Add compliance policy for empty name and version (#3257)

* add policy for empty name and version

Signed-off-by: Alex Goodman <wagoodman@users.noreply.github.com>

* default stub version

Signed-off-by: Alex Goodman <wagoodman@users.noreply.github.com>

* modifying ids requires augmenting relationships

Signed-off-by: Alex Goodman <wagoodman@users.noreply.github.com>

---------

Signed-off-by: Alex Goodman <wagoodman@users.noreply.github.com>

internal/relationship/index.go

@@ -17,20 +17,16 @@ type Index struct {
 
 // NewIndex returns a new relationship Index
 func NewIndex(relationships ...artifact.Relationship) *Index {
-	out := Index{}
+	out := Index{
+		fromID: make(map[artifact.ID]*mappedRelationships),
+		toID:   make(map[artifact.ID]*mappedRelationships),
+	}
 	out.Add(relationships...)
 	return &out
 }
 
 // Add adds all the given relationships to the index, without adding duplicates
 func (i *Index) Add(relationships ...artifact.Relationship) {
-	if i.fromID == nil {
-		i.fromID = map[artifact.ID]*mappedRelationships{}
-	}
-	if i.toID == nil {
-		i.toID = map[artifact.ID]*mappedRelationships{}
-	}
-
 	// store appropriate indexes for stable ordering to minimize ID() calls
 	for _, r := range relationships {
 		// prevent duplicates
@@ -68,6 +64,39 @@ func (i *Index) Add(relationships ...artifact.Relationship) {
 	}
 }
 
+func (i *Index) Remove(id artifact.ID) {
+	delete(i.fromID, id)
+	delete(i.toID, id)
+
+	for idx := 0; idx < len(i.all); {
+		if i.all[idx].from == id || i.all[idx].to == id {
+			i.all = append(i.all[:idx], i.all[idx+1:]...)
+		} else {
+			idx++
+		}
+	}
+}
+
+func (i *Index) Replace(ogID artifact.ID, replacement artifact.Identifiable) {
+	for _, mapped := range fromMappedByID(i.fromID, ogID) {
+		i.Add(artifact.Relationship{
+			From: replacement,
+			To:   mapped.relationship.To,
+			Type: mapped.relationship.Type,
+		})
+	}
+
+	for _, mapped := range fromMappedByID(i.toID, ogID) {
+		i.Add(artifact.Relationship{
+			From: mapped.relationship.From,
+			To:   replacement,
+			Type: mapped.relationship.Type,
+		})
+	}
+
+	i.Remove(ogID)
+}
+
 // From returns all relationships from the given identifiable, with specified types
 func (i *Index) From(identifiable artifact.Identifiable, types ...artifact.RelationshipType) []artifact.Relationship {
 	return toSortedSlice(fromMapped(i.fromID, identifiable), types)
@@ -110,10 +139,17 @@ func (i *Index) All(types ...artifact.RelationshipType) []artifact.Relationship
 }
 
 func fromMapped(idMap map[artifact.ID]*mappedRelationships, identifiable artifact.Identifiable) []*sortableRelationship {
-	if identifiable == nil || idMap == nil {
+	if identifiable == nil {
 		return nil
 	}
-	mapped := idMap[identifiable.ID()]
+	return fromMappedByID(idMap, identifiable.ID())
+}
+
+func fromMappedByID(idMap map[artifact.ID]*mappedRelationships, id artifact.ID) []*sortableRelationship {
+	if idMap == nil {
+		return nil
+	}
+	mapped := idMap[id]
 	if mapped == nil {
 		return nil
 	}

internal/relationship/index_test.go

@@ -3,6 +3,7 @@ package relationship
 import (
 	"testing"
 
+	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"
 
