Add cataloger list command (#2366)

* add cataloger list command

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

* add tests

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

* chore: tidy go mod

Signed-off-by: Christopher Phillips <christopher.phillips@anchore.com>

---------

Signed-off-by: Alex Goodman <wagoodman@users.noreply.github.com>
Signed-off-by: Christopher Phillips <christopher.phillips@anchore.com>
Co-authored-by: Christopher Angelo Phillips <32073428+spiffcs@users.noreply.github.com>
Co-authored-by: Christopher Phillips <christopher.phillips@anchore.com>

cmd/syft/cli/cli.go

@@ -86,6 +86,7 @@ func create(id clio.Identification, out io.Writer) (clio.Application, *cobra.Com
 	rootCmd.AddCommand(
 		scanCmd,
 		commands.Packages(app, scanCmd), // this is currently an alias for the scan command
+		commands.Cataloger(app),
 		commands.Attest(app),
 		commands.Convert(app),
 		clio.VersionCommand(id),

cmd/syft/cli/commands/attest.go

@@ -30,7 +30,7 @@ import (
 )
 
 const (
-	attestExample = `  {{.appName}} {{.command}} --output [FORMAT] alpine:latest defaults to using images from a Docker daemon. If Docker is not present, the image is pulled directly from the registry
+	attestExample = `  {{.appName}} {{.command}} --output [FORMAT] alpine:latest            defaults to using images from a Docker daemon. If Docker is not present, the image is pulled directly from the registry
 `
 	attestSchemeHelp = "\n  " + schemeHelpHeader + "\n" + imageSchemeHelp
 	attestHelp       = attestExample + attestSchemeHelp
@@ -232,7 +232,7 @@ func attestCommand(sbomFilepath string, opts *attestOptions, userInput string) (
 }
 
 func predicateType(outputName string) string {
-	// Select Cosign predicate type based on defined output type
+	// select the Cosign predicate type based on defined output type
 	// As orientation, check: https://github.com/sigstore/cosign/blob/main/pkg/cosign/attestation/attestation.go
 	switch strings.ToLower(outputName) {
 	case "cyclonedx-json":

cmd/syft/cli/commands/cataloger.go

@@ -0,0 +1,20 @@
+package commands
+
+import (
+	"github.com/spf13/cobra"
+
+	"github.com/anchore/clio"
+)
+
+func Cataloger(app clio.Application) *cobra.Command {
+	cmd := &cobra.Command{
+		Use:   "cataloger",
+		Short: "Show available catalogers and configuration",
+	}
+
+	cmd.AddCommand(
+		CatalogerList(app),
+	)
+
+	return cmd
+}

