Migrate format definitions to sbom package (#864)

syft/encode_decode.go

@@ -6,18 +6,10 @@ import (
 	"io"
 
 	"github.com/anchore/syft/syft/sbom"
-
-	"github.com/anchore/syft/internal/formats"
-	"github.com/anchore/syft/syft/format"
 )
 
 // Encode takes all SBOM elements and a format option and encodes an SBOM document.
-func Encode(s sbom.SBOM, option format.Option) ([]byte, error) {
-	f := formats.ByOption(option)
-	if f == nil {
-		return nil, fmt.Errorf("unsupported format: %+v", option)
-	}
-
+func Encode(s sbom.SBOM, f sbom.Format) ([]byte, error) {
 	buff := bytes.Buffer{}
 
 	if err := f.Encode(&buff, s); err != nil {
@@ -28,19 +20,17 @@ func Encode(s sbom.SBOM, option format.Option) ([]byte, error) {
 }
 
 // Decode takes a reader for an SBOM and generates all internal SBOM elements.
-func Decode(reader io.Reader) (*sbom.SBOM, format.Option, error) {
+func Decode(reader io.Reader) (*sbom.SBOM, sbom.Format, error) {
 	by, err := io.ReadAll(reader)
 	if err != nil {
-		return nil, format.UnknownFormatOption, fmt.Errorf("unable to read sbom: %w", err)
+		return nil, nil, fmt.Errorf("unable to read sbom: %w", err)
 	}
 
-	f, err := formats.Identify(by)
-	if err != nil {
-		return nil, format.UnknownFormatOption, fmt.Errorf("unable to detect format: %w", err)
-	}
+	f := IdentifyFormat(by)
 	if f == nil {
-		return nil, format.UnknownFormatOption, fmt.Errorf("unable to identify format")
+		return nil, nil, fmt.Errorf("unable to identify format")
 	}
+
 	s, err := f.Decode(bytes.NewReader(by))
-	return s, f.Option, err
+	return s, f, err
 }

syft/format/decoder.go

@@ -1,10 +0,0 @@
-package format
-
-import (
-	"io"
-
-	"github.com/anchore/syft/syft/sbom"
-)
-
-// Decoder is a function that can convert an SBOM document of a specific format from a reader into Syft native objects.
-type Decoder func(reader io.Reader) (*sbom.SBOM, error)

syft/format/encoder.go

@@ -1,10 +0,0 @@
-package format
-
-import (
-	"io"
-
-	"github.com/anchore/syft/syft/sbom"
-)
-
-// Encoder is a function that can transform Syft native objects into an SBOM document of a specific format written to the given writer.
-type Encoder func(io.Writer, sbom.SBOM) error

syft/format/format.go

@@ -1,52 +0,0 @@
-package format
-
-import (
-	"errors"
-	"io"
-
-	"github.com/anchore/syft/syft/sbom"
-)
-
-var (
-	ErrEncodingNotSupported   = errors.New("encoding not supported")
-	ErrDecodingNotSupported   = errors.New("decoding not supported")
-	ErrValidationNotSupported = errors.New("validation not supported")
-)
-
-type Format struct {
-	Option    Option
-	encoder   Encoder
-	decoder   Decoder
-	validator Validator
-}
-
-func NewFormat(option Option, encoder Encoder, decoder Decoder, validator Validator) Format {
-	return Format{
-		Option:    option,
-		encoder:   encoder,
-		decoder:   decoder,
-		validator: validator,
-	}
-}
-
-func (f Format) Encode(output io.Writer, s sbom.SBOM) error {
-	if f.encoder == nil {
-		return ErrEncodingNotSupported
-	}
-	return f.encoder(output, s)
-}
-
-func (f Format) Decode(reader io.Reader) (*sbom.SBOM, error) {
-	if f.decoder == nil {
-		return nil, ErrDecodingNotSupported
-	}
-	return f.decoder(reader)
-}
-
-func (f Format) Validate(reader io.Reader) error {
-	if f.validator == nil {
-		return ErrValidationNotSupported
-	}
-
-	return f.validator(reader)
-}

