diff --git a/pkg/detectors/accuweather/accuweather.go b/pkg/detectors/accuweather/v1/accuweather.go similarity index 81% rename from pkg/detectors/accuweather/accuweather.go rename to pkg/detectors/accuweather/v1/accuweather.go index c7a5d75e7..525a6d2a4 100644 --- a/pkg/detectors/accuweather/accuweather.go +++ b/pkg/detectors/accuweather/v1/accuweather.go @@ -13,7 +13,7 @@ import ( ) type Scanner struct { - client *http.Client + Client *http.Client } const accuweatherURL = "https://dataservice.accuweather.com" @@ -35,27 +35,24 @@ func (s Scanner) Keywords() []string { return []string{"accuweather"} } +func (s Scanner) Version() int { return 1 } + func (s Scanner) getClient() *http.Client { - if s.client != nil { - return s.client + if s.Client != nil { + return s.Client } return defaultClient } // FromData will find and optionally verify Accuweather secrets in a given set of bytes. func (s Scanner) FromData(ctx context.Context, verify bool, data []byte) (results []detectors.Result, err error) { - dataStr := string(data) + matches := keyPat.FindAllStringSubmatch(string(data), -1) + return s.ProcessMatches(ctx, matches, verify) +} - matches := make(map[string]struct{}) - for _, match := range keyPat.FindAllStringSubmatch(dataStr, -1) { - k := match[1] - if detectors.StringShannonEntropy(k) < requiredShannonEntropy { - continue - } - matches[k] = struct{}{} - } - - for key := range matches { +func (s Scanner) ProcessMatches(ctx context.Context, matches [][]string, verify bool) (results []detectors.Result, err error) { + uniqueMatches := getUniqueMatches(matches) + for key := range uniqueMatches { s1 := detectors.Result{ DetectorType: detectorspb.DetectorType_Accuweather, Raw: []byte(key), @@ -71,7 +68,19 @@ func (s Scanner) FromData(ctx context.Context, verify bool, data []byte) (result results = append(results, s1) } - return results, nil + return +} + +func getUniqueMatches(allMatches [][]string) map[string]struct{} { + uniqueMatches := map[string]struct{}{} + for _, match := range allMatches { + k := match[1] + if detectors.StringShannonEntropy(k) < requiredShannonEntropy { + continue + } + uniqueMatches[k] = struct{}{} + } + return uniqueMatches } func verifyAccuweather(ctx context.Context, client *http.Client, key string) (bool, error) { diff --git a/pkg/detectors/accuweather/accuweather_integration_test.go b/pkg/detectors/accuweather/v1/accuweather_integration_test.go similarity index 97% rename from pkg/detectors/accuweather/accuweather_integration_test.go rename to pkg/detectors/accuweather/v1/accuweather_integration_test.go index 8dc6f8421..68075082d 100644 --- a/pkg/detectors/accuweather/accuweather_integration_test.go +++ b/pkg/detectors/accuweather/v1/accuweather_integration_test.go @@ -57,7 +57,7 @@ func TestAccuweather_FromChunk(t *testing.T) { }, { name: "found, real secrets, verification error due to timeout", - s: Scanner{client: common.SaneHttpClientTimeOut(1 * time.Microsecond)}, + s: Scanner{Client: common.SaneHttpClientTimeOut(1 * time.Microsecond)}, args: args{ ctx: context.Background(), data: []byte(fmt.Sprintf("You can find a accuweather secret %s within", secret)), @@ -75,7 +75,7 @@ func TestAccuweather_FromChunk(t *testing.T) { }, { name: "found, real secrets, verification error due to unexpected api surface", - s: Scanner{client: common.ConstantResponseHttpClient(500, "{}")}, + s: Scanner{Client: common.ConstantResponseHttpClient(500, "{}")}, args: args{ ctx: context.Background(), data: []byte(fmt.Sprintf("You can find a accuweather secret %s within", secret)), diff --git a/pkg/detectors/accuweather/accuweather_test.go b/pkg/detectors/accuweather/v1/accuweather_test.go similarity index 100% rename