cmd/syft/cli/commands/cataloger_list.go

@@ -0,0 +1,267 @@
+package commands
+
+import (
+	"encoding/json"
+	"fmt"
+	"sort"
+	"strings"
+
+	"github.com/charmbracelet/lipgloss"
+	"github.com/jedib0t/go-pretty/v6/table"
+	"github.com/scylladb/go-set/strset"
+	"github.com/spf13/cobra"
+
+	"github.com/anchore/clio"
+	"github.com/anchore/syft/internal/bus"
+	"github.com/anchore/syft/internal/task"
+	"github.com/anchore/syft/syft/cataloging/pkgcataloging"
+)
+
+type catalogerListOptions struct {
+	Output            string   `yaml:"output" json:"output" mapstructure:"output"`
+	DefaultCatalogers []string `yaml:"default-catalogers" json:"default-catalogers" mapstructure:"default-catalogers"`
+	SelectCatalogers  []string `yaml:"select-catalogers" json:"select-catalogers" mapstructure:"select-catalogers"`
+	ShowHidden        bool     `yaml:"show-hidden" json:"show-hidden" mapstructure:"show-hidden"`
+}
+
+func (o *catalogerListOptions) AddFlags(flags clio.FlagSet) {
+	flags.StringVarP(&o.Output, "output", "o", "format to output the cataloger list (available: table, json)")
+
+	flags.StringArrayVarP(&o.DefaultCatalogers, "override-default-catalogers", "", "override the default catalogers with an expression")
+
+	flags.StringArrayVarP(&o.SelectCatalogers, "select-catalogers", "", "select catalogers with an expression")
+
+	flags.BoolVarP(&o.ShowHidden, "show-hidden", "s", "show catalogers that have been de-selected")
+}
+
+func defaultCatalogerListOptions() *catalogerListOptions {
+	return &catalogerListOptions{
+		DefaultCatalogers: []string{"all"},
+	}
+}
+
+func CatalogerList(app clio.Application) *cobra.Command {
+	opts := defaultCatalogerListOptions()
+
+	return app.SetupCommand(&cobra.Command{
+		Use:   "list [OPTIONS]",
+		Short: "List available catalogers",
+		RunE: func(cmd *cobra.Command, args []string) error {
+			return runCatalogerList(opts)
+		},
+	}, opts)
+}
+
+func runCatalogerList(opts *catalogerListOptions) error {
+	factories := task.DefaultPackageTaskFactories()
+	allTasks, err := factories.Tasks(task.DefaultCatalogingFactoryConfig())
+	if err != nil {
+		return fmt.Errorf("unable to create cataloger tasks: %w", err)
+	}
+
+	report, err := catalogerListReport(opts, allTasks)
+	if err != nil {
+		return fmt.Errorf("unable to generate cataloger list report: %w", err)
+	}
+
+	bus.Report(report)
+
+	return nil
+}
+
+func catalogerListReport(opts *catalogerListOptions, allTasks []task.Task) (string, error) {
+	selectedTasks, selectionEvidence, err := task.Select(allTasks,
+		pkgcataloging.NewSelectionRequest().
+			WithDefaults(opts.DefaultCatalogers...).
+			WithExpression(opts.SelectCatalogers...),
+	)
+	if err != nil {
+		return "", fmt.Errorf("unable to select catalogers: %w", err)
+	}
+	var report string
+
+	switch opts.Output {
+	case "json":
+		report, err = renderCatalogerListJSON(selectedTasks, selectionEvidence, opts.DefaultCatalogers, opts.SelectCatalogers)
+	case "table", "":
+		if opts.ShowHidden {
+			report = renderCatalogerListTable(allTasks, selectionEvidence, opts.DefaultCatalogers, opts.SelectCatalogers)
+		} else {
+			report = renderCatalogerListTable(selectedTasks, selectionEvidence, opts.DefaultCatalogers, opts.SelectCatalogers)
+		}
+	}
+
+	if err != nil {
+		return "", fmt.Errorf("unable to render cataloger list: %w", err)
+	}
+
+	return report, nil
+}
+
+func renderCatalogerListJSON(tasks []task.Task, selection task.Selection, defaultSelections, selections []string) (string, error) {
+	type node struct {
+		Name string   `json:"name"`
+		Tags []string `json:"tags"`
+	}
+
+	names, tagsByName := extractTaskInfo(tasks)
+
+	nodesByName := make(map[string]node)
+
+	for name := range tagsByName {
