WeKnora deployment source with the reviewed wiki identity, summary and recovery fixes

internal/application/repository/knowledge_version_graph.go

@@ -0,0 +1,119 @@
+package repository
+
+import (
+	"context"
+	"crypto/sha256"
+	"encoding/hex"
+	"fmt"
+
+	"github.com/Tencent/WeKnora/internal/knowledgeversion"
+	"github.com/Tencent/WeKnora/internal/types"
+	"gorm.io/gorm"
+)
+
+// CopyVersionGraph copies imported PostgreSQL evidence without rerunning AI
+// extraction. Provenance remains immutable; only proven native pointers change.
+func (r *knowledgeRepository) CopyVersionGraph(ctx context.Context, source, target string, documents, chunks map[string]string) error {
+	tenant, ok := types.TenantIDFromContext(ctx)
+	if !knowledgeversion.IsWorker(ctx) || !ok || tenant == 0 || source == target {
+		return fmt.Errorf("graph version copy requires an isolated workspace worker")
+	}
+	db := r.db.WithContext(ctx)
+	if !knowledgeversion.GraphTablesExist(db) {
+		return nil
+	}
+	return db.Transaction(func(tx *gorm.DB) error {
+		var bases []types.KnowledgeBase
+		if err := tx.Where("tenant_id=? AND id IN ?", tenant, []string{source, target}).Find(&bases).Error; err != nil {
+			return err
+		}
+		validTarget := false
+		for _, kb := range bases {
+			if kb.ID == target && kb.IsTemporary && kb.VersionState == "copying" {
+				validTarget = true
+			}
+		}
+		if len(bases) != 2 || !validTarget {
+			return fmt.Errorf("graph version target must be a temporary workspace copy")
+		}
+		for _, model := range []any{&knowledgeversion.GraphRecord{}, &knowledgeversion.GraphSource{}} {
+			var count int64
+			if err := tx.Model(model).Where("tenant_id=? AND knowledge_base_id=?", tenant, target).Count(&count).Error; err != nil {
+				return err
+			}
+			if count != 0 {
+				return fmt.Errorf("graph version target must be empty")
+			}
+		}
+		for offset := 0; ; offset += 200 {
+			var records []knowledgeversion.GraphRecord
+			if err := tx.Where("tenant_id=? AND knowledge_base_id=?", tenant, source).Order("source_table,legacy_id").Limit(200).Offset(offset).Find(&records).Error; err != nil {
+				return err
+			}
+			if len(records) == 0 {
+				break
+			}
+			for i := range records {
+				records[i].KnowledgeBaseID = target
+			}
+			if err := tx.Create(&records).Error; err != nil {
+				return err
+			}
+		}
+		for offset := 0; ; offset += 200 {
+			var evidence []knowledgeversion.GraphSource
+			if err := tx.Where("tenant_id=? AND knowledge_base_id=?", tenant, source).Order("source_chunk_id").Limit(200).Offset(offset).Find(&evidence).Error; err != nil {
+				return err
+			}
+			if len(evidence) == 0 {
+				break
+			}
+			ids := []string{}
+			for _, e := range evidence {
+				if e.NativeChunkID != nil {
+					ids = append(ids, *e.NativeChunkID)
+					if dst := chunks[*e.NativeChunkID]; dst != "" {
+						ids = append(ids, dst)
+					}
+				}
+			}
+			loaded := []types.Chunk{}
+			if len(ids) > 0 {
+				if err := tx.Where("tenant_id=? AND knowledge_base_id IN ? AND id IN ?", tenant, []string{source, target}, ids).Find(&loaded).Error; err != nil {
+					return err
+				}
+			}
+			byID := map[string]types.Chunk{}
+			for _, c := range loaded {
+				byID[c.ID] = c
+			}
+			for i := range evidence {
+				e := &evidence[i]
+				e.KnowledgeBaseID = target
+				oldID, oldDoc, oldHash := e.NativeChunkID, e.NativeKnowledgeID, e.NativeContentSHA256
+				e.NativeChunkID, e.NativeKnowledgeID, e.NativeContentSHA256 = nil, nil, nil
+				if oldID == nil {
+					continue
+				}
+				e.MappingState = "stale"
+				src, srcOK := byID[*oldID]
+				dst, dstOK := byID[chunks[*oldID]]
+				if !srcOK || !dstOK || oldDoc == nil || oldHash == nil || src.KnowledgeBaseID != source || dst.KnowledgeBaseID != target || src.KnowledgeID != *oldDoc || documents[*oldDoc] != dst.KnowledgeID {
+					continue
+				}
+				sourceHash := sha256.Sum256([]byte(src.Content))
+				if hex.EncodeToString(sourceHash[:]) != *oldHash {
+					continue
+				}
+				targetHash := sha256.Sum256([]byte(dst.Content))
+				hash := hex.EncodeToString(targetHash[:])
+				e.NativeChunkID, e.NativeKnowledgeID, e.NativeContentSHA256 = &dst.ID, &dst.KnowledgeID, &hash
+				e.MappingState = "exact_content_hash"
+			}
+			if err := tx.Create(&evidence).Error; err != nil {
+				return err
+			}
+		}
+		return nil
+	})
+}