 	"github.com/anchore/syft/syft/artifact"
@@ -231,3 +232,179 @@ func (i id) ID() artifact.ID {
 func slice[T any](values ...T) []T {
 	return values
 }
+
+func TestRemove(t *testing.T) {
+	p1 := pkg.Package{Name: "pkg-1"}
+	p2 := pkg.Package{Name: "pkg-2"}
+	p3 := pkg.Package{Name: "pkg-3"}
+	c1 := file.Coordinates{RealPath: "/coords/1"}
+	c2 := file.Coordinates{RealPath: "/coords/2"}
+	c3 := file.Coordinates{RealPath: "/coords/3"}
+	c4 := file.Coordinates{RealPath: "/coords/4"}
+
+	for _, p := range []*pkg.Package{&p1, &p2, &p3} {
+		p.SetID()
+	}
+
+	r1 := artifact.Relationship{
+		From: p1,
+		To:   p2,
+		Type: artifact.DependencyOfRelationship,
+	}
+	r2 := artifact.Relationship{
+		From: p1,
+		To:   p3,
+		Type: artifact.DependencyOfRelationship,
+	}
+	r3 := artifact.Relationship{
+		From: p1,
+		To:   c1,
+		Type: artifact.ContainsRelationship,
+	}
+	r4 := artifact.Relationship{
+		From: p2,
+		To:   c2,
+		Type: artifact.ContainsRelationship,
+	}
+	r5 := artifact.Relationship{
+		From: p3,
+		To:   c1,
+		Type: artifact.ContainsRelationship,
+	}
+	r6 := artifact.Relationship{
+		From: p3,
+		To:   c2,
+		Type: artifact.ContainsRelationship,
+	}
+	r7 := artifact.Relationship{
+		From: c1,
+		To:   c3,
+		Type: artifact.ContainsRelationship,
+	}
+	r8 := artifact.Relationship{
+		From: c3,
+		To:   c4,
+		Type: artifact.ContainsRelationship,
+	}
+
+	index := NewIndex(r1, r2, r3, r4, r5, r6, r7, r8)
+
+	assert.Equal(t, 8, len(index.All()))
+
+	// removal of p1 should remove r1, r2, and r3
+	index.Remove(p1.ID())
+	remaining := index.All()
+
+	assert.Equal(t, 5, len(remaining))
+	assert.NotContains(t, remaining, r1)
+	assert.NotContains(t, remaining, r2)
+	assert.NotContains(t, remaining, r3)
+
+	assert.Empty(t, index.From(p1))
+	assert.Empty(t, index.To(p1))
+
+	// removal of c1 should remove r5 and r7
+	index.Remove(c1.ID())
+	remaining = index.All()
+
+	// r8 remains since c3->c4 should still exist
+	assert.Equal(t, 3, len(remaining))
+	assert.NotContains(t, remaining, r5)
+	assert.NotContains(t, remaining, r7)
+	assert.Contains(t, remaining, r8)
+
+	assert.Empty(t, index.From(c1))
+	assert.Empty(t, index.To(c1))
+
+	// removal of c3 should remove r8
+	index.Remove(c3.ID())
+	remaining = index.All()
+
+	assert.Equal(t, 2, len(remaining))
+	assert.Contains(t, remaining, r4)
+	assert.Contains(t, remaining, r6)
+
+	assert.Empty(t, index.From(c3))
+	assert.Empty(t, index.To(c3))
+}
+
+func TestReplace(t *testing.T) {
+	p1 := pkg.Package{Name: "pkg-1"}
+	p2 := pkg.Package{Name: "pkg-2"}
+	p3 := pkg.Package{Name: "pkg-3"}
+	p4 := pkg.Package{Name: "pkg-4"}
+
+	for _, p := range []*pkg.Package{&p1, &p2, &p3, &p4} {
+		p.SetID()
+	}
+
+	r1 := artifact.Relationship{
+		From: p1,
+		To:   p2,
+		Type: artifact.DependencyOfRelationship,
+	}
+	r2 := artifact.Relationship{
+		From: p3,
+		To:   p1,
+		Type: artifact.DependencyOfRelationship,
+	}
+	r3 := artifact.Relationship{
+		From: p2,
+		To:   p3,
+		Type: artifact.ContainsRelationship,
+	}
+
+	index := NewIndex(r1, r2, r3)
+
+	// replace p1 with p4 in the relationships
+	index.Replace(p1.ID(), &p4)
+
+	expectedRels := []artifact.Relationship{
+		{
+			From: p4, // replaced
+			To:   p2,
+			Type: artifact.DependencyOfRelationship,
+		},
+		{
+			From: p3,
+			To:   p4, // replaced
+			Type: artifact.DependencyOfRelationship,
+		},
+		{
+			From: p2,
+			To:   p3,
+			Type: artifact.ContainsRelationship,
+		},
+	}
+
+	compareRelationships(t, expectedRels, index.All())
+}
+
+func compareRelationships(t testing.TB, expected, actual []artifact.Relationship) {
+	assert.Equal(t, len(expected), len(actual), "number of relationships should match")
+	for _, e := range expected {
+		found := false
+		for _, a := range actual {
+			if a.From.ID() == e.From.ID() && a.To.ID() == e.To.ID() && a.Type == e.Type {
+				found = true
+				break
+			}
+		}
+		assert.True(t, found, "expected relationship not found: %+v", e)
+	}
+}
+
+func TestReplace_NoExistingRelations(t *testing.T) {
+	p1 := pkg.Package{Name: "pkg-1"}
+	p2 := pkg.Package{Name: "pkg-2"}
+
+	p1.SetID()
+	p2.SetID()
+
+	index := NewIndex()
+
+	index.Replace(p1.ID(), &p2)
+
+	allRels := index.All()
+	assert.Len(t, allRels, 0)
+}

