ultisuite-backend/internal/migration/drive_delta.go

package migration

import (
	"context"
	"encoding/json"
	"fmt"
	"net/url"
	"path"
	"strings"
)

func (d *DriveImporter) hasDriveDeltaCursor(job *Job, provider string) bool {
	if provider == "google" {
		token, _ := job.CursorJSON["driveChangeToken"].(string)
		return strings.TrimSpace(token) != ""
	}
	link, _ := job.CursorJSON["driveDeltaLink"].(string)
	return strings.TrimSpace(link) != ""
}

func (d *DriveImporter) bootstrapDriveDelta(ctx context.Context, accessToken, provider string, cursor map[string]any) error {
	switch provider {
	case "google":
		body, err := apiGet(ctx, d.client, "https://www.googleapis.com/drive/v3/changes/startPageToken?spaces=drive", accessToken)
		if err != nil {
			return err
		}
		var parsed struct {
			StartPageToken string `json:"startPageToken"`
		}
		if err := json.Unmarshal(body, &parsed); err != nil {
			return err
		}
		if parsed.StartPageToken != "" {
			cursor["driveChangeToken"] = parsed.StartPageToken
		}
		return nil
	default:
		link, err := d.walkMicrosoftDriveDelta(ctx, accessToken, graphMicrosoftURL(d.userUPN, "/drive/root/delta?$select=id,name,folder,file,size,parentReference,deleted"))
		if err != nil {
			return err
		}
		if link != "" {
			cursor["driveDeltaLink"] = link
		}
		return nil
	}
}

func (d *DriveImporter) walkMicrosoftDriveDelta(ctx context.Context, accessToken, listURL string) (string, error) {
	for listURL != "" {
		body, err := apiGet(ctx, d.client, listURL, accessToken)
		if err != nil {
			return "", err
		}
		var parsed struct {
			NextLink  string `json:"@odata.nextLink"`
			DeltaLink string `json:"@odata.deltaLink"`
		}
		if err := json.Unmarshal(body, &parsed); err != nil {
			return "", err
		}
		if parsed.DeltaLink != "" {
			return parsed.DeltaLink, nil
		}
		listURL = parsed.NextLink
	}
	return "", nil
}

func (d *DriveImporter) importDriveDelta(ctx context.Context, job *Job, accessToken, provider, ncUserID, root string, items *ImportedItemStore, update progressUpdater) error {
	switch provider {
	case "google":
		return d.importGoogleDriveDelta(ctx, job, accessToken, ncUserID, root, items, update)
	default:
		return d.importMicrosoftDriveDelta(ctx, job, accessToken, ncUserID, root, items, update)
	}
}

func (d *DriveImporter) importGoogleDriveDelta(ctx context.Context, job *Job, accessToken, ncUserID, root string, items *ImportedItemStore, update progressUpdater) error {
	pageToken, _ := job.CursorJSON["driveChangeToken"].(string)
	if pageToken == "" {
		return fmt.Errorf("google drive delta token missing")
	}

	listURL := "https://www.googleapis.com/drive/v3/changes?pageSize=100&spaces=drive&includeRemoved=true&fields=" +
		url.QueryEscape("nextPageToken,newStartPageToken,changes(fileId,removed,file(id,name,mimeType,size,parents,trashed))") +
		"&pageToken=" + url.QueryEscape(pageToken)

	body, err := apiGet(ctx, d.client, listURL, accessToken)
	if err != nil {
		return err
	}
	var parsed struct {
		Changes          []googleDriveChange `json:"changes"`
		NextPageToken    string              `json:"nextPageToken"`
		NewStartPageToken string             `json:"newStartPageToken"`
	}
	if err := json.Unmarshal(body, &parsed); err != nil {
		return err
	}