syft/format/option.go

@@ -1,49 +0,0 @@
-package format
-
-import "strings"
-
-const (
-	UnknownFormatOption Option = "UnknownFormatOption"
-	JSONOption          Option = "json"
-	TextOption          Option = "text"
-	TableOption         Option = "table"
-	CycloneDxXMLOption  Option = "cyclonedx"
-	CycloneDxJSONOption Option = "cyclonedx-json"
-	SPDXTagValueOption  Option = "spdx-tag-value"
-	SPDXJSONOption      Option = "spdx-json"
-)
-
-var AllOptions = []Option{
-	JSONOption,
-	TextOption,
-	TableOption,
-	CycloneDxXMLOption,
-	CycloneDxJSONOption,
-	SPDXTagValueOption,
-	SPDXJSONOption,
-}
-
-type Option string
-
-func ParseOption(userStr string) Option {
-	switch strings.ToLower(userStr) {
-	case string(JSONOption):
-		return JSONOption
-	case string(TextOption):
-		return TextOption
-	case string(TableOption):
-		return TableOption
-	case string(CycloneDxXMLOption), "cyclone", "cyclone-dx", "cyclone-dx-xml", "cyclone-xml":
-		// NOTE(jonasagx): setting "cyclone" to XML by default for retro-compatibility.
-		// If we want to show no preference between XML and JSON please remove it.
-		return CycloneDxXMLOption
-	case string(CycloneDxJSONOption), "cyclone-json", "cyclone-dx-json":
-		return CycloneDxJSONOption
-	case string(SPDXTagValueOption), "spdx", "spdx-tagvalue", "spdxtagvalue", "spdx-tv", "spdxtv":
-		return SPDXTagValueOption
-	case string(SPDXJSONOption), "spdxjson":
-		return SPDXJSONOption
-	default:
-		return UnknownFormatOption
-	}
-}

syft/format/validator.go

@@ -1,12 +0,0 @@
-package format
-
-import "io"
-
-// Validator reads the SBOM from the given reader and assesses whether the document conforms to the specific SBOM format.
-// The validator should positively confirm if the SBOM is not only the format but also has the minimal set of values
-// that the format requires. For example, all syftjson formatted documents have a schema section which should have
-// "anchore/syft" within the version --if this isn't found then the validator should raise an error. These active
-// assertions protect against "simple" format decoding validations that may lead to false positives (e.g. I decoded
-// json successfully therefore this must be the target format, however, all values are their default zero-value and
-// really represent a different format that also uses json)
-type Validator func(reader io.Reader) error