from pkg/detectors/accuweather/accuweather_test.go rename to pkg/detectors/accuweather/v1/accuweather_test.go diff --git a/pkg/detectors/accuweather/v2/accuweather.go b/pkg/detectors/accuweather/v2/accuweather.go new file mode 100644 index 000000000..f99875680 --- /dev/null +++ b/pkg/detectors/accuweather/v2/accuweather.go @@ -0,0 +1,30 @@ +package accuweather + +import ( + "context" + + regexp "github.com/wasilibs/go-re2" + + "github.com/trufflesecurity/trufflehog/v3/pkg/detectors" + v1 "github.com/trufflesecurity/trufflehog/v3/pkg/detectors/accuweather/v1" +) + +type Scanner struct { + v1.Scanner +} + +func (s Scanner) Version() int { return 2 } + +var ( + // Ensure the Scanner satisfies the interface at compile time. + _ detectors.Detector = (*Scanner)(nil) + + // Make sure that your group is surrounded in boundary characters such as below to reduce false positives. + keyPat = regexp.MustCompile(detectors.PrefixRegex([]string{"accuweather"}) + `\b([a-zA-Z0-9]{32})\b`) +) + +// FromData will find and optionally verify Accuweather secrets in a given set of bytes. +func (s Scanner) FromData(ctx context.Context, verify bool, data []byte) (results []detectors.Result, err error) { + matches := keyPat.FindAllStringSubmatch(string(data), -1) + return s.ProcessMatches(ctx, matches, verify) +} diff --git a/pkg/detectors/accuweather/v2/accuweather_integration_test.go b/pkg/detectors/accuweather/v2/accuweather_integration_test.go new file mode 100644 index 000000000..2c377fd52 --- /dev/null +++ b/pkg/detectors/accuweather/v2/accuweather_integration_test.go @@ -0,0 +1,169 @@ +//go:build detectors +// +build detectors + +package accuweather + +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" + v1 "github.com/trufflesecurity/trufflehog/v3/pkg/detectors/accuweather/v1" + "github.com/trufflesecurity/trufflehog/v3/pkg/pb/detectorspb" +) + +func TestAccuweather_FromChunk(t *testing.T) { + ctx, cancel := context.WithTimeout(context.Background(), time.Second*5) + 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("ACCUWEATHER") + inactiveSecret := testSecrets.MustGetField("ACCUWEATHER_INACTIVE") + + 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, verified", + s: Scanner{}, + args: args{ + ctx: context.Background(), + data: []byte(fmt.Sprintf("You can find a accuweather secret %s within but verified", secret)), + verify: true, + }, + want: []detectors.Result{ + { + DetectorType: detectorspb.DetectorType_Accuweather, + Verified: true, + }, + }, + wantErr: false, + }, + { + name: "found, real secrets, verification error due to timeout", + s: Scanner{Scanner: v1.Scanner{Client: common.SaneHttpClientTimeOut(1 * time.Microsecond)}}, + args: args{ + ctx: context.Background(), + data: []byte(fmt.Sprintf("You can find a accuweather secret %s within", secret)), + verify: true, + }, + want: func() []detectors.Result { + r := detectors.Result{ + DetectorType: detectorspb.DetectorType_Accuweather, + Verified: false, + } + r.SetVerificationError(context.DeadlineExceeded) + return []detectors.Result{r} + }(), + wantErr: false, + }, + { + name: "found, real secrets, verification error due to unexpected api surface", + s: Scanner{Scanner: v1.Scanner{Client: common.ConstantResponseHttpClient(500, "{}")}}, + args: args{ + ctx: context.Background(), + data: []byte(fmt.Sprintf("You can find a accuweather secret %s within", secret)), + verify: true, + }, + want: func() []detectors.Result { + r := detectors.Result{ + DetectorType: detectorspb.DetectorType_Accuweather, + Verified: false, + } + r.SetVerificationError(fmt.Errorf("unexpected HTTP response status 500")) + return []detectors.Result{r} + }(), + wantErr: false, + }, + { + name: "found, unverified", + s: Scanner{}, + args: args{ + ctx: context.Background(), + data: []byte(fmt.Sprintf("You can find a accuweather secret %s within but verified", inactiveSecret)), + verify: true, + }, + want: []detectors.Result{ + { + DetectorType: detectorspb.DetectorType_Accuweather, + 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("Accuweather.