internal/task/package_task_factory.go

@@ -11,6 +11,7 @@ import (
 
 	"github.com/anchore/syft/internal/bus"
 	"github.com/anchore/syft/internal/log"
+	"github.com/anchore/syft/internal/relationship"
 	"github.com/anchore/syft/internal/sbomsync"
 	"github.com/anchore/syft/syft/artifact"
 	"github.com/anchore/syft/syft/cataloging"
@@ -27,6 +28,7 @@ type packageTaskFactory func(cfg CatalogingFactoryConfig) Task
 type PackageTaskFactories []packageTaskFactory
 
 type CatalogingFactoryConfig struct {
+	ComplianceConfig     cataloging.ComplianceConfig
 	SearchConfig         cataloging.SearchConfig
 	RelationshipsConfig  cataloging.RelationshipsConfig
 	DataGenerationConfig cataloging.DataGenerationConfig
@@ -35,6 +37,7 @@ type CatalogingFactoryConfig struct {
 
 func DefaultCatalogingFactoryConfig() CatalogingFactoryConfig {
 	return CatalogingFactoryConfig{
+		ComplianceConfig:     cataloging.DefaultComplianceConfig(),
 		SearchConfig:         cataloging.DefaultSearchConfig(),
 		RelationshipsConfig:  cataloging.DefaultRelationshipsConfig(),
 		DataGenerationConfig: cataloging.DefaultDataGenerationConfig(),
@@ -82,7 +85,7 @@ func (f PackageTaskFactories) Tasks(cfg CatalogingFactoryConfig) ([]Task, error)
 }
 
 // NewPackageTask creates a Task function for a generic pkg.Cataloger, honoring the common configuration options.
-func NewPackageTask(cfg CatalogingFactoryConfig, c pkg.Cataloger, tags ...string) Task { //nolint: funlen
+func NewPackageTask(cfg CatalogingFactoryConfig, c pkg.Cataloger, tags ...string) Task {
 	fn := func(ctx context.Context, resolver file.Resolver, sbom sbomsync.Builder) error {
 		catalogerName := c.Name()
 		log.WithFields("name", catalogerName).Trace("starting package cataloger")
@@ -106,42 +109,9 @@ func NewPackageTask(cfg CatalogingFactoryConfig, c pkg.Cataloger, tags ...string
 
 		log.WithFields("cataloger", catalogerName).Debugf("discovered %d packages", len(pkgs))
 