syft/formats.go

@@ -0,0 +1,100 @@
+package syft
+
+import (
+	"bytes"
+	"strings"
+
+	"github.com/anchore/syft/internal/formats/cyclonedx13json"
+	"github.com/anchore/syft/internal/formats/cyclonedx13xml"
+	"github.com/anchore/syft/internal/formats/spdx22json"
+	"github.com/anchore/syft/internal/formats/spdx22tagvalue"
+	"github.com/anchore/syft/internal/formats/syftjson"
+	"github.com/anchore/syft/internal/formats/table"
+	"github.com/anchore/syft/internal/formats/text"
+	"github.com/anchore/syft/syft/sbom"
+)
+
+// these have been exported for the benefit of API users
+const (
+	JSONFormatID          = syftjson.ID
+	TextFormatID          = text.ID
+	TableFormatID         = table.ID
+	CycloneDxXMLFormatID  = cyclonedx13xml.ID
+	CycloneDxJSONFormatID = cyclonedx13json.ID
+	SPDXTagValueFormatID  = spdx22tagvalue.ID
+	SPDXJSONFormatID      = spdx22json.ID
+)
+
+var formats []sbom.Format
+
+func init() {
+	formats = []sbom.Format{
+		syftjson.Format(),
+		cyclonedx13xml.Format(),
+		cyclonedx13json.Format(),
+		spdx22tagvalue.Format(),
+		spdx22json.Format(),
+		table.Format(),
+		text.Format(),
+	}
+}
+
+func FormatIDs() (ids []sbom.FormatID) {
+	for _, f := range formats {
+		ids = append(ids, f.ID())
+	}
+	return ids
+}
+
+func FormatByID(id sbom.FormatID) sbom.Format {
+	for _, f := range formats {
+		if f.ID() == id {
+			return f
+		}
+	}
+	return nil
+}
+
+func FormatByName(name string) sbom.Format {
+	cleanName := cleanFormatName(name)
+	for _, f := range formats {
+		if cleanFormatName(string(f.ID())) == cleanName {
+			return f
+		}
+	}
+
+	// handle any aliases for any supported format
+	switch cleanName {
+	case "json", "syftjson":
+		return FormatByID(syftjson.ID)
+	case "cyclonedx", "cyclone", "cyclonedxxml":
+		return FormatByID(cyclonedx13xml.ID)
+	case "cyclonedxjson":
+		return FormatByID(cyclonedx13json.ID)
+	case "spdx", "spdxtv", "spdxtagvalue":
+		return FormatByID(spdx22tagvalue.ID)
+	case "spdxjson":
+		return FormatByID(spdx22json.ID)
+	case "table":
+		return FormatByID(table.ID)
+	case "text":
+		return FormatByID(text.ID)
+	}
+
+	return nil
+}
+
+func cleanFormatName(name string) string {
+	r := strings.NewReplacer("-", "", "_", "")
+	return strings.ToLower(r.Replace(name))
+}
+
+func IdentifyFormat(by []byte) sbom.Format {
+	for _, f := range formats {
+		if err := f.Validate(bytes.NewReader(by)); err != nil {
+			continue
+		}
+		return f
+	}
+	return nil
+}

syft/formats_test.go

@@ -0,0 +1,157 @@
+package syft
+
+import (
+	"github.com/anchore/syft/internal/formats/cyclonedx13json"
+	"github.com/anchore/syft/internal/formats/cyclonedx13xml"
+	"github.com/anchore/syft/internal/formats/spdx22json"
+	"github.com/anchore/syft/internal/formats/spdx22tagvalue"
+	"github.com/anchore/syft/internal/formats/syftjson"
+	"github.com/anchore/syft/internal/formats/table"
+	"github.com/anchore/syft/internal/formats/text"
+	"github.com/anchore/syft/syft/sbom"
+	"github.com/stretchr/testify/require"
+	"io"
+	"os"
+	"testing"
+
+	"github.com/stretchr/testify/assert"
+)
+
+func TestIdentify(t *testing.T) {
+	tests := []struct {
+		fixture  string
+		expected sbom.FormatID
+	}{
+		{
+			fixture:  "test-fixtures/alpine-syft.json",
+			expected: syftjson.ID,
+		},
+	}
+	for _, test := range tests {
+		t.Run(test.fixture, func(t *testing.T) {
+			f, err := os.Open(test.fixture)
+			assert.NoError(t, err)
+			by, err := io.ReadAll(f)
+			assert.NoError(t, err)
+			frmt := IdentifyFormat(by)
+			assert.NotNil(t, frmt)
+			assert.Equal(t, test.expected, frmt.ID())
+		})
+	}
+}
+
+func TestFormatByName(t *testing.T) {
+
+	tests := []struct {
+		name string
+		want sbom.FormatID
+	}{
+		// SPDX Tag-Value
+		{
+			name: "spdx",
+			want: spdx22tagvalue.ID,
+		},
+		{
+			name: "spdx-tag-value",
+			want: spdx22tagvalue.ID,
+		},
+		{
+			name: "spdx-tv",
+			want: spdx22tagvalue.ID,
+		},
+		{
+			name: "spdxtv", // clean variant
+			want: spdx22tagvalue.ID,
+		},
+		{
+			name: "spdx-2-tag-value", // clean variant
+			want: spdx22tagvalue.ID,
+		},
+		{
+			name: "spdx-2-tagvalue", // clean variant
+			want: spdx22tagvalue.ID,
+		},
+		{
+			name: "spdx2-tagvalue", // clean variant
+			want: spdx22tagvalue.ID,
+		},
+
+		// SPDX JSON
+		{
+			name: "spdx-json",
+			want: spdx22json.ID,
+		},
+		{
+			name: "spdx-2-json",
+			want: spdx22json.ID,
+		},
+
+		// Cyclonedx JSON
+		{
+			name: "cyclonedx-json",
+			want: cyclonedx13json.ID,
+		},
+		{
+			name: "cyclonedx-1-json",
+			want: cyclonedx13json.ID,
+		},
+
+		// Cyclonedx XML
+		{
+			name: "cyclonedx",
+			want: cyclonedx13xml.ID,
+		},
+		{
+			name: "cyclonedx-xml",
+			want: cyclonedx13xml.ID,
+		},
+		{
+			name: "cyclonedx-1-xml",
+			want: cyclonedx13xml.ID,
+		},
+
+		// Syft Table
+		{
+			name: "table",
+			want: table.ID,
+		},
+
+		{
+			name: "syft-table",
+			want: table.ID,
+		},
+
+		// Syft Text
+		{
+			name: "text",
+			want: text.ID,
+		},
+
+		{
+			name: "syft-text",
+			want: text.ID,
+		},
+
+		// Syft JSON
+		{
+			name: "json",
+			want: syftjson.ID,
+		},
+
+		{
+			name: "syft-json",
+			want: syftjson.ID,
+		},
+	}
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			f := FormatByName(tt.name)
+			if tt.want == "" {
+				require.Nil(t, f)
+				return
+			}
+			require.NotNil(t, f)
+			assert.Equal(t, tt.want, f.ID())
+		})
+	}
+}