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]) + } + gotErr := "" + if got[i].VerificationError() != nil { + gotErr = got[i].VerificationError().Error() + } + wantErr := "" + if tt.want[i].VerificationError() != nil { + wantErr = tt.want[i].VerificationError().Error() + } + if gotErr != wantErr { + t.Fatalf("wantVerificationError = %v, verification error = %v", tt.want[i].VerificationError(), got[i].VerificationError()) + } + got[i].Raw = nil + } + ignoreOpts := cmpopts.IgnoreFields(detectors.Result{}, "Raw", "RawV2", "verificationError") + if diff := cmp.Diff(got, tt.want, ignoreOpts); diff != "" { + t.Errorf("Accuweather.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) + } + } + }) + } +} diff --git a/pkg/detectors/accuweather/v2/accuweather_test.go b/pkg/detectors/accuweather/v2/accuweather_test.go new file mode 100644 index 000000000..b4c309f54 --- /dev/null +++ b/pkg/detectors/accuweather/v2/accuweather_test.go @@ -0,0 +1,86 @@ +package accuweather + +import ( + "context" + "fmt" + "testing" + + "github.com/google/go-cmp/cmp" + + "github.com/trufflesecurity/trufflehog/v3/pkg/detectors" + "github.com/trufflesecurity/trufflehog/v3/pkg/engine/ahocorasick" +) + +var ( + validPattern = "dqftwc490oPcxae67sBSF741M56d091a" + invalidPattern = "dqftwc49%oPcxae67sBSF741M56f091a" +) + +func TestAccuWeather_Pattern(t *testing.T) { + d := Scanner{} + ahoCorasickCore := ahocorasick.NewAhoCorasickCore([]detectors.Detector{d}) + + tests := []struct { + name string + input string + want []string + }{ + { + name: "valid pattern", + input: fmt.Sprintf("accuweather token = '%s'", validPattern), + want: []string{validPattern}, + }, + { + name: "valid pattern - out of prefix range", + input: fmt.Sprintf("accuweather token keyword is not close to the real token = '%s'", validPattern), + want: nil, + }, + { + name: "invalid pattern", + input: fmt.Sprintf("accuweather = '%s'", invalidPattern), + want: nil, + }, + } + + for _, test := range tests { + t.Run(test.name, func(t *testing.T) { + matchedDetectors := ahoCorasickCore.FindDetectorMatches([]byte(test.input)) + if len(matchedDetectors) == 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) + } + }) + } +} diff --git a/pkg/engine/defaults/defaults.go b/pkg/engine/defaults/defaults.go index 8dc34ddb0..581cc1e22 100644 --- a/pkg/engine/defaults/defaults.go +++ b/pkg/engine/defaults/defaults.go @@ -4,7 +4,8 @@ import ( "github.com/trufflesecurity/trufflehog/v3/pkg/detectors" "github.com/trufflesecurity/trufflehog/v3/pkg/detectors/abuseipdb" "github.com/trufflesecurity/trufflehog/v3/pkg/detectors/abyssale" - "github.com/trufflesecurity/trufflehog/v3/pkg/detectors/accuweather" + accuweatherv1 "github.com/trufflesecurity/trufflehog/v3/pkg/detectors/accuweather/v1" + accuweatherv2 "github.com/trufflesecurity/trufflehog/v3/pkg/detectors/accuweather/v2" "github.com/trufflesecurity/trufflehog/v3/pkg/detectors/adafruitio" "github.com/trufflesecurity/trufflehog/v3/pkg/detectors/adzuna" "github.com/trufflesecurity/trufflehog/v3/pkg/detectors/aeroworkflow" @@ -843,7 +844,8 @@ func buildDetectorList() []detectors.Detector { &abyssale.Scanner{}, // &abstract.Scanner{}, &abuseipdb.Scanner{}, - &accuweather.Scanner{}, + &accuweatherv1.Scanner{}, + &accuweatherv2.Scanner{}, &adafruitio.Scanner{}, // &adobeio.Scanner{}, &adzuna.Scanner{},