Add PutMetadata function to storage backends (#171)

* Add PutMetadata function to storage backends

This function is not currently used, but it will be useful for helper
scripts that need to regenerate metadata on the fly, especially scripts
to migrate between storage backends. In the future, we can also use it
to automatically regenerate metadata if it is found to be missing or
corrupted.

* Add PutMetadata function to storage backend interface and
  implementations
* Rework metadata generation to be more efficient and work better with
  the PutMetadata function
* Add a basic test for metadata generation

* Change PutMetadata to take a Metadata type instead

It's unlikely that this function is useful if it always regenerates the
metadata. Instead, the caller should do that if it needs.

backends/localfs/localfs.go

@@ -126,11 +126,16 @@ func (b LocalfsBackend) Put(key string, r io.Reader, expiry time.Time, deleteKey
 		return m, err
 	}
 
+	dst.Seek(0 ,0)
+	m, err = helpers.GenerateMetadata(dst)
+	if err != nil {
+		os.Remove(filePath)
+		return
+	}
+	dst.Seek(0 ,0)
+
 	m.Expiry = expiry
 	m.DeleteKey = deleteKey
-	m.Size = bytes
-	m.Mimetype, _ = helpers.DetectMime(dst)
-	m.Sha256sum, _ = helpers.Sha256sum(dst)
 	m.ArchiveFiles, _ = helpers.ListArchiveFiles(m.Mimetype, m.Size, dst)
 
 	err = b.writeMetadata(key, m)
@@ -142,6 +147,15 @@ func (b LocalfsBackend) Put(key string, r io.Reader, expiry time.Time, deleteKey
 	return
 }
 
+func (b LocalfsBackend) PutMetadata(key string, m backends.Metadata) (err error) {
+	err = b.writeMetadata(key, m)
+	if err != nil {
+		return
+	}
+
+	return
+}
+
 func (b LocalfsBackend) Size(key string) (int64, error) {
 	fileInfo, err := os.Stat(path.Join(b.filesPath, key))
 	if err != nil {

backends/s3/s3.go

@@ -122,11 +122,12 @@ func (b S3Backend) Put(key string, r io.Reader, expiry time.Time, deleteKey stri
 		return m, err
 	}
 
+	m, err = helpers.GenerateMetadata(r)
+	if err != nil {
+		return
+	}
 	m.Expiry = expiry
 	m.DeleteKey = deleteKey
-	m.Size = bytes
-	m.Mimetype, _ = helpers.DetectMime(tmpDst)
-	m.Sha256sum, _ = helpers.Sha256sum(tmpDst)
 	// XXX: we may not be able to write this to AWS easily
 	//m.ArchiveFiles, _ = helpers.ListArchiveFiles(m.Mimetype, m.Size, tmpDst)
 
@@ -145,6 +146,20 @@ func (b S3Backend) Put(key string, r io.Reader, expiry time.Time, deleteKey stri
 	return
 }
 
+func (b S3Backend) PutMetadata(key string, m backends.Metadata) (err error) {
+	_, err = b.svc.CopyObject(&s3.CopyObjectInput{
+		Bucket:     aws.String(b.bucket),
+		Key:        aws.String(key),
+		CopySource: aws.String("/" + b.bucket + "/" + key),
+		Metadata:   mapMetadata(m),
+	})
+	if err != nil {
+		return
+	}
+
+	return
+}
+
 func (b S3Backend) Size(key string) (int64, error) {
 	input := &s3.HeadObjectInput{
 		Bucket: aws.String(b.bucket),
@@ -169,7 +184,6 @@ func (b S3Backend) List() ([]string, error) {
 		return nil, err
 	}
 
-
 	for _, object := range results.Contents {
 		output = append(output, *object.Key)
 	}

backends/storage.go

@@ -12,6 +12,7 @@ type StorageBackend interface {
 	Head(key string) (Metadata, error)
 	Get(key string) (Metadata, io.ReadCloser, error)
 	Put(key string, r io.Reader, expiry time.Time, deleteKey string) (Metadata, error)
+	PutMetadata(key string, m Metadata) error
 	Size(key string) (int64, error)
 }
 

helpers/helpers.go

@@ -1,49 +1,65 @@
 package helpers
 
 import (
+	"bytes"
 	"encoding/hex"
 	"io"
 	"unicode"
 