syft/pkg/cataloger/ruby/parse_gemfile_lock.go

@@ -14,7 +14,7 @@ import (
 // integrity check
 var _ common.ParserFn = parseGemFileLockEntries
 
-var sectionsOfInterest = internal.NewStringSetFromSlice([]string{"GEM"})
+var sectionsOfInterest = internal.NewStringSet("GEM")
 
 // parseGemFileLockEntries is a parser function for Gemfile.lock contents, returning all Gems discovered.
 func parseGemFileLockEntries(_ string, reader io.Reader) ([]*pkg.Package, []artifact.Relationship, error) {

syft/sbom/format.go

@@ -0,0 +1,78 @@
+package sbom
+
+import (
+	"errors"
+	"io"
+)
+
+var (
+	ErrEncodingNotSupported   = errors.New("encoding not supported")
+	ErrDecodingNotSupported   = errors.New("decoding not supported")
+	ErrValidationNotSupported = errors.New("validation not supported")
+)
+
+type FormatID string
+
+type Format interface {
+	ID() FormatID
+	Encode(io.Writer, SBOM) error
+	Decode(io.Reader) (*SBOM, error)
+	Validate(io.Reader) error
+}
+
+type format struct {
+	id        FormatID
+	encoder   Encoder
+	decoder   Decoder
+	validator Validator
+}
+
+// Decoder is a function that can convert an SBOM document of a specific format from a reader into Syft native objects.
+type Decoder func(reader io.Reader) (*SBOM, error)
+
+// Encoder is a function that can transform Syft native objects into an SBOM document of a specific format written to the given writer.
+type Encoder func(io.Writer, SBOM) error
+
+// Validator reads the SBOM from the given reader and assesses whether the document conforms to the specific SBOM format.
+// The validator should positively confirm if the SBOM is not only the format but also has the minimal set of values
+// that the format requires. For example, all syftjson formatted documents have a schema section which should have
+// "anchore/syft" within the version --if this isn't found then the validator should raise an error. These active
+// assertions protect against "simple" format decoding validations that may lead to false positives (e.g. I decoded
+// json successfully therefore this must be the target format, however, all values are their default zero-value and
+// really represent a different format that also uses json)
+type Validator func(reader io.Reader) error
+
+func NewFormat(id FormatID, encoder Encoder, decoder Decoder, validator Validator) Format {
+	return &format{
+		id:        id,
+		encoder:   encoder,
+		decoder:   decoder,
+		validator: validator,
+	}
+}
+
+func (f format) ID() FormatID {
+	return f.id
+}
+
+func (f format) Encode(output io.Writer, s SBOM) error {
+	if f.encoder == nil {
+		return ErrEncodingNotSupported
+	}
+	return f.encoder(output, s)
+}
+
+func (f format) Decode(reader io.Reader) (*SBOM, error) {
+	if f.decoder == nil {
+		return nil, ErrDecodingNotSupported
+	}
+	return f.decoder(reader)
+}
+
+func (f format) Validate(reader io.Reader) error {
+	if f.validator == nil {
+		return ErrValidationNotSupported
+	}
+
+	return f.validator(reader)
+}

