feat: CPEs format decoder (#4207)

Signed-off-by: Adam Chovanec <git@adamchovanec.cz>

syft/format/cpes/decoder.go

@@ -0,0 +1,95 @@
+package cpes
+
+import (
+	"bufio"
+	"errors"
+	"fmt"
+	"io"
+	"strings"
+
+	"github.com/anchore/syft/internal/log"
+	"github.com/anchore/syft/syft/cpe"
+	"github.com/anchore/syft/syft/format/internal"
+	"github.com/anchore/syft/syft/pkg"
+	"github.com/anchore/syft/syft/sbom"
+)
+
+const ID sbom.FormatID = "cpes"
+const version = "1"
+
+var _ sbom.FormatDecoder = (*decoder)(nil)
+
+type decoder struct{}
+
+func NewFormatDecoder() sbom.FormatDecoder {
+	return decoder{}
+}
+
+func (d decoder) Decode(r io.Reader) (*sbom.SBOM, sbom.FormatID, string, error) {
+	if r == nil {
+		return nil, "", "", fmt.Errorf("no reader provided")
+	}
+	s, err := toSyftModel(r)
+	return s, ID, version, err
+}
+
+func (d decoder) Identify(r io.Reader) (sbom.FormatID, string) {
+	if r == nil {
+		return "", ""
+	}
+
+	scanner := bufio.NewScanner(r)
+	for scanner.Scan() {
+		line := strings.TrimSpace(scanner.Text())
+		if line == "" {
+			// skip whitespace only lines
+			continue
+		}
+
+		err := cpe.ValidateString(line)
+		if err != nil {
+			return "", ""
+		}
+
+		return ID, version
+	}
+
+	return "", ""
+}
+
+func toSyftModel(r io.Reader) (*sbom.SBOM, error) {
+	var errs []error
+	pkgs := pkg.NewCollection()
+
+	scanner := bufio.NewScanner(r)
+	for scanner.Scan() {
+		line := strings.TrimSpace(scanner.Text())
+
+		if line == "" {
+			continue
+		}
+
+		// skip invalid CPEs
+		c, err := cpe.New(line, "")
+		if err != nil {
+			log.WithFields("error", err, "line", line).Debug("unable to parse cpe")
+			continue
+		}
+
+		p := pkg.Package{
+			Name:    c.Attributes.Product,
+			Version: c.Attributes.Version,
+			CPEs:    []cpe.CPE{c},
+		}
+
+		internal.Backfill(&p)
+		p.SetID()
+		pkgs.Add(p)
+	}
+
+	return &sbom.SBOM{
+		Artifacts: sbom.Artifacts{
+			Packages: pkgs,
+		},
+	}, errors.Join(errs...)
+}

