feat: add identifiable field to source object (#1243)

go.mod

@@ -60,6 +60,7 @@ require (
 	github.com/google/go-containerregistry v0.11.0
 	github.com/in-toto/in-toto-golang v0.3.4-0.20220709202702-fa494aaa0add
 	github.com/knqyf263/go-rpmdb v0.0.0-20220629110411-9a3bd2ebb923
+	github.com/opencontainers/go-digest v1.0.0
 	github.com/sassoftware/go-rpmutils v0.2.0
 	github.com/sigstore/cosign v1.12.1
 	github.com/sigstore/rekor v0.12.1-0.20220915152154-4bb6f441c1b2
@@ -218,7 +219,6 @@ require (
 	github.com/mozillazg/docker-credential-acr-helper v0.3.0 // indirect
 	github.com/nwaples/rardecode v1.1.0 // indirect
 	github.com/oklog/ulid v1.3.1 // indirect
-	github.com/opencontainers/go-digest v1.0.0 // indirect
 	github.com/opencontainers/image-spec v1.0.3-0.20220114050600-8b9d41f48198 // indirect
 	github.com/opentracing/opentracing-go v1.2.0 // indirect
 	github.com/pelletier/go-toml/v2 v2.0.5 // indirect

syft/source/source.go

@@ -15,16 +15,19 @@ import (
 
 	"github.com/bmatcuk/doublestar/v4"
 	"github.com/mholt/archiver/v3"
+	digest "github.com/opencontainers/go-digest"
 	"github.com/spf13/afero"
 
 	"github.com/anchore/stereoscope"
 	"github.com/anchore/stereoscope/pkg/image"
 	"github.com/anchore/syft/internal/log"
+	"github.com/anchore/syft/syft/artifact"
 )
 
 // Source is an object that captures the data source to be cataloged, configuration, and a specific resolver used
 // in cataloging (based on the data source and configuration)
 type Source struct {
+	id                artifact.ID
 	Image             *image.Image // the image object to be cataloged (image only)
 	Metadata          Metadata
 	directoryResolver *directoryResolver
@@ -304,6 +307,76 @@ func NewFromImage(img *image.Image, userImageStr string) (Source, error) {
 	}, nil
 }
 
+func (s Source) ID() artifact.ID {
+	if s.id == "" {
+		s.SetID()
+	}
+	return s.id
+}
+
+func (s *Source) SetID() {
+	var d string
+	switch s.Metadata.Scheme {
+	case DirectoryScheme:
+		d = digest.FromString(s.Metadata.Path).String()
+	case FileScheme:
+		// attempt to use the digest of the contents of the file as the ID
+		file, err := os.Open(s.Metadata.Path)
+		if err != nil {
+			d = digest.FromString(s.Metadata.Path).String()
+			break
+		}
+		di, err := digest.FromReader(file)
+		if err != nil {
+			d = digest.FromString(s.Metadata.Path).String()
+			break
+		}
+		d = di.String()
+	case ImageScheme:
+		manifestDigest := digest.FromBytes(s.Image.Metadata.RawManifest).String()
+		if manifestDigest != "" {
+			d = manifestDigest
+			break
+		}
+
+		// calcuate chain ID for image sources where manifestDigest is not available
+		// https://github.com/opencontainers/image-spec/blob/main/config.md#layer-chainid
+		d = calculateChainID(s.Image)
+		if d == "" {
+			// TODO what happens here if image has no layers?
+			// Is this case possible
+			d = digest.FromString(s.Metadata.ImageMetadata.UserInput).String()
+		}
+	default: // for UnknownScheme we hash the struct
+		id, _ := artifact.IDByHash(s)
+		d = string(id)
+	}
+
+	s.id = artifact.ID(strings.TrimPrefix(d, "sha256:"))
+}
+
+func calculateChainID(img *image.Image) string {
+	if len(img.Layers) < 1 {
+		return ""
+	}
+
+	// DiffID(L0) = digest of layer 0
+	// https://github.com/anchore/stereoscope/blob/1b1b744a919964f38d14e1416fb3f25221b761ce/pkg/image/layer_metadata.go#L19-L32
+	chainID := img.Layers[0].Metadata.Digest
+	id := chain(chainID, img.Layers[1:])
+
+	return id
+}
+
+func chain(chainID string, layers []*image.Layer) string {
+	if len(layers) < 1 {
+		return chainID
+	}
+
+	chainID = digest.FromString(layers[0].Metadata.Digest + " " + chainID).String()
+	return chain(chainID, layers[1:])
+}
+
 func (s *Source) FileResolver(scope Scope) (FileResolver, error) {
 	switch s.Metadata.Scheme {
 	case DirectoryScheme, FileScheme:

syft/source/source_test.go

@@ -18,6 +18,7 @@ import (
 
 	"github.com/anchore/stereoscope/pkg/image"
 	"github.com/anchore/stereoscope/pkg/imagetest"
+	"github.com/anchore/syft/syft/artifact"
 )
 
 func TestParseInput(t *testing.T) {
@@ -65,6 +66,72 @@ func TestNewFromImageFails(t *testing.T) {
 	})
 }
 
+func TestSetID(t *testing.T) {
+	layer := image.NewLayer(nil)
+	layer.Metadata = image.LayerMetadata{
+		Digest: "sha256:6f4fb385d4e698647bf2a450749dfbb7bc2831ec9a730ef4046c78c08d468e89",
+	}
+	img := image.Image{
+		Layers: []*image.Layer{layer},
+	}
+
+	tests := []struct {
+		name     string
+		input    *Source
+		expected artifact.ID
+	}{
+		{
+			name: "source.SetID sets the ID for FileScheme",
+			input: &Source{
+				Metadata: Metadata{
+					Scheme: FileScheme,
+					Path:   "test-fixtures/image-simple/file-1.txt",
+				},
+			},
+			expected: artifact.ID("55096713247489add592ce977637be868497132b36d1e294a3831925ec64319a"),
+		},
+		{
+			name: "source.SetID sets the ID for ImageScheme",
+			input: &Source{
+				Image: &img,
+				Metadata: Metadata{
+					Scheme: ImageScheme,
+				},
+			},
+			expected: artifact.ID("e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855"),
+		},
+		{
+			name: "source.SetID sets the ID for DirectoryScheme",
+			input: &Source{
+				Image: &img,
+				Metadata: Metadata{
+					Scheme: DirectoryScheme,
+					Path:   "test-fixtures/image-simple",
+				},
+			},
+			expected: artifact.ID("91db61e5e0ae097ef764796ce85e442a93f2a03e5313d4c7307e9b413f62e8c4"),
+		},
+		{
+			name: "source.SetID sets the ID for UnknownScheme",
+			input: &Source{
+				Image: &img,
+				Metadata: Metadata{
+					Scheme: UnknownScheme,
+					Path:   "test-fixtures/image-simple",
+				},
+			},
+			expected: artifact.ID("febd2d6148dc327d"),
+		},
+	}
+
+	for _, test := range tests {
+		t.Run(test.name, func(t *testing.T) {
+			test.input.SetID()
+			assert.Equal(t, test.expected, test.input.ID())
+		})
+	}
+}
+
 func TestNewFromImage(t *testing.T) {
 	layer := image.NewLayer(nil)
 	img := image.Image{