syft/sbom/multi_writer.go

@@ -0,0 +1,95 @@
+package sbom
+
+import (
+	"fmt"
+	"os"
+	"path"
+
+	"github.com/anchore/syft/internal/log"
+
+	"github.com/hashicorp/go-multierror"
+)
+
+// multiWriter holds a list of child sbom.Writers to apply all Write and Close operations to
+type multiWriter struct {
+	writers []Writer
+}
+
+// WriterOption Format and path strings used to create sbom.Writer
+type WriterOption struct {
+	Format Format
+	Path   string
+}
+
+// NewWriter create all report writers from input options; if a file is not specified, os.Stdout is used
+func NewWriter(options ...WriterOption) (Writer, error) {
+	if len(options) == 0 {
+		return nil, fmt.Errorf("no output options provided")
+	}
+
+	out := &multiWriter{}
+
+	defer func() {
+		// close any previously opened files; we can't really recover from any errors
+		if err := out.Close(); err != nil {
+			log.Warnf("unable to close sbom writers: %+v", err)
+		}
+	}()
+
+	for _, option := range options {
+		switch len(option.Path) {
+		case 0:
+			out.writers = append(out.writers, &streamWriter{
+				format: option.Format,
+				out:    os.Stdout,
+			})
+		default:
+			// create any missing subdirectories
+			dir := path.Dir(option.Path)
+			if dir != "" {
+				s, err := os.Stat(dir)
+				if err != nil {
+					err = os.MkdirAll(dir, 0755) // maybe should be os.ModePerm ?
+					if err != nil {
+						return nil, err
+					}
+				} else if !s.IsDir() {
+					return nil, fmt.Errorf("output path does not contain a valid directory: %s", option.Path)
+				}
+			}
+			fileOut, err := os.OpenFile(option.Path, os.O_RDWR|os.O_CREATE|os.O_TRUNC, 0644)
+			if err != nil {
+				return nil, fmt.Errorf("unable to create report file: %w", err)
+			}
+			out.writers = append(out.writers, &streamWriter{
+				format: option.Format,
+				out:    fileOut,
+				close:  fileOut.Close,
+			})
+		}
+	}
+
+	return out, nil
+}
+
+// Write writes the SBOM to all writers
+func (m *multiWriter) Write(s SBOM) (errs error) {
+	for _, w := range m.writers {
+		err := w.Write(s)
+		if err != nil {
+			errs = multierror.Append(errs, err)
+		}
+	}
+	return errs
+}
+
+// Close closes all writers
+func (m *multiWriter) Close() (errs error) {
+	for _, w := range m.writers {
+		err := w.Close()
+		if err != nil {
+			errs = multierror.Append(errs, err)
+		}
+	}
+	return errs
+}