-		for i, p := range pkgs {
-			if p.FoundBy == "" {
-				p.FoundBy = catalogerName
-			}
+		pkgs, relationships = finalizePkgCatalogerResults(cfg, resolver, catalogerName, pkgs, relationships)
 
-			if cfg.DataGenerationConfig.GenerateCPEs && !hasAuthoritativeCPE(p.CPEs) {
-				// generate CPEs (note: this is excluded from package ID, so is safe to mutate)
-				// we might have binary classified CPE already with the package so we want to append here
-				dictionaryCPEs, ok := cpeutils.DictionaryFind(p)
-				if ok {
-					log.Tracef("used CPE dictionary to find CPEs for %s package %q: %s", p.Type, p.Name, dictionaryCPEs)
-					p.CPEs = append(p.CPEs, dictionaryCPEs...)
-				} else {
-					p.CPEs = append(p.CPEs, cpeutils.Generate(p)...)
-				}
-			}
-
-			// if we were not able to identify the language we have an opportunity
-			// to try and get this value from the PURL. Worst case we assert that
-			// we could not identify the language at either stage and set UnknownLanguage
-			if p.Language == "" {
-				p.Language = pkg.LanguageFromPURL(p.PURL)
-			}
-
-			if cfg.RelationshipsConfig.PackageFileOwnership {
-				// create file-to-package relationships for files owned by the package
-				owningRelationships, err := packageFileOwnershipRelationships(p, resolver)
-				if err != nil {
-					log.Warnf("unable to create any package-file relationships for package name=%q type=%q: %w", p.Name, p.Type, err)
-				} else {
-					relationships = append(relationships, owningRelationships...)
-				}
-			}
-
-			pkgs[i] = p
-		}
+		pkgs, relationships = applyCompliance(cfg.ComplianceConfig, pkgs, relationships)
 
 		sbom.AddPackages(pkgs...)
 		sbom.AddRelationships(relationships...)
@@ -157,6 +127,136 @@ func NewPackageTask(cfg CatalogingFactoryConfig, c pkg.Cataloger, tags ...string
 	return NewTask(c.Name(), fn, tags...)
 }
 