+		tagsSelected := selection.TokensByTask[name].SelectedOn.List()
+
+		if len(tagsSelected) == 1 && tagsSelected[0] == "all" {
+			tagsSelected = tagsByName[name]
+		}
+
+		sort.Strings(tagsSelected)
+
+		if tagsSelected == nil {
+			// ensure collections are not null
+			tagsSelected = []string{}
+		}
+
+		nodesByName[name] = node{
+			Name: name,
+			Tags: tagsSelected,
+		}
+	}
+
+	type document struct {
+		DefaultSelection []string `json:"default"`
+		Selection        []string `json:"selection"`
+		Catalogers       []node   `json:"catalogers"`
+	}
+
+	if selections == nil {
+		// ensure collections are not null
+		selections = []string{}
+	}
+
+	doc := document{
+		DefaultSelection: defaultSelections,
+		Selection:        selections,
+	}
+
+	for _, name := range names {
+		doc.Catalogers = append(doc.Catalogers, nodesByName[name])
+	}
+
+	by, err := json.Marshal(doc)
+
+	return string(by), err
+}
+
+func renderCatalogerListTable(tasks []task.Task, selection task.Selection, defaultSelections, selections []string) string {
+	t := table.NewWriter()
+	t.SetStyle(table.StyleLight)
+	t.AppendHeader(table.Row{"Cataloger", "Tags"})
+
+	names, tagsByName := extractTaskInfo(tasks)
+
+	rowsByName := make(map[string]table.Row)
+
+	for name, tags := range tagsByName {
+		rowsByName[name] = formatRow(name, tags, selection)
+	}
+
+	for _, name := range names {
+		t.AppendRow(rowsByName[name])
+	}
+
+	report := t.Render()
+
+	if len(selections) > 0 {
+		header := "Selected by expressions:\n"
+		for _, expr := range selections {
+			header += fmt.Sprintf("  - %q\n", expr)
+		}
+		report = header + report
+	}
+
+	if len(defaultSelections) > 0 {
+		header := "Default selections:\n"
+		for _, expr := range defaultSelections {
+			header += fmt.Sprintf("  - %q\n", expr)
+		}
+		report = header + report
+	}
+
+	return report
+}
+
+func formatRow(name string, tags []string, selection task.Selection) table.Row {
+	isIncluded := selection.Result.Has(name)
+	var selections *task.TokenSelection
+	if s, exists := selection.TokensByTask[name]; exists {
+		selections = &s
+	}
+
+	var formattedTags []string
+	for _, tag := range tags {
+		formattedTags = append(formattedTags, formatToken(tag, selections, isIncluded))
+	}
+
+	var tagStr string
+	if isIncluded {
+		tagStr = strings.Join(formattedTags, ", ")
+	} else {
+		tagStr = strings.Join(formattedTags, grey.Render(", "))
+	}
+
+	// TODO: selection should keep warnings (non-selections) in struct
+
+	return table.Row{
+		formatToken(name, selections, isIncluded),
+		tagStr,
+	}
+}
+
+var (
+	green = lipgloss.NewStyle().Foreground(lipgloss.Color("10")) // hi green
+	grey  = lipgloss.NewStyle().Foreground(lipgloss.Color("8"))  // dark grey
+	red   = lipgloss.NewStyle().Foreground(lipgloss.Color("9"))  // high red
+)
+
+func formatToken(token string, selection *task.TokenSelection, included bool) string {
+	if included && selection != nil {
+		// format all tokens in selection in green
+		if selection.SelectedOn.Has(token) {
+			return green.Render(token)
+		}
+
+		return token
+	}
+
+	// format all tokens in selection in red, all others in grey
+	if selection != nil && selection.DeselectedOn.Has(token) {
+		return red.Render(token)
+	}
+
+	return grey.Render(token)
+}
+
+func extractTaskInfo(tasks []task.Task) ([]string, map[string][]string) {
+	tagsByName := make(map[string][]string)
+	var names []string
+
+	for _, tsk := range tasks {
+		var tags []string
+		name := tsk.Name()
+
+		if s, ok := tsk.(task.Selector); ok {
+			set := strset.New(s.Selectors()...)
+			set.Remove(name)
+			tags = set.List()
+			sort.Strings(tags)
+		}
+
+		tagsByName[name] = tags
+		names = append(names, name)
+	}
+
+	sort.Strings(names)
+
+	return names, tagsByName
+}