	imported, _ := job.StatsJSON["delta_imported"].(float64)
	deleted, _ := job.StatsJSON["delta_deleted"].(float64)
	exported, _ := job.StatsJSON["exported"].(float64)
	skipped, _ := job.StatsJSON["skipped"].(float64)
	batch := 0
	listIndex := int(jsonNumber(job.CursorJSON["listIndex"]))

	for i := listIndex; i < len(parsed.Changes) && batch < driveImportBatchSize(); i++ {
		change := parsed.Changes[i]
		if change.Removed || change.File == nil || change.File.Trashed {
			if err := d.deleteDriveItem(ctx, ncUserID, root, items, change.FileID); err != nil {
				return err
			}
			deleted++
			batch++
			continue
		}
		item := googleFileToDriveItem(*change.File)
		relPath := d.resolveDriveRelPath(items, item)
		if err := d.uploadDriveItem(ctx, accessToken, ncUserID, root, relPath, item, items, &imported, &exported, &skipped, job.StatsJSON); err != nil {
			return err
		}
		batch++
	}

	job.StatsJSON["delta_imported"] = imported
	job.StatsJSON["delta_deleted"] = deleted
	job.StatsJSON["exported"] = exported
	job.StatsJSON["skipped"] = skipped

	if listIndex+batch < len(parsed.Changes) {
		job.CursorJSON["listIndex"] = float64(listIndex + batch)
		return update("pending", job.CursorJSON, job.StatsJSON, "")
	}
	delete(job.CursorJSON, "listIndex")

	if parsed.NextPageToken != "" {
		job.CursorJSON["driveChangeToken"] = parsed.NextPageToken
		return update("pending", job.CursorJSON, job.StatsJSON, "")
	}
	if parsed.NewStartPageToken != "" {
		job.CursorJSON["driveChangeToken"] = parsed.NewStartPageToken
	}
	job.StatsJSON["phase"] = "delta"
	return update("completed", job.CursorJSON, job.StatsJSON, "")
}

type googleDriveChange struct {
	FileID  string          `json:"fileId"`
	Removed bool            `json:"removed"`
	File    *googleDriveFile `json:"file"`
}

type googleDriveFile struct {
	ID       string   `json:"id"`
	Name     string   `json:"name"`
	MimeType string   `json:"mimeType"`
	Size     string   `json:"size"`
	Parents  []string `json:"parents"`
	Trashed  bool     `json:"trashed"`
}

func googleFileToDriveItem(f googleDriveFile) driveItem {
	size := int64(0)
	if f.Size != "" {
		fmt.Sscan(f.Size, &size)
	}
	item := driveItem{
		ID:       f.ID,
		Name:     f.Name,
		IsFolder: f.MimeType == "application/vnd.google-apps.folder",
		Size:     size,
		MimeType: f.MimeType,
	}
	if len(f.Parents) > 0 {
		item.ParentID = f.Parents[0]
	}
	if item.IsFolder {
		return item
	}
	if exportMime, ext, ok := googleWorkspaceExport(f.MimeType); ok {
		item.Export = true
		item.ExportMime = exportMime
		item.ExportExt = ext
		item.Name = driveExportFileName(f.Name, ext)
	} else {
		item.Download = "https://www.googleapis.com/drive/v3/files/" + url.PathEscape(f.ID) + "?alt=media"
	}
	return item
}

func (d *DriveImporter) importMicrosoftDriveDelta(ctx context.Context, job *Job, accessToken, ncUserID, root string, items *ImportedItemStore, update progressUpdater) error {
	deltaLink, _ := job.CursorJSON["driveDeltaLink"].(string)
	if deltaLink == "" {
		return fmt.Errorf("microsoft drive delta link missing")
	}

	body, err := apiGet(ctx, d.client, deltaLink, accessToken)
	if err != nil {
		return err
	}
	var parsed struct {
		Value     []graphDriveItem `json:"value"`
		NextLink  string           `json:"@odata.nextLink"`
		DeltaLink string           `json:"@odata.deltaLink"`
	}
	if err := json.Unmarshal(body, &parsed); err != nil {
		return err
	}