+func finalizePkgCatalogerResults(cfg CatalogingFactoryConfig, resolver file.PathResolver, catalogerName string, pkgs []pkg.Package, relationships []artifact.Relationship) ([]pkg.Package, []artifact.Relationship) {
+	for i, p := range pkgs {
+		if p.FoundBy == "" {
+			p.FoundBy = catalogerName
+		}
+
+		if cfg.DataGenerationConfig.GenerateCPEs && !hasAuthoritativeCPE(p.CPEs) {
+			// generate CPEs (note: this is excluded from package ID, so is safe to mutate)
+			// we might have binary classified CPE already with the package so we want to append here
+			dictionaryCPEs, ok := cpeutils.DictionaryFind(p)
+			if ok {
+				log.Tracef("used CPE dictionary to find CPEs for %s package %q: %s", p.Type, p.Name, dictionaryCPEs)
+				p.CPEs = append(p.CPEs, dictionaryCPEs...)
+			} else {
+				p.CPEs = append(p.CPEs, cpeutils.Generate(p)...)
+			}
+		}
+
+		// if we were not able to identify the language we have an opportunity
+		// to try and get this value from the PURL. Worst case we assert that
+		// we could not identify the language at either stage and set UnknownLanguage
+		if p.Language == "" {
+			p.Language = pkg.LanguageFromPURL(p.PURL)
+		}
+
+		if cfg.RelationshipsConfig.PackageFileOwnership {
+			// create file-to-package relationships for files owned by the package
+			owningRelationships, err := packageFileOwnershipRelationships(p, resolver)
+			if err != nil {
+				log.Warnf("unable to create any package-file relationships for package name=%q type=%q: %v", p.Name, p.Type, err)
+			} else {
+				relationships = append(relationships, owningRelationships...)
+			}
+		}
+
+		pkgs[i] = p
+	}
+	return pkgs, relationships
+}
+
+type packageReplacement struct {
+	original artifact.ID
+	pkg      pkg.Package
+}
+
+func applyCompliance(cfg cataloging.ComplianceConfig, pkgs []pkg.Package, relationships []artifact.Relationship) ([]pkg.Package, []artifact.Relationship) {
+	remainingPkgs, droppedPkgs, replacements := filterNonCompliantPackages(pkgs, cfg)
+
+	relIdx := relationship.NewIndex(relationships...)
+	for _, p := range droppedPkgs {
+		relIdx.Remove(p.ID())
+	}
+
+	for _, replacement := range replacements {
+		relIdx.Replace(replacement.original, replacement.pkg)
+	}
+
+	return remainingPkgs, relIdx.All()
+}
+
+func filterNonCompliantPackages(pkgs []pkg.Package, cfg cataloging.ComplianceConfig) ([]pkg.Package, []pkg.Package, []packageReplacement) {
+	var remainingPkgs, droppedPkgs []pkg.Package
+	var replacements []packageReplacement
+	for _, p := range pkgs {
+		keep, replacement := applyComplianceRules(&p, cfg)
+		if keep {
+			remainingPkgs = append(remainingPkgs, p)
+		} else {
+			droppedPkgs = append(droppedPkgs, p)
+		}
+		if replacement != nil {
+			replacements = append(replacements, *replacement)
+		}
+	}
+
+	return remainingPkgs, droppedPkgs, replacements
+}
+
+func applyComplianceRules(p *pkg.Package, cfg cataloging.ComplianceConfig) (bool, *packageReplacement) {
+	var drop bool
+	var replacement *packageReplacement
+
+	applyComplianceRule := func(value, fieldName string, action cataloging.ComplianceAction) bool {
+		if strings.TrimSpace(value) != "" {
+			return false
+		}
+
+		loc := "unknown"
+		locs := p.Locations.ToSlice()
+		if len(locs) > 0 {
+			loc = locs[0].Path()
+		}
+		switch action {
+		case cataloging.ComplianceActionDrop:
+			log.WithFields("pkg", p.String(), "location", loc).Debugf("package with missing %s, dropping", fieldName)
+			drop = true
+
+		case cataloging.ComplianceActionStub:
+			log.WithFields("pkg", p.String(), "location", loc).Debugf("package with missing %s, stubbing with default value", fieldName)
+			return true
+
+		case cataloging.ComplianceActionKeep:
+			log.WithFields("pkg", p.String(), "location", loc).Tracef("package with missing %s, taking no action", fieldName)
+		}
+		return false
+	}
+
+	ogID := p.ID()
+
+	if applyComplianceRule(p.Name, "name", cfg.MissingName) {
+		p.Name = cataloging.UnknownStubValue
+		p.SetID()
+	}
+
+	if applyComplianceRule(p.Version, "version", cfg.MissingVersion) {
+		p.Version = cataloging.UnknownStubValue
+		p.SetID()
+	}
+
+	newID := p.ID()
+	if newID != ogID {
+		replacement = &packageReplacement{
+			original: ogID,
+			pkg:      *p,
+		}
+	}
+
+	return !drop, replacement
+}
+
 func hasAuthoritativeCPE(cpes []cpe.CPE) bool {
 	for _, c := range cpes {
 		if c.Source != cpe.GeneratedSource {

internal/task/package_task_factory_test.go

@@ -4,8 +4,13 @@ import (
 	"testing"
 