syft/sbom/multi_writer_test.go

@@ -0,0 +1,172 @@
+package sbom
+
+import (
+	"io"
+	"strings"
+	"testing"
+
+	"github.com/stretchr/testify/assert"
+)
+
+func dummyEncoder(io.Writer, SBOM) error {
+	return nil
+}
+
+func dummyFormat(name string) Format {
+	return NewFormat(FormatID(name), dummyEncoder, nil, nil)
+}
+
+type writerConfig struct {
+	format string
+	file   string
+}
+
+func TestOutputWriter(t *testing.T) {
+	tmp := t.TempDir()
+
+	testName := func(options []WriterOption, err bool) string {
+		var out []string
+		for _, opt := range options {
+			out = append(out, string(opt.Format.ID())+"="+opt.Path)
+		}
+		errs := ""
+		if err {
+			errs = "(err)"
+		}
+		return strings.Join(out, ", ") + errs
+	}
+
+	tests := []struct {
+		outputs  []WriterOption
+		err      bool
+		expected []writerConfig
+	}{
+		{
+			outputs: []WriterOption{},
+			err:     true,
+		},
+		{
+			outputs: []WriterOption{
+				{
+					Format: dummyFormat("table"),
+					Path:   "",
+				},
+			},
+			expected: []writerConfig{
+				{
+					format: "table",
+				},
+			},
+		},
+		{
+			outputs: []WriterOption{
+				{
+					Format: dummyFormat("json"),
+				},
+			},
+			expected: []writerConfig{
+				{
+					format: "json",
+				},
+			},
+		},
+		{
+			outputs: []WriterOption{
+				{
+					Format: dummyFormat("json"),
+					Path:   "test-2.json",
+				},
+			},
+			expected: []writerConfig{
+				{
+					format: "json",
+					file:   "test-2.json",
+				},
+			},
+		},
+		{
+			outputs: []WriterOption{
+				{
+					Format: dummyFormat("json"),
+					Path:   "test-3/1.json",
+				},
+				{
+					Format: dummyFormat("spdx-json"),
+					Path:   "test-3/2.json",
+				},
+			},
+			expected: []writerConfig{
+				{
+					format: "json",
+					file:   "test-3/1.json",
+				},
+				{
+					format: "spdx-json",
+					file:   "test-3/2.json",
+				},
+			},
+		},
+		{
+			outputs: []WriterOption{
+				{
+					Format: dummyFormat("text"),
+				},
+				{
+					Format: dummyFormat("spdx-json"),
+					Path:   "test-4.json",
+				},
+			},
+			expected: []writerConfig{
+				{
+					format: "text",
+				},
+				{
+					format: "spdx-json",
+					file:   "test-4.json",
+				},
+			},
+		},
+	}
+
+	for _, test := range tests {
+		t.Run(testName(test.outputs, test.err), func(t *testing.T) {
+			outputs := test.outputs
+			for i := range outputs {
+				if outputs[i].Path != "" {
+					outputs[i].Path = tmp + outputs[i].Path
+				}
+			}
+
+			writer, err := NewWriter(outputs...)
+
+			if test.err {
+				assert.Error(t, err)
+				return
+			} else {
+				assert.NoError(t, err)
+			}
+
+			mw := writer.(*multiWriter)
+
+			assert.Len(t, mw.writers, len(test.expected))
+
+			for i, e := range test.expected {
+				w := mw.writers[i].(*streamWriter)
+
+				assert.Equal(t, string(w.format.ID()), e.format)
+
+				if e.file != "" {
+					assert.FileExists(t, tmp+e.file)
+				}
+
+				if e.file != "" {
+					assert.NotNil(t, w.out)
+					assert.NotNil(t, w.close)
+				} else {
+					assert.NotNil(t, w.out)
+					assert.Nil(t, w.close)
+				}
+			}
+		})
+	}
+}

syft/sbom/stream_writer.go

@@ -0,0 +1,25 @@
+package sbom
+
+import (
+	"io"
+)
+
+// streamWriter implements sbom.Writer for a given format and io.Writer, also providing a close function for cleanup
+type streamWriter struct {
+	format Format
+	out    io.Writer
+	close  func() error
+}
+
+// Write the provided SBOM to the data stream
+func (w *streamWriter) Write(s SBOM) error {
+	return w.format.Encode(w.out, s)
+}
+
+// Close any resources, such as open files
+func (w *streamWriter) Close() error {
+	if w.close != nil {
+		return w.close()
+	}
+	return nil
+}