	imported, _ := job.StatsJSON["delta_imported"].(float64)
	deleted, _ := job.StatsJSON["delta_deleted"].(float64)
	skipped, _ := job.StatsJSON["skipped"].(float64)
	batch := 0
	listIndex := int(jsonNumber(job.CursorJSON["listIndex"]))

	for i := listIndex; i < len(parsed.Value) && batch < driveImportBatchSize(); i++ {
		item := parsed.Value[i]
		sourceID := strings.TrimSpace(item.ID)
		if sourceID == "" {
			continue
		}
		if item.Removed != nil || item.Deleted != nil {
			if err := d.deleteDriveItem(ctx, ncUserID, root, items, sourceID); err != nil {
				return err
			}
			deleted++
			batch++
			continue
		}
		driveItem := graphDriveToItem(d.userUPN, item)
		relPath := d.resolveDriveRelPath(items, driveItem)
		if err := d.uploadDriveItem(ctx, accessToken, ncUserID, root, relPath, driveItem, items, &imported, nil, &skipped, job.StatsJSON); err != nil {
			return err
		}
		batch++
	}

	job.StatsJSON["delta_imported"] = imported
	job.StatsJSON["delta_deleted"] = deleted
	job.StatsJSON["skipped"] = skipped

	if listIndex+batch < len(parsed.Value) {
		job.CursorJSON["listIndex"] = float64(listIndex + batch)
		return update("pending", job.CursorJSON, job.StatsJSON, "")
	}
	delete(job.CursorJSON, "listIndex")

	if parsed.NextLink != "" {
		job.CursorJSON["driveDeltaLink"] = parsed.NextLink
		return update("pending", job.CursorJSON, job.StatsJSON, "")
	}
	if parsed.DeltaLink != "" {
		job.CursorJSON["driveDeltaLink"] = parsed.DeltaLink
	}
	job.StatsJSON["phase"] = "delta"
	return update("completed", job.CursorJSON, job.StatsJSON, "")
}

type graphDriveItem struct {
	ID              string `json:"id"`
	Name            string `json:"name"`
	Size            int64  `json:"size"`
	Folder          *struct{ ChildCount int `json:"childCount"` } `json:"folder"`
	File            *struct{ MimeType string `json:"mimeType"` }   `json:"file"`
	ParentReference *struct {
		ID string `json:"id"`
	} `json:"parentReference"`
	Removed *struct {
		Reason string `json:"reason"`
	} `json:"@removed"`
	Deleted *struct{} `json:"deleted"`
}

func graphDriveToItem(userUPN string, item graphDriveItem) driveItem {
	out := driveItem{
		ID:   item.ID,
		Name: item.Name,
		Size: item.Size,
	}
	if item.ParentReference != nil {
		out.ParentID = item.ParentReference.ID
	}
	if item.Folder != nil {
		out.IsFolder = true
		return out
	}
	mime := ""
	if item.File != nil {
		mime = item.File.MimeType
	}
	out.MimeType = mime
	out.Download = graphMicrosoftURL(userUPN, "/drive/items/"+url.PathEscape(item.ID)+"/content")
	return out
}

func (d *DriveImporter) resolveDriveRelPath(items *ImportedItemStore, item driveItem) string {
	if stored := items.Path(item.ID); stored != "" {
		return stored
	}
	parentRel := ""
	if item.ParentID != "" {
		parentRel = items.Path(item.ParentID)
	}
	return path.Join(parentRel, sanitizeDrivePath(item.Name))
}