 	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
 
+	"github.com/anchore/syft/syft/artifact"
+	"github.com/anchore/syft/syft/cataloging"
 	"github.com/anchore/syft/syft/cpe"
+	"github.com/anchore/syft/syft/file"
+	"github.com/anchore/syft/syft/pkg"
 )
 
 func Test_hasAuthoritativeCPE(t *testing.T) {
@@ -53,3 +58,83 @@ func Test_hasAuthoritativeCPE(t *testing.T) {
 		})
 	}
 }
+
+func TestApplyCompliance(t *testing.T) {
+	p1 := pkg.Package{Name: "pkg-1", Version: "1.0"}
+	p2 := pkg.Package{Name: "", Version: "1.0"}   // missing name
+	p3 := pkg.Package{Name: "pkg-3", Version: ""} // missing version
+	p4 := pkg.Package{Name: "pkg-4", Version: ""} // missing version
+	c1 := file.Coordinates{RealPath: "/coords/1"}
+	c2 := file.Coordinates{RealPath: "/coords/2"}
+
+	for _, p := range []*pkg.Package{&p1, &p2, &p3, &p4} {
+		p.SetID()
+	}
+
+	r1 := artifact.Relationship{
+		From: p1,
+		To:   c1,
+		Type: artifact.ContainsRelationship,
+	}
+	r2 := artifact.Relationship{
+		From: p2,
+		To:   c2,
+		Type: artifact.ContainsRelationship,
+	}
+
+	cfg := cataloging.ComplianceConfig{
+		MissingName:    cataloging.ComplianceActionDrop,
+		MissingVersion: cataloging.ComplianceActionStub,
+	}
+
+	remainingPkgs, remainingRels := applyCompliance(cfg, []pkg.Package{p1, p2, p3, p4}, []artifact.Relationship{r1, r2})
+
+	// p2 should be dropped because it has a missing name, p3 and p4 should pass with a warning for the missing version
+	assert.Len(t, remainingPkgs, 3) // p1, p3, p4 should remain
+	assert.Len(t, remainingRels, 1) // only r1 should remain (relationship involving p1)
+}
+
+func TestFilterNonCompliantPackages(t *testing.T) {
+	p1 := pkg.Package{Name: "pkg-1", Version: "1.0"}
+	p2 := pkg.Package{Name: "", Version: "1.0"}   // missing name
+	p3 := pkg.Package{Name: "pkg-3", Version: ""} // missing version
+
+	for _, p := range []*pkg.Package{&p1, &p2, &p3} {
+		p.SetID()
+	}
+
+	cfg := cataloging.ComplianceConfig{
+		MissingName:    cataloging.ComplianceActionDrop,
+		MissingVersion: cataloging.ComplianceActionKeep,
+	}
+
+	remainingPkgs, droppedPkgs, replacement := filterNonCompliantPackages([]pkg.Package{p1, p2, p3}, cfg)
+	require.Nil(t, replacement)
+
+	// p2 should be dropped because it has a missing name
+	assert.Len(t, remainingPkgs, 2)
+	assert.Len(t, droppedPkgs, 1)
+	assert.Equal(t, p2, droppedPkgs[0])
+}
+
+func TestApplyComplianceRules_DropAndStub(t *testing.T) {
+	p := pkg.Package{Name: "", Version: ""}
+	p.SetID()
+	ogID := p.ID()
+
+	cfg := cataloging.ComplianceConfig{
+		MissingName:    cataloging.ComplianceActionDrop,
+		MissingVersion: cataloging.ComplianceActionStub,
+	}
+
+	isCompliant, replacement := applyComplianceRules(&p, cfg)
+	require.NotNil(t, replacement)
+	assert.Equal(t, packageReplacement{
+		original: ogID,
+		pkg:      p,
+	}, *replacement)
+
+	// the package should be dropped due to missing name (drop action) and its version should be stubbed
+	assert.False(t, isCompliant)
+	assert.Equal(t, cataloging.UnknownStubValue, p.Version)
+}