syft/format/cpes/decoder_test.go

@@ -0,0 +1,171 @@
+package cpes
+
+import (
+	"strings"
+	"testing"
+
+	"github.com/google/go-cmp/cmp"
+	"github.com/google/go-cmp/cmp/cmpopts"
+	"github.com/stretchr/testify/require"
+
+	"github.com/anchore/syft/syft/cpe"
+	"github.com/anchore/syft/syft/file"
+	"github.com/anchore/syft/syft/pkg"
+	"github.com/anchore/syft/syft/sbom"
+)
+
+func Test_CPEProvider(t *testing.T) {
+	tests := []struct {
+		name      string
+		userInput string
+		sbom      *sbom.SBOM
+	}{
+		{
+			name:      "takes a single cpe",
+			userInput: "cpe:/a:apache:log4j:2.14.1",
+			sbom: &sbom.SBOM{
+				Artifacts: sbom.Artifacts{
+					Packages: pkg.NewCollection(pkg.Package{
+						Name:    "log4j",
+						Version: "2.14.1",
+						CPEs: []cpe.CPE{
+							cpe.Must("cpe:/a:apache:log4j:2.14.1", ""),
+						},
+					}),
+				},
+			},
+		},
+		{
+			name: "takes multiple cpes",
+			userInput: `cpe:/a:apache:log4j:2.14.1
+						cpe:2.3:a:f5:nginx:*:*:*:*:*:*:*:*;
+						cpe:2.3:a:f5:nginx:0.5.2:*:*:*:*:*:*:*;
+						cpe:2.3:a:f5:nginx:0.5.3:*:*:*:*:*:*:*;`,
+			sbom: &sbom.SBOM{
+				Artifacts: sbom.Artifacts{
+					Packages: pkg.NewCollection(
+						pkg.Package{
+							Name:    "log4j",
+							Version: "2.14.1",
+							CPEs: []cpe.CPE{
+								cpe.Must("cpe:/a:apache:log4j:2.14.1", ""),
+							},
+						},
+						pkg.Package{
+							Name:    "nginx",
+							Version: "",
+							CPEs: []cpe.CPE{
+								cpe.Must("cpe:2.3:a:f5:nginx:*:*:*:*:*:*:*:*;", ""),
+							},
+						},
+						pkg.Package{
+							Name:    "nginx",
+							Version: "0.5.2",
+							CPEs: []cpe.CPE{
+								cpe.Must("cpe:2.3:a:f5:nginx:0.5.2:*:*:*:*:*:*:*;", ""),
+							},
+						},
+						pkg.Package{
+							Name:    "nginx",
+							Version: "0.5.3",
+							CPEs: []cpe.CPE{
+								cpe.Must("cpe:2.3:a:f5:nginx:0.5.3:*:*:*:*:*:*:*;", ""),
+							},
+						},
+					),
+				},
+			},
+		},
+		{
+			name:      "takes cpe with no version",
+			userInput: "cpe:/a:apache:log4j",
+			sbom: &sbom.SBOM{
+				Artifacts: sbom.Artifacts{
+					Packages: pkg.NewCollection(pkg.Package{
+						Name: "log4j",
+						CPEs: []cpe.CPE{
+							cpe.Must("cpe:/a:apache:log4j", ""),
+						},
+					}),
+				},
+			},
+		},
+		{
+			name:      "takes CPE 2.3 format",
+			userInput: "cpe:2.3:a:apache:log4j:2.14.1:*:*:*:*:*:*:*",
+			sbom: &sbom.SBOM{
+				Artifacts: sbom.Artifacts{
+					Packages: pkg.NewCollection(pkg.Package{
+						Name:    "log4j",
+						Version: "2.14.1",
+						CPEs: []cpe.CPE{
+							cpe.Must("cpe:2.3:a:apache:log4j:2.14.1:*:*:*:*:*:*:*", ""),
+						},
+					}),
+				},
+			},
+		},
+		{
+			name:      "deduces target SW from CPE - known target_sw",
+			userInput: "cpe:2.3:a:amazon:opensearch:*:*:*:*:*:ruby:*:*",
+			sbom: &sbom.SBOM{
+				Artifacts: sbom.Artifacts{
+					Packages: pkg.NewCollection(pkg.Package{
+						Name: "opensearch",
+						Type: pkg.GemPkg,
+						CPEs: []cpe.CPE{
+							cpe.Must("cpe:2.3:a:amazon:opensearch:*:*:*:*:*:ruby:*:*", ""),
+						},
+					}),
+				},
+			},
+		},
+		{
+			name:      "handles unknown target_sw CPE field",
+			userInput: "cpe:2.3:a:amazon:opensearch:*:*:*:*:*:loremipsum:*:*",
+			sbom: &sbom.SBOM{
+				Artifacts: sbom.Artifacts{
+					Packages: pkg.NewCollection(pkg.Package{
+						Name: "opensearch",
+						Type: "",
+						CPEs: []cpe.CPE{
+							cpe.Must("cpe:2.3:a:amazon:opensearch:*:*:*:*:*:loremipsum:*:*", ""),
+						},
+					}),
+				},
+			},
+		},
+		{
+			name:      "invalid prefix",
+			userInput: "dir:test-fixtures/cpe",
+			sbom: &sbom.SBOM{
+				Artifacts: sbom.Artifacts{
+					Packages: pkg.NewCollection(),
+				},
+			},
+		},
+	}
+
+	syftPkgOpts := []cmp.Option{
+		cmpopts.IgnoreFields(pkg.Package{}, "id", "Language"),
+		cmpopts.IgnoreUnexported(pkg.Package{}, file.LocationSet{}, pkg.LicenseSet{}),
+	}
+
+	for _, tc := range tests {
+		t.Run(tc.name, func(t *testing.T) {
+			dec := NewFormatDecoder()
+
+			decodedSBOM, _, _, err := dec.Decode(strings.NewReader(tc.userInput))
+			require.NoError(t, err)
+
+			gotSyftPkgs := decodedSBOM.Artifacts.Packages.Sorted()
+			wantSyftPkgs := tc.sbom.Artifacts.Packages.Sorted()
+			require.Equal(t, len(gotSyftPkgs), len(wantSyftPkgs))
+			for idx, wantPkg := range wantSyftPkgs {
+				if d := cmp.Diff(wantPkg, gotSyftPkgs[idx], syftPkgOpts...); d != "" {
+					t.Errorf("unexpected Syft Package (-want +got):\n%s", d)
+				}
+			}
+		})
+	}
+}

syft/format/decoders.go

@@ -3,6 +3,7 @@ package format
 import (
 	"io"
 
+	"github.com/anchore/syft/syft/format/cpes"
 	"github.com/anchore/syft/syft/format/cyclonedxjson"
 	"github.com/anchore/syft/syft/format/cyclonedxxml"
 	"github.com/anchore/syft/syft/format/purls"
@@ -26,6 +27,7 @@ func Decoders() []sbom.FormatDecoder {
 		spdxtagvalue.NewFormatDecoder(),
 		spdxjson.NewFormatDecoder(),
 		purls.NewFormatDecoder(),
+		cpes.NewFormatDecoder(),
 	}
 }
 

syft/format/internal/backfill.go

@@ -10,13 +10,31 @@ import (
 	"github.com/anchore/syft/internal/log"
 	"github.com/anchore/syft/syft/cpe"
 	"github.com/anchore/syft/syft/pkg"
+	cataloger "github.com/anchore/syft/syft/pkg/cataloger/common/cpe"
 )
 