+	"github.com/andreimarcu/linx-server/backends"
 	"github.com/minio/sha256-simd"
 	"gopkg.in/h2non/filetype.v1"
 )
 
-func DetectMime(r io.ReadSeeker) (string, error) {
+func GenerateMetadata(r io.Reader) (m backends.Metadata, err error) {
+	// Since we don't have the ability to seek within a file, we can use a
+	// Buffer in combination with a TeeReader to keep a copy of the bytes
+	// we read when detecting the file type. These bytes are still needed
+	// to hash the file and determine its size and cannot be discarded.
+	var buf bytes.Buffer
+	teeReader := io.TeeReader(r, &buf)
+
 	// Get first 512 bytes for mimetype detection
 	header := make([]byte, 512)
+	_, err = teeReader.Read(header)
+	if err != nil {
+		return
+	}
 
-	r.Seek(0, 0)
-	r.Read(header)
-	r.Seek(0, 0)
+	// Create a Hash and a MultiReader that includes the Buffer we created
+	// above along with the original Reader, which will have the rest of
+	// the file.
+	hasher := sha256.New()
+	multiReader := io.MultiReader(&buf, r)
 
+	// Copy everything into the Hash, then use the number of bytes written
+	// as the file size.
+	var readLen int64
+	readLen, err = io.Copy(hasher, multiReader)
+	if err != nil {
+		return
+	} else {
+		m.Size += readLen
+	}
+
+	// Get the hex-encoded string version of the Hash checksum
+	m.Sha256sum = hex.EncodeToString(hasher.Sum(nil))
+
+	// Use the bytes we extracted earlier and attempt to determine the file
+	// type
 	kind, err := filetype.Match(header)
 	if err != nil {
-		return "application/octet-stream", err
+		m.Mimetype = "application/octet-stream"
+		return m, err
 	} else if kind.MIME.Value != "" {
-		return kind.MIME.Value, nil
-	}
-
-	// Check if the file seems anything like text
-	if printable(header) {
-		return "text/plain", nil
+		m.Mimetype = kind.MIME.Value
+	} else if printable(header) {
+		m.Mimetype = "text/plain"
 	} else {
-		return "application/octet-stream", nil
-	}
+		m.Mimetype = "application/octet-stream"
 	}
 
-func Sha256sum(r io.ReadSeeker) (string, error) {
-	hasher := sha256.New()
-
-	r.Seek(0, 0)
-	_, err := io.Copy(hasher, r)
-	if err != nil {
-		return "", err
-	}
-
-	r.Seek(0, 0)
-
-	return hex.EncodeToString(hasher.Sum(nil)), nil
+	return
 }
 
 func printable(data []byte) bool {

helpers/helpers_test.go

@@ -0,0 +1,29 @@
+package helpers
+
+import (
+	"strings"
+	"testing"
+)
+
+func TestGenerateMetadata(t *testing.T) {
+	r := strings.NewReader("This is my test content")
+	m, err := GenerateMetadata(r)
+	if err != nil {
+		t.Fatal(err)
+	}
+
+	expectedSha256sum := "966152d20a77e739716a625373ee15af16e8f4aec631a329a27da41c204b0171"
+	if m.Sha256sum != expectedSha256sum {
+		t.Fatalf("Sha256sum was %q instead of expected value of %q", m.Sha256sum, expectedSha256sum)
+	}
+
+	expectedMimetype := "text/plain"
+	if m.Mimetype != expectedMimetype {
+		t.Fatalf("Mimetype was %q instead of expected value of %q", m.Mimetype, expectedMimetype)
+	}
+
+	expectedSize := int64(23)
+	if m.Size != expectedSize {
+		t.Fatalf("Size was %d instead of expected value of %d", m.Size, expectedSize)
+	}
+}