Initial commit: knox knowledge index

2026-08-29 02:52:32 -07:00
commit 8eecde18b4
31 changed files with 6855 additions and 0 deletions
@@ -0,0 +1,71 @@
+package cmd
+
+import (
+	"fmt"
+	"log"
+
+	"github.com/spf13/cobra"
+	"github.com/david/knox/internal/db"
+	"github.com/david/knox/internal/ingest"
+)
+
+func NewGiteaCmd(kdb *db.KnoxDB) *cobra.Command {
+	cmd := &cobra.Command{
+		Use:   "gitea",
+		Short: "Ingest Gitea repositories, issues, and PRs into the knowledge index",
+		Long: `Uses the 'tea' CLI to fetch repositories, open issues, and open
+pull requests from Gitea. Fingerprints each by ID for dedup.`,
+		RunE: func(c *cobra.Command, args []string) error {
+			ing := ingest.NewGiteaIngester()
+			log.Printf("[knox] ingesting Gitea data via tea...")
+
+			results, err := ing.IngestAll()
+			if err != nil {
+				return fmt.Errorf("gitea ingest: %w", err)
+			}
+
+			if len(results) == 0 {
+				fmt.Println("No Gitea data found.")
+				return nil
+			}
+
+			var repos, issues, pulls int
+			var newCount int
+			for _, result := range results {
+				_, isNew, err := kdb.RecordObservation(db.ObservationRecord{
+					Fingerprint:     result.Fingerprint,
+					SourceID:        result.SourceID,
+					SourcePath:      result.SourcePath,
+					Project:         result.Project,
+					ContentType:     result.ContentType,
+					Title:           result.Title,
+					Summary:         result.Summary,
+					CreatedAt:       result.CreatedAt,
+					Confidence:      result.Confidence,
+					IngesterVersion: result.IngesterVersion,
+					Trigger:         "gitea_ingest",
+					Provenance:      ingest.ProvenanceJSON(result.Provenance),
+				})
+				if err != nil {
+					log.Printf("[knox] db error: %v", err)
+					continue
+				}
+				if isNew {
+					newCount++
+				}
+				switch result.ContentType {
+				case "repo":
+					repos++
+				case "issue":
+					issues++
+				case "pull":
+					pulls++
+				}
+			}
+
+			fmt.Printf("Gitea: %d repos, %d issues, %d PRs (%d new)\n", repos, issues, pulls, newCount)
+			return nil
+		},
+	}
+	return cmd
+}