cmd/syft/cli/commands/cataloger_list_test.go

@@ -0,0 +1,201 @@
+package commands
+
+import (
+	"context"
+	"strings"
+	"testing"
+
+	"github.com/scylladb/go-set/strset"
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+
+	"github.com/anchore/syft/internal/sbomsync"
+	"github.com/anchore/syft/internal/task"
+	"github.com/anchore/syft/syft/file"
+)
+
+var _ interface {
+	task.Task
+	task.Selector
+} = (*dummyTask)(nil)
+
+type dummyTask struct {
+	name      string
+	selectors []string
+}
+
+func (d dummyTask) HasAllSelectors(s ...string) bool {
+	return strset.New(d.selectors...).Has(s...)
+}
+
+func (d dummyTask) Selectors() []string {
+	return d.selectors
+}
+
+func (d dummyTask) Name() string {
+	return d.name
+}
+
+func (d dummyTask) Execute(_ context.Context, _ file.Resolver, _ sbomsync.Builder) error {
+	panic("implement me")
+}
+
+func testTasks() []task.Task {
+	return []task.Task{
+		dummyTask{
+			name:      "task1",
+			selectors: []string{"image", "a", "b", "1"},
+		},
+		dummyTask{
+			name:      "task2",
+			selectors: []string{"image", "b", "c", "2"},
+		},
+		dummyTask{
+			name:      "task3",
+			selectors: []string{"directory", "c", "d", "3"},
+		},
+		dummyTask{
+			name:      "task4",
+			selectors: []string{"directory", "d", "e", "4"},
+		},
+	}
+}
+
+func Test_catalogerListReport(t *testing.T) {
+	tests := []struct {
+		name    string
+		options *catalogerListOptions
+		want    string
+		wantErr require.ErrorAssertionFunc
+	}{
+		{
+			name: "no expressions, table",
+			options: func() *catalogerListOptions {
+				c := defaultCatalogerListOptions()
+				c.Output = "table"
+				return c
+			}(),
+			want: `
+Default selections:
+  - "all"
+┌───────────┬────────────────────┐
+│ CATALOGER │ TAGS               │
+├───────────┼────────────────────┤
+│ task1     │ 1, a, b, image     │
+│ task2     │ 2, b, c, image     │
+│ task3     │ 3, c, d, directory │
+│ task4     │ 4, d, directory, e │
+└───────────┴────────────────────┘
+`,
+		},
+		{
+			name: "no expressions, json",
+			options: func() *catalogerListOptions {
+				c := defaultCatalogerListOptions()
+				c.Output = "json"
+				return c
+			}(),
+			want: `
+{"default":["all"],"selection":[],"catalogers":[{"name":"task1","tags":["1","a","b","image"]},{"name":"task2","tags":["2","b","c","image"]},{"name":"task3","tags":["3","c","d","directory"]},{"name":"task4","tags":["4","d","directory","e"]}]}
+`,
+		},
+		{
+			name: "no expressions, default selection, table",
+			options: func() *catalogerListOptions {
+				c := defaultCatalogerListOptions()
+				c.Output = "table"
+				c.DefaultCatalogers = []string{
+					"image",
+				}
+				return c
+			}(),
+			want: `
+Default selections:
+  - "image"
+┌───────────┬────────────────┐
+│ CATALOGER │ TAGS           │
+├───────────┼────────────────┤
+│ task1     │ 1, a, b, image │
+│ task2     │ 2, b, c, image │
+└───────────┴────────────────┘
+`,
+		},
+		{
+			name: "no expressions, default selection, json",
+			options: func() *catalogerListOptions {
+				c := defaultCatalogerListOptions()
+				c.Output = "json"
+				c.DefaultCatalogers = []string{
+					"image",
+				}
+				return c
+			}(),
+			want: `
+{"default":["image"],"selection":[],"catalogers":[{"name":"task1","tags":["image"]},{"name":"task2","tags":["image"]}]}
+`,
+		},
+		{
+			name: "with expressions, default selection, table",
+			options: func() *catalogerListOptions {
+				c := defaultCatalogerListOptions()
+				c.Output = "table"
+				c.DefaultCatalogers = []string{
+					"image",
+				}
+				c.SelectCatalogers = []string{
+					"-directory",
+					"+task3",
+					"-c",
+					"b",
+				}
+				return c
+			}(),
+			want: `
+Default selections:
+  - "image"
+Selected by expressions:
+  - "-directory"
+  - "+task3"
+  - "-c"
+  - "b"
+┌───────────┬────────────────────┐
+│ CATALOGER │ TAGS               │
+├───────────┼────────────────────┤
+│ task1     │ 1, a, b, image     │
+│ task3     │ 3, c, d, directory │
+└───────────┴────────────────────┘
+`,
+		},
+		{
+			name: "with expressions, default selection, json",
+			options: func() *catalogerListOptions {
+				c := defaultCatalogerListOptions()
+				c.Output = "json"
+				c.DefaultCatalogers = []string{
+					"image",
+				}
+				c.SelectCatalogers = []string{
+					"-directory",
+					"+task3",
+					"-c",
+					"b",
+				}
+				return c
+			}(),
+			want: `
+{"default":["image"],"selection":["-directory","+task3","-c","b"],"catalogers":[{"name":"task1","tags":["b","image"]},{"name":"task3","tags":["task3"]}]}
+`,
+		},
+	}
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			if tt.wantErr == nil {
+				tt.wantErr = require.NoError
+			}
+
+			got, err := catalogerListReport(tt.options, testTasks())
+			tt.wantErr(t, err)
+			assert.Equal(t, strings.TrimSpace(tt.want), strings.TrimSpace(got))
+		})
+	}
+}

cmd/syft/cli/commands/packages.go

@@ -5,6 +5,7 @@ import (
 
 	"github.com/anchore/clio"
 	"github.com/anchore/syft/cmd/syft/internal/ui"
+	"github.com/anchore/syft/internal"
 )
 
 func Packages(app clio.Application, scanCmd *cobra.Command) *cobra.Command {
@@ -13,11 +14,14 @@ func Packages(app clio.Application, scanCmd *cobra.Command) *cobra.Command {
 	opts := defaultScanOptions()
 
 	cmd := app.SetupCommand(&cobra.Command{
-		Use:     "packages [SOURCE]",
+		Use:   "packages [SOURCE]",
-		Short:   scanCmd.Short,
+		Short: scanCmd.Short,
-		Long:    scanCmd.Long,
+		Long:  scanCmd.Long,
-		Args:    scanCmd.Args,
+		Args:  scanCmd.Args,
-		Example: scanCmd.Example,
+		Example: internal.Tprintf(scanHelp, map[string]interface{}{
+			"appName": id.Name,
+			"command": "packages",
+		}),
 		PreRunE: applicationUpdateCheck(id, &opts.UpdateCheck),
 		RunE: func(cmd *cobra.Command, args []string) error {
 			restoreStdout := ui.CaptureStdoutToTraceLog()