 // Backfill takes all information present in the package and attempts to fill in any missing information
-// from any available sources, such as the Metadata and PURL.
+// from any available sources, such as the Metadata, PURL, or CPEs.
 //
 // Backfill does not call p.SetID(), but this needs to be called later to ensure it's up to date
 func Backfill(p *pkg.Package) {
+	backfillFromPurl(p)
+	backfillFromCPE(p)
+}
+
+func backfillFromCPE(p *pkg.Package) {
+	if len(p.CPEs) == 0 {
+		return
+	}
+
+	c := p.CPEs[0]
+
+	if p.Type == "" {
+		p.Type = cataloger.TargetSoftwareToPackageType(c.Attributes.TargetSW)
+	}
+}
+
+func backfillFromPurl(p *pkg.Package) {
 	if p.PURL == "" {
 		return
 	}

syft/format/internal/backfill_test.go

@@ -121,6 +121,20 @@ func Test_Backfill(t *testing.T) {
 				Metadata: pkg.JavaArchive{},
 			},
 		},
+		{
+			name: "target-sw from CPE",
+			in: pkg.Package{
+				CPEs: []cpe.CPE{
+					cpe.Must("cpe:2.3:a:amazon:opensearch:*:*:*:*:*:ruby:*:*", ""),
+				},
+			},
+			expected: pkg.Package{
+				CPEs: []cpe.CPE{
+					cpe.Must("cpe:2.3:a:amazon:opensearch:*:*:*:*:*:ruby:*:*", ""),
+				},
+				Type: pkg.GemPkg,
+			},
+		},
 	}
 	for _, tt := range tests {
 		t.Run(tt.name, func(t *testing.T) {

syft/pkg/cataloger/common/cpe/target_software_to_pkg_type.go

@@ -0,0 +1,58 @@
+package cpe
+
+import (
+	"strings"
+
+	"github.com/anchore/syft/syft/pkg"
+)
+
+// TargetSoftwareToPackageType is derived from looking at target_software attributes in the NVD dataset
+// TODO: ideally this would be driven from the store, where we can resolve ecosystem aliases directly
+func TargetSoftwareToPackageType(tsw string) pkg.Type {
+	tsw = strings.NewReplacer("-", "_", " ", "_").Replace(strings.ToLower(tsw))
+	switch tsw {
+	case "alpine", "apk":
+		return pkg.ApkPkg
+	case "debian", "dpkg":
+		return pkg.DebPkg
+	case "java", "maven", "ant", "gradle", "jenkins", "jenkins_ci", "kafka", "logstash", "mule", "nifi", "solr", "spark", "storm", "struts", "tomcat", "zookeeper", "log4j":
+		return pkg.JavaPkg
+	case "javascript", "node", "nodejs", "node.js", "npm", "yarn", "apache", "jquery", "next.js", "prismjs":
+		return pkg.NpmPkg
+	case "c", "c++", "c/c++", "conan", "gnu_c++", "qt":
+		return pkg.ConanPkg
+	case "dart":
+		return pkg.DartPubPkg
+	case "redhat", "rpm", "redhat_enterprise_linux", "rhel", "suse", "suse_linux", "opensuse", "opensuse_linux", "fedora", "centos", "oracle_linux", "ol":
+		return pkg.RpmPkg
+	case "elixir", "hex":
+		return pkg.HexPkg
+	case "erlang":
+		return pkg.ErlangOTPPkg
+	case ".net", ".net_framework", "asp", "asp.net", "dotnet", "dotnet_framework", "c#", "csharp", "nuget":
+		return pkg.DotnetPkg
+	case "ruby", "gem", "nokogiri", "ruby_on_rails":
+		return pkg.GemPkg
+	case "rust", "cargo", "crates":
+		return pkg.RustPkg
+	case "python", "pip", "pypi", "flask":
+		return pkg.PythonPkg
+	case "kb", "knowledgebase", "msrc", "mskb", "microsoft":
+		return pkg.KbPkg
+	case "portage", "gentoo":
+		return pkg.PortagePkg
+	case "go", "golang", "gomodule":
+		return pkg.GoModulePkg
+	case "linux_kernel", "linux", "z/linux":
+		return pkg.LinuxKernelPkg
+	case "php":
+		return pkg.PhpComposerPkg
+	case "swift":
+		return pkg.SwiftPkg
+	case "wordpress", "wordpress_plugin", "wordpress_":
+		return pkg.WordpressPluginPkg
+	case "lua", "luarocks":
+		return pkg.LuaRocksPkg
+	}
+	return ""
+}