Detector-Competition-Fix: fixed regex for databricks domain and fixed…

… tests (#1965) * fixed regex for domain and fixed tests * fixed regex * fixed an issue with regex subgrouping * made recommended changes * made recommended changed * fixed RawV2
trufflesecurity · Nov 2, 2023 · 965a274 · 965a274
1 parent b6469f2
commit 965a274
Show file tree

Hide file tree

Showing 2 changed files with 80 additions and 34 deletions.
diff --git a/pkg/detectors/databrickstoken/databrickstoken.go b/pkg/detectors/databrickstoken/databrickstoken.go
@@ -12,17 +12,19 @@ import (
 	"github.com/trufflesecurity/trufflehog/v3/pkg/pb/detectorspb"
 )
 
-type Scanner struct{}
+type Scanner struct{
+	client *http.Client
+}
 
 // Ensure the Scanner satisfies the interface at compile time.
 var _ detectors.Detector = (*Scanner)(nil)
 
 var (
-	client = common.SaneHttpClient()
+	defaultClient = common.SaneHttpClient()
 
 	// Make sure that your group is surrounded in boundary characters such as below to reduce false positives.
-	domain = regexp.MustCompile(`\b(https:\/\/[a-z0-9-]+\.cloud\.databricks\.com)\b`)
-	keyPat = regexp.MustCompile(`\b(dapi[a-z0-9]{32})\b`)
+	domain = regexp.MustCompile(`\b([a-z0-9-]+(?:\.[a-z0-9-]+)*\.(cloud\.databricks\.com|gcp\.databricks\.com|azurewebsites\.net))\b`)
+	keyPat    = regexp.MustCompile(`\b(dapi[0-9a-f]{32})(-\d)?\b`)
 )
 
 // Keywords are used for efficiently pre-filtering chunks.
@@ -39,15 +41,9 @@ func (s Scanner) FromData(ctx context.Context, verify bool, data []byte) (result
 	domainMatches := domain.FindAllStringSubmatch(dataStr, -1)
 
 	for _, match := range matches {
-		if len(match) != 2 {
-			continue
-		}
 		resMatch := strings.TrimSpace(match[1])
 
 		for _, domainmatch := range domainMatches {
-			if len(domainmatch) != 2 {
-				continue
-			}
 			resDomainMatch := strings.TrimSpace(domainmatch[1])
 
 			s1 := detectors.Result{
@@ -57,7 +53,11 @@ func (s Scanner) FromData(ctx context.Context, verify bool, data []byte) (result
 			}
 
 			if verify {
-				req, err := http.NewRequestWithContext(ctx, "GET", resDomainMatch + "/api/2.0/clusters/list", nil)
+				client := s.client
+                        	if client == nil {
+                                	client = defaultClient
+                        	}
+				req, err := http.NewRequestWithContext(ctx, "GET", "https://" + resDomainMatch + "/api/2.0/clusters/list", nil)
 				if err != nil {
 					continue
 				}
@@ -67,14 +67,18 @@ func (s Scanner) FromData(ctx context.Context, verify bool, data []byte) (result
 					defer res.Body.Close()
 					if res.StatusCode >= 200 && res.StatusCode < 300 {
 						s1.Verified = true
+					} else if res.StatusCode == 403 {
+						// nothing to do here
 					} else {
-						// This function will check false positives for common test words, but also it will make sure the key appears 'random' enough to be a real key.
-						if detectors.IsKnownFalsePositive(resMatch, detectors.DefaultFalsePositives, true) {
-							continue
-						}
+						s1.VerificationError = fmt.Errorf("unexpected HTTP response status %d", res.StatusCode)
 					}
+				} else {
+					s1.VerificationError = err
 				}
 			}
+			if !s1.Verified && detectors.IsKnownFalsePositive(string(s1.Raw), detectors.DefaultFalsePositives, true) {
+				continue
+			}
 
 			results = append(results, s1)
 		}

diff --git a/pkg/detectors/databrickstoken/databrickstoken_test.go b/pkg/detectors/databrickstoken/databrickstoken_test.go
@@ -6,17 +6,18 @@ package databrickstoken
 import (
 	"context"
 	"fmt"
+	"github.com/google/go-cmp/cmp"
+	"github.com/google/go-cmp/cmp/cmpopts"
 	"testing"
 	"time"
 
