	}
	now := h.now()
	scan, err := h.store.StartSourceScan(ctx, payload.ScanID, now)
	if err != nil {
		return err
	}
	if scan.State == governance.SourceScanCompleted {
		return nil
	}
	if scan.State != governance.SourceScanRunning {
		return errors.New("source scan is not running")
	}
	if scan.SourceType != governance.SourceTypePP {
		return h.store.FailSourceScan(ctx, scan.ID, "source_type_unsupported", now)
	}
	snapshot, snapshotSHA, err := h.scanner.Scan(ctx)
	if err != nil || !governance.ValidSHA256(snapshotSHA) {
		return h.store.FailSourceScan(ctx, scan.ID, "pp_snapshot_invalid", now)
	}
	current, err := h.store.LoadPPCurrentItems(ctx, scan.SourceConnectionID)
	if err != nil {
		return err
	}
	reconciliation, err := ppconnector.PlanReconciliation(snapshot, current)
	if err != nil {
		return h.store.FailSourceScan(ctx, scan.ID, "pp_snapshot_invalid", now)
	}
	observed := make(map[string]struct{}, len(snapshot.Items))
	changes := make([]governance.SourceScanChange, 0)
	for _, item := range snapshot.Items {
		observed[item.SourceURLHash] = struct{}{}
		if item.Status == ppconnector.StatusOverLimit {
			continue
		}
		size := item.Bytes
		stored, exists := current[item.SourceURLHash]
		if !exists {
			changes = append(changes, governance.SourceScanChange{
				StableExternalID: item.SourceURLHash, ChangeType: governance.SourceScanAdded,
				ObservedRevision: item.SHA256, ObservedSizeBytes: &size,
			})
		} else if stored.Revision != item.SHA256 {
			changes = append(changes, governance.SourceScanChange{
				StableExternalID: item.SourceURLHash, ChangeType: governance.SourceScanUpdated,
				PreviousRevision: stored.Revision, ObservedRevision: item.SHA256, ObservedSizeBytes: &size,
			})
		}
	}
	missingIDs := make([]string, 0)
	for sourceID := range current {
		if _, ok := observed[sourceID]; !ok {
			missingIDs = append(missingIDs, sourceID)
		}
	}
	sort.Strings(missingIDs)
	for _, sourceID := range missingIDs {
		changes = append(changes, governance.SourceScanChange{
			StableExternalID: sourceID, ChangeType: governance.SourceScanUnavailable,
			PreviousRevision: current[sourceID].Revision,
		})
	}
	result := governance.SourceScanResult{
		SnapshotSHA256: snapshotSHA,
		ObservedItems:  len(snapshot.Items),
		UnchangedItems: reconciliation.Unchanged,
		AddedItems:     reconciliation.Adopted,
		UpdatedItems:   reconciliation.UploadRequired,
		OverLimitItems: reconciliation.OverLimit,
		MissingItems:   len(missingIDs),
		PlannedUploads: reconciliation.UploadRequired,
		Changes:        changes,
	}
	return h.store.CompleteSourceScan(ctx, scan.ID, result, now)
}