go.mod

@@ -17,10 +17,12 @@ require (
 	github.com/anchore/go-version v1.2.2-0.20200701162849-18adb9c92b9b
 	github.com/anchore/packageurl-go v0.1.1-0.20230104203445-02e0a6721501
 	github.com/anchore/stereoscope v0.0.0-20231220161148-590920dabc54
+	github.com/anmitsu/go-shlex v0.0.0-20200514113438-38f4b401e2be
 	// we are hinting brotli to latest due to warning when installing archiver v3:
 	// go: warning: github.com/andybalholm/brotli@v1.0.1: retracted by module author: occasional panics and data corruption
 	github.com/aquasecurity/go-pep440-version v0.0.0-20210121094942-22b2f8951d46
 	github.com/bmatcuk/doublestar/v4 v4.6.1
+	github.com/charmbracelet/bubbles v0.17.1
 	github.com/charmbracelet/bubbletea v0.25.0
 	github.com/charmbracelet/lipgloss v0.9.1
 	github.com/dave/jennifer v1.7.0
@@ -42,6 +44,7 @@ require (
 	github.com/hashicorp/go-multierror v1.1.1
 	github.com/iancoleman/strcase v0.3.0
 	github.com/invopop/jsonschema v0.7.0
+	github.com/jedib0t/go-pretty/v6 v6.5.2
 	github.com/jinzhu/copier v0.4.0
 	github.com/kastenhq/goversion v0.0.0-20230811215019-93b2f8823953
 	github.com/knqyf263/go-rpmdb v0.0.0-20230301153543-ba94b245509b
@@ -76,12 +79,6 @@ require (
 	modernc.org/sqlite v1.28.0
 )
 
-require (
-	github.com/anmitsu/go-shlex v0.0.0-20200514113438-38f4b401e2be
-	github.com/charmbracelet/bubbles v0.17.1
-	github.com/jedib0t/go-pretty/v6 v6.5.3
-)
-
 require (
 	dario.cat/mergo v1.0.0 // indirect
 	github.com/AdaLogics/go-fuzz-headers v0.0.0-20230811130428-ced1acdcaa24 // indirect

go.sum

@@ -466,8 +466,8 @@ github.com/invopop/jsonschema v0.7.0 h1:2vgQcBz1n256N+FpX3Jq7Y17AjYt46Ig3zIWyy77
 github.com/invopop/jsonschema v0.7.0/go.mod h1:O9uiLokuu0+MGFlyiaqtWxwqJm41/+8Nj0lD7A36YH0=
 github.com/jbenet/go-context v0.0.0-20150711004518-d14ea06fba99 h1:BQSFePA1RWJOlocH6Fxy8MmwDt+yVQYULKfN0RoTN8A=
 github.com/jbenet/go-context v0.0.0-20150711004518-d14ea06fba99/go.mod h1:1lJo3i6rXxKeerYnT8Nvf0QmHCRC1n8sfWVwXF2Frvo=
-github.com/jedib0t/go-pretty/v6 v6.5.3 h1:GIXn6Er/anHTkVUoufs7ptEvxdD6KIhR7Axa2wYCPF0=
+github.com/jedib0t/go-pretty/v6 v6.5.2 h1:1zphkAo77tdoCkdqIYsMHoXmEGTnTy3GZ6Mn+NyIro0=
-github.com/jedib0t/go-pretty/v6 v6.5.3/go.mod h1:5LQIxa52oJ/DlDSLv0HEkWOFMDGoWkJb9ss5KqPpJBg=
+github.com/jedib0t/go-pretty/v6 v6.5.2/go.mod h1:5LQIxa52oJ/DlDSLv0HEkWOFMDGoWkJb9ss5KqPpJBg=
 github.com/jinzhu/copier v0.4.0 h1:w3ciUoD19shMCRargcpm0cm91ytaBhDvuRpz1ODO/U8=
 github.com/jinzhu/copier v0.4.0/go.mod h1:DfbEm0FYsaqBcKcFuvmOZb218JkPGtvSHsKg8S8hyyg=
 github.com/json-iterator/go v1.1.6/go.mod h1:+SdeFBvtyEkXs7REEP0seUULqWtbJapLOCVDaaPEHmU=