-	"github.com/kylelemons/godebug/pretty"
+	"github.com/trufflesecurity/trufflehog/v3/pkg/detectors"
 
 	"github.com/trufflesecurity/trufflehog/v3/pkg/common"
-	"github.com/trufflesecurity/trufflehog/v3/pkg/detectors"
 	"github.com/trufflesecurity/trufflehog/v3/pkg/pb/detectorspb"
 )
 
-func TestDatabrickstoken_FromChunk(t *testing.T) {
+func TestDatabricksToken_FromChunk(t *testing.T) {
 	ctx, cancel := context.WithTimeout(context.Background(), time.Second*5)
 	defer cancel()
 	testSecrets, err := common.GetSecret(ctx, "trufflehog-testing", "detectors2")
@@ -25,25 +26,27 @@ func TestDatabrickstoken_FromChunk(t *testing.T) {
 	}
 	secret := testSecrets.MustGetField("DATABRICKSTOKEN")
 	inactiveSecret := testSecrets.MustGetField("DATABRICKSTOKEN_INACTIVE")
+	domain := testSecrets.MustGetField("DATABRICKSTOKEN_DOMAIN")
 
 	type args struct {
 		ctx    context.Context
 		data   []byte
 		verify bool
 	}
 	tests := []struct {
-		name    string
-		s       Scanner
-		args    args
-		want    []detectors.Result
-		wantErr bool
+		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 a databrickstoken secret %s within", secret)),
+				data:   []byte(fmt.Sprintf("You can find a databrickstoken secret %s within %s", secret, domain)),
 				verify: true,
 			},
 			want: []detectors.Result{
@@ -52,14 +55,15 @@ func TestDatabrickstoken_FromChunk(t *testing.T) {
 					Verified:     true,
 				},
 			},
-			wantErr: false,
+			wantErr:             false,
+			wantVerificationErr: false,
 		},
 		{
 			name: "found, unverified",
 			s:    Scanner{},
 			args: args{
 				ctx:    context.Background(),
-				data:   []byte(fmt.Sprintf("You can find a databrickstoken secret %s within but not valid", inactiveSecret)), // the secret would satisfy the regex but not pass validation
+				data:   []byte(fmt.Sprintf("You can find a databrickstoken secret %s within %s but not valid", inactiveSecret, domain)), // the secret would satisfy the regex but not pass validation
 				verify: true,
 			},
 			want: []detectors.Result{
@@ -68,7 +72,8 @@ func TestDatabrickstoken_FromChunk(t *testing.T) {
 					Verified:     false,
 				},
 			},
-			wantErr: false,
+			wantErr:             false,
+			wantVerificationErr: false,
 		},
 		{
 			name: "not found",
@@ -78,14 +83,48 @@ func TestDatabrickstoken_FromChunk(t *testing.T) {
 				data:   []byte("You cannot find the secret within"),
 				verify: true,
 			},
-			want:    nil,
-			wantErr: false,
+			want:                nil,
+			wantErr:             false,
+			wantVerificationErr: 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 a databrickstoken secret %s within %s", secret, domain)),
+				verify: true,
+			},
+			want: []detectors.Result{
+				{
+					DetectorType: detectorspb.DetectorType_DatabricksToken,
+					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 a databrickstoken secret %s within %s", secret, domain)),
+				verify: true,
+			},
+			want: []detectors.Result{
+				{
+					DetectorType: detectorspb.DetectorType_DatabricksToken,
+					Verified:     false,
+				},
+			},
+			wantErr:             false,
+			wantVerificationErr: true,
 		},
 	}
 	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)
+			got, err := tt.s.FromData(tt.args.ctx, tt.args.verify, tt.args.data)
 			if (err != nil) != tt.wantErr {
 				t.Errorf("Databrickstoken.FromData() error = %v, wantErr %v", err, tt.wantErr)
 				return
@@ -94,10 +133,13 @@ func TestDatabrickstoken_FromChunk(t *testing.T) {
 				if len(got[i].Raw) == 0 {
 					t.Fatalf("no raw secret present: \n %+v", got[i])
 				}
-				got[i].Raw = nil
+				if (got[i].VerificationError != nil) != tt.wantVerificationErr {
+					t.Fatalf("wantVerificationError