func (d *DriveImporter) uploadDriveItem(ctx context.Context, accessToken, ncUserID, root, relPath string, item driveItem, items *ImportedItemStore, imported, exported, skipped *float64, stats map[string]any) error {
	targetPath := path.Join(root, relPath)
	if item.IsFolder {
		if err := d.nc.CreateFolder(ctx, ncUserID, targetPath); err != nil {
			if markErr := items.MarkFailed(ctx, item.ID, err.Error(), relPath); markErr != nil {
				return markErr
			}
			incJobStat(stats, "failed")
			return nil
		}
		if err := items.MarkPath(ctx, item.ID, relPath); err != nil {
			return err
		}
		if imported != nil {
			*imported++
		}
		return nil
	}
	if item.Export {
		content, contentType, fileName, err := d.downloadGoogleExport(ctx, accessToken, item)
		if err != nil {
			if skipped != nil {
				*skipped++
			}
			return items.MarkSkipped(ctx, item.ID, "export: "+err.Error(), relPath)
		}
		targetPath = path.Join(path.Dir(targetPath), fileName)
		relPath = path.Join(path.Dir(relPath), fileName)
		if err := d.nc.Upload(ctx, ncUserID, targetPath, content, contentType); err != nil {
			if markErr := items.MarkFailed(ctx, item.ID, err.Error(), relPath); markErr != nil {
				return markErr
			}
			incJobStat(stats, "failed")
			return nil
		}
		if exported != nil {
			*exported++
		}
		if pdfMime, pdfExt, ok := googleSlidesPDFExport(item.MimeType); ok {
			pdfItem := item
			pdfItem.ExportMime = pdfMime
			pdfItem.ExportExt = pdfExt
			pdfContent, pdfType, pdfName, err := d.downloadGoogleExport(ctx, accessToken, pdfItem)
			if err == nil {
				pdfRel := path.Join(path.Dir(relPath), pdfName)
				pdfTarget := path.Join(root, pdfRel)
				if err := d.nc.Upload(ctx, ncUserID, pdfTarget, pdfContent, pdfType); err == nil {
					if err := items.MarkPath(ctx, item.ID+"_pdf", pdfRel); err != nil {
						return err
					}
				}
			}
		}
	} else {
		if item.Size > maxDriveFileBytes {
			if skipped != nil {
				*skipped++
			}
			reason := fmt.Sprintf("file exceeds %d byte limit", maxDriveFileBytes)
			return items.MarkSkipped(ctx, item.ID, reason, relPath)
		}
		content, contentType, err := d.downloadDriveFile(ctx, accessToken, item)
		if err != nil {
			if markErr := items.MarkFailed(ctx, item.ID, err.Error(), relPath); markErr != nil {
				return markErr
			}
			incJobStat(stats, "failed")
			return nil
		}
		if err := d.nc.Upload(ctx, ncUserID, targetPath, content, contentType); err != nil {
			if markErr := items.MarkFailed(ctx, item.ID, err.Error(), relPath); markErr != nil {
				return markErr
			}
			incJobStat(stats, "failed")
			return nil
		}
	}
	if err := items.MarkImported(ctx, item.ID); err != nil {
		return err
	}
	if err := items.MarkPath(ctx, item.ID, relPath); err != nil {
		return err
	}
	if imported != nil {
		*imported++
	}
	return nil
}

func (d *DriveImporter) deleteDriveItem(ctx context.Context, ncUserID, root string, items *ImportedItemStore, fileID string) error {
	if fileID == "" {
		return nil
	}
	for _, suffix := range []string{"", "_pdf"} {
		rel := items.Path(fileID + suffix)
		if rel == "" {
			continue
		}
		target := path.Join(root, rel)
		if err := d.nc.Delete(ctx, ncUserID, target); err != nil && !isDeleteNotFound(err) {
			return err
		}
	}
	if err := items.Unmark(ctx, fileID); err != nil {
		return err
	}
	return items.Unmark(ctx, fileID+"_pdf")
}

func isDeleteNotFound(err error) bool {
	if err == nil {
		return false
	}
	msg := strings.ToLower(err.Error())
	return strings.Contains(msg, "404") || strings.Contains(msg, "not found")
}