trufflehog/pkg/detectors/intra42/intra42_test.go

//go:build detectors
// +build detectors

package intra42

import (
	"context"
	"fmt"
	"testing"
	"time"

	"github.com/google/go-cmp/cmp"
	"github.com/google/go-cmp/cmp/cmpopts"

	"github.com/trufflesecurity/trufflehog/v3/pkg/common"
	"github.com/trufflesecurity/trufflehog/v3/pkg/detectors"
	"github.com/trufflesecurity/trufflehog/v3/pkg/engine/ahocorasick"
	"github.com/trufflesecurity/trufflehog/v3/pkg/pb/detectorspb"
)

func TestIntra42_Pattern(t *testing.T) {
	d := Scanner{}
	ahoCorasickCore := ahocorasick.NewAhoCorasickCore([]detectors.Detector{d})
	tests := []struct {
		name  string
		input string
		want  []string
	}{
		{
			name: "typical pattern",
			input: `
intra_client_id = 'u-s4t2ud-d91c558a2ba6b47f60f690efc20a33d28c252d5bed8400343246f3eb68f490d2'
intra_client_secret = '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'
`,
			want: []string{
				"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",
			},
		},
	}

	for _, test := range tests {
		t.Run(test.name, func(t *testing.T) {
			detectorMatches := ahoCorasickCore.FindDetectorMatches([]byte(test.input))
			if len(detectorMatches) == 0 {
				t.Errorf("keywords '%v' not matched by: %s", d.Keywords(), test.input)
				return
			}

			results, err := d.FromData(context.Background(), false, []byte(test.input))
			if err != nil {
				t.Errorf("error = %v", err)
				return
			}

			if len(results) != len(test.want) {
				if len(results) == 0 {
					t.Errorf("did not receive result")
				} else {
					t.Errorf("expected %d results, only received %d", len(test.want), len(results))
				}
				return
			}

			actual := make(map[string]struct{}, len(results))
			for _, r := range results {
				if len(r.RawV2) > 0 {
					actual[string(r.RawV2)] = struct{}{}
				} else {
					actual[string(r.Raw)] = struct{}{}
				}
			}
			expected := make(map[string]struct{}, len(test.want))
			for _, v := range test.want {
				expected[v] = struct{}{}
			}

			if diff := cmp.Diff(expected, actual); diff != "" {
				t.Errorf("%s diff: (-want +got)\n%s", test.name, diff)
			}
		})
	}
}

func TestIntra42_FromChunk(t *testing.T) {
	ctx, cancel := context.WithTimeout(context.Background(), time.Second*10)
	defer cancel()
	testSecrets, err := common.GetSecret(ctx, "trufflehog-testing", "detectors5")
	if err != nil {
		t.Fatalf("could not get test secrets from GCP: %s", err)
	}
	secret := testSecrets.MustGetField("INTRA42_SECRET")
	id := testSecrets.MustGetField("INTRA42_ID")
	inactiveSecret := testSecrets.MustGetField("INTRA42_INACTIVE")

	type args struct {
		ctx    context.Context
		data   []byte
		verify bool
	}
	tests := []struct {
		name                string
		s                   Scanner
		args                args
		want                []detectors.Result
		wantErr             bool
		wantVerificationErr bool
	}{
		{
			name: "found, verified",
			s:    Scanner{},
			args: args{
				ctx:    context.Background(),
				data:   []byte(fmt.Sprintf("You can find an intra42 secret %s within intra42 %s", secret, id)),
				verify: true,
			},
			want: []detectors.Result{
				{
					DetectorType: detectorspb.DetectorType_Intra42,
					Verified:     true,
				},
			},
			wantErr: false,
		},
		{
			name: "found, would be verified if not for timeout",
			s:    Scanner{client: common.SaneHttpClientTimeOut(1 * time.Microsecond)},
			args: args{
				ctx:    context.Background(),
				data:   []byte(fmt.Sprintf("You can find an intra42 secret %s within intra42 %s", secret, id)),
				verify: true,
			},
			want: []detectors.Result{
				{
					DetectorType: detectorspb.DetectorType_Intra42,
					Verified:     false,
				},
			},
			wantErr:             false,
			wantVerificationErr: true,
		},
		{
			name: "found, verified but unexpected api surface",
			s:    Scanner{client: common.ConstantResponseHttpClient(404, "")},
			args: args{
				ctx:    context.Background(),
				data:   []byte(fmt.Sprintf("You can find an intra42 secret %s within intra42 %s", secret, id)),
				verify: true,
			},
			want: []detectors.Result{
				{
					DetectorType: detectorspb.DetectorType_Intra42,
					Verified:     false,
				},
			},
			wantErr:             false,
			wantVerificationErr: true,
		},
		{
			name: "found, unverified",
			s:    Scanner{},
			args: args{
				ctx:    context.Background(),
				data:   []byte(fmt.Sprintf("You can find an intra42 secret %s within intra42 %s but not valid", inactiveSecret, id)), // the secret would satisfy the regex but not pass validation
				verify: true,
			},
			want: []detectors.Result{
				{
					DetectorType: detectorspb.DetectorType_Intra42,
					Verified:     false,
				},
			},
			wantErr: false,
		},
		{
			name: "not found",
			s:    Scanner{},
			args: args{
				ctx:    context.Background(),
				data:   []byte("You cannot find the secret within"),
				verify: true,
			},
			want:    nil,
			wantErr: false,
		},
	}
	for _, tt := range tests {
		t.Run(tt.name, func(t *testing.T) {
			got, err := tt.s.FromData(tt.args.ctx, tt.args.verify, tt.args.data)
			if (err != nil) != tt.wantErr {
				t.Errorf("Intra42.FromData() error = %v, wantErr %v", err, tt.wantErr)
				return
			}
			for i := range got {
				if len(got[i].Raw) == 0 {
					t.Fatalf("no raw secret present: \n %+v", got[i])
				}
				if (got[i].VerificationError() != nil) != tt.wantVerificationErr {
					t.Errorf("Intra42.FromData() verificationError = %v, wantVerificationErr %v", got[i].VerificationError(), tt.wantVerificationErr)
				}
			}
			ignoreOpts := cmpopts.IgnoreFields(detectors.Result{}, "Raw", "verificationError")
			if diff := cmp.Diff(got, tt.want, ignoreOpts); diff != "" {
				t.Errorf("Intra42.FromData() %s diff: (-got +want)\n%s", tt.name, diff)
			}
		})
	}
}

func BenchmarkFromData(benchmark *testing.B) {
	ctx := context.Background()
	s := Scanner{}
	for name, data := range detectors.MustGetBenchmarkData() {
		benchmark.Run(name, func(b *testing.B) {
			b.ResetTimer()
			for n := 0; n < b.N; n++ {
				_, err := s.FromData(ctx, false, data)
				if err != nil {
					b.Fatal(err)
				}
			}
		})
	}
}