func TestOwnerOf(t *testing.T) {
if util.OwnerOf("aptrust.receiving.unc.edu") != "unc.edu" {
t.Error("OwnerOf misidentified receiving bucket owner")
}
if util.OwnerOf("aptrust.receiving.test.unc.edu") != "unc.edu" {
t.Error("OwnerOf misidentified receiving bucket owner")
}
if util.OwnerOf("aptrust.restore.unc.edu") != "unc.edu" {
t.Error("OwnerOf misidentified restoration bucket owner")
}
}
func (reader *APTBucketReader) createWorkItem(bucket string, s3Object *s3.Object) *models.WorkItem {
// Create a WorkItem in Pharos
institution := reader.Institutions[util.OwnerOf(bucket)]
if institution == nil {
errMsg := fmt.Sprintf("Cannot find institution record for item %s/%s. "+
"Owner computes to '%s'", bucket, *s3Object.Key, util.OwnerOf(bucket))
reader.Context.MessageLog.Error(errMsg)
if reader.stats != nil {
reader.stats.AddError(errMsg)
}
return nil
}
workItem := &models.WorkItem{}
workItem.Name = *s3Object.Key
workItem.Bucket = bucket
workItem.ETag = strings.Replace(*s3Object.ETag, "\"", "", -1)
workItem.Size = *s3Object.Size
workItem.BagDate = *s3Object.LastModified
workItem.InstitutionId = institution.Id
workItem.User = constants.APTrustSystemUser
workItem.Date = time.Now().UTC()
workItem.Note = "Bag is in receiving bucket"
workItem.Action = constants.ActionIngest
workItem.Stage = constants.StageReceive
workItem.Status = constants.StatusPending
workItem.Outcome = "Item is pending ingest"
workItem.Retry = true
resp := reader.Context.PharosClient.WorkItemSave(workItem)
if resp.Error != nil {
errMsg := fmt.Sprintf("Error creating WorkItem for name '%s', etag '%s', time '%s': %v",
workItem.Name, workItem.ETag, workItem.BagDate, resp.Error)
reader.Context.MessageLog.Debug("%s", resp.Request.URL.String())
reader.Context.MessageLog.Error(errMsg)
if reader.stats != nil {
reader.stats.AddError(errMsg)
}
return nil
}
if resp.Response.StatusCode != 201 {
reader.processPharosError(resp)
return nil
}
savedWorkItem := resp.WorkItem()
reader.Context.MessageLog.Debug("Created WorkItem with id %d for %s/%s in Pharos",
savedWorkItem.Id, bucket, *s3Object.Key)
if reader.stats != nil {
reader.stats.AddWorkItem("WorkItemsCreated", savedWorkItem)
}
return savedWorkItem
}
// SetBasicObjectInfo sets initial essential properties on the
// IntellectualObject associated with an ingestState
// (ingestState.IngestManifest.Object). This is only used by
// apt_fetcher and is only ever called during the fetch stage.
func SetBasicObjectInfo(ingestState *models.IngestState, _context *context.Context) {
// instIdentifier is, e.g., virginia.edu, ncsu.edu, etc.
// We'll download the tar file from the receiving bucket to
// something like /mnt/apt/data/virginia.edu/name_of_bag.tar
// See IngestTarFilePath below.
obj := ingestState.IngestManifest.Object
instIdentifier := util.OwnerOf(ingestState.IngestManifest.S3Bucket)
obj.BagName = util.CleanBagName(ingestState.IngestManifest.S3Key)
obj.Institution = instIdentifier
obj.InstitutionId = ingestState.WorkItem.InstitutionId
obj.IngestS3Bucket = ingestState.IngestManifest.S3Bucket
obj.IngestS3Key = ingestState.IngestManifest.S3Key
obj.IngestTarFilePath = filepath.Join(
_context.Config.TarDirectory,
instIdentifier, ingestState.IngestManifest.S3Key)
// If this IntellectualObject was created by our validator and VirtualBag,
// the identifier will be the bag name (minus the .tar extension).
// That's fine for cases where depositors or other organizations are
// using the validator outside of APTrust's repository environment, but
// APTrust requires that we add the Institution name and a slash to
// the beginning of the identifier. So make sure it's there, and propagate
// the change all the way down to the GenericFiles.
if !strings.HasPrefix(obj.Identifier, obj.Institution+"/") {
obj.Identifier = fmt.Sprintf("%s/%s", obj.Institution, obj.Identifier)
for _, gf := range obj.GenericFiles {
if !strings.HasPrefix(gf.Identifier, obj.Identifier) {
gf.IntellectualObjectIdentifier = obj.Identifier
gf.Identifier = fmt.Sprintf("%s/%s", obj.Institution, gf.Identifier)
}
}
}
}
// The name of the owning institution, followed by a slash, followed
// by the name of the tar file. This differs from the ObjectName,
// because it will have the .tar or bag.001.of030.tar suffix.
func (s3File *S3File) BagName() string {
return fmt.Sprintf("%s/%s", util.OwnerOf(s3File.BucketName), s3File.Key.Key)
}
// Returns the object identifier that will identify this bag
// in fedora. That's the institution identifier, followed by
// a slash and the tar file name, minus the .tar extension
// and the ".bag1of12" multipart extension. So for BucketName
// "aptrust.receiving.unc.edu" and Key.Key "nc_bag.b001.of030.tar",
// this would return "unc.edu/nc_bag"
func (s3File *S3File) ObjectName() (string, error) {
institution := util.OwnerOf(s3File.BucketName)
cleanBagName := util.CleanBagName(s3File.Key.Key)
return fmt.Sprintf("%s/%s", institution, cleanBagName), nil
}