trufflehog/pkg/detectors/openai/openai_test.go

//go:build detectors
// +build detectors

package openai

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

	"github.com/google/go-cmp/cmp"
	"github.com/kylelemons/godebug/pretty"

	"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 TestOpenAI_Pattern(t *testing.T) {
	d := Scanner{}
	ahoCorasickCore := ahocorasick.NewAhoCorasickCore([]detectors.Detector{d})
	tests := []struct {
		name  string
		input string
		want  []string
	}{
		{
			name:  "user API key",
			input: "openai.api-key: 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",
			want:  []string{"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"},
		},
		{
			name:  "project API key",
			input: `OPENAI_API_KEY = "sk-proj-mpjtr05CFsJqs4TAeKlCT3BlbkFJsh1KtN0SUjTPeJiagE8K"`,
			want:  []string{"sk-proj-mpjtr05CFsJqs4TAeKlCT3BlbkFJsh1KtN0SUjTPeJiagE8K"},
		},
		{
			name:  "service account API key",
			input: `OPENAI_API_KEY = "sk-service-account-name-Ofbtr05CFsJqs4TAeKlCT3BlbkFJsh1KtN0SUjTPeJiaglyC"`,
			want:  []string{"sk-service-account-name-Ofbtr05CFsJqs4TAeKlCT3BlbkFJsh1KtN0SUjTPeJiaglyC"},
		},
		{
			name:  "newer user API key",
			input: `"OPENAI_API_KEY = "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"`,
			want:  []string{"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"},
		},
		{
			name:  "newer service account API key",
			input: `OPENAI_API_KEY = "sk-svcacct-IUXtc5gIZK-2cBfB-nTgEWbD8mi-fi-gc20oGtq8ve51sET3BlbkFJCg8iQkCVz_nmE_q1dCWlMpemoaoMqHzQ6D-FnWGqlz4C8A"`,
			want:  []string{"sk-svcacct-IUXtc5gIZK-2cBfB-nTgEWbD8mi-fi-gc20oGtq8ve51sET3BlbkFJCg8iQkCVz_nmE_q1dCWlMpemoaoMqHzQ6D-FnWGqlz4C8A"},
		},
	}

	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 TestOpenAI_FromChunk(t *testing.T) {
	ctx, cancel := context.WithTimeout(context.Background(), time.Second*5)
	defer cancel()
	testSecrets, err := common.GetSecret(ctx, "trufflehog-testing", "detectors4")
	if err != nil {
		t.Fatalf("could not get test secrets from GCP: %s", err)
	}

	oaiUnverified := testSecrets.MustGetField("OPENAI_UNVERIFIED")
	oaiVerified := testSecrets.MustGetField("OPENAI_VERIFIED")

	type args struct {
		ctx    context.Context
		data   []byte
		verify bool
	}

	tests := []struct {
		name    string
		s       Scanner
		args    args
		want    []detectors.Result
		wantErr bool
	}{
		{
			name: "Found, unverified OpenAI token sk-",
			s:    Scanner{},
			args: args{
				ctx:    context.Background(),
				data:   []byte(fmt.Sprintf("You can find an OpenAI secret %s within", oaiUnverified)),
				verify: true,
			},
			want: []detectors.Result{
				{
					DetectorType: detectorspb.DetectorType_OpenAI,
					Redacted:     "sk-...gOPc",
					Verified:     false,
				},
			},
			wantErr: false,
		},
		{
			name: "Found, verified OpenAI token sk-",
			s:    Scanner{},
			args: args{
				ctx:    context.Background(),
				data:   []byte(fmt.Sprintf("You can find an OpenAI secret %s within", oaiVerified)),
				verify: true,
			},
			want: []detectors.Result{
				{
					DetectorType: detectorspb.DetectorType_OpenAI,
					Verified:     true,
					Redacted:     "sk-...gOPb",
				},
			},
			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) {
			s := Scanner{}
			got, err := s.FromData(tt.args.ctx, tt.args.verify, tt.args.data)
			if (err != nil) != tt.wantErr {
				t.Errorf("OpenAI.FromData() error = %v, wantErr %v", err, tt.wantErr)
				return
			}
			for i := range got {
				if len(got[i].Raw) == 0 {
					t.Fatal("no raw secret present")
				}
				got[i].Raw = nil
				got[i].ExtraData = nil
				got[i].AnalysisInfo = nil
			}
			if diff := pretty.Compare(got, tt.want); diff != "" {
				t.Errorf("OpenAI.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)
				}
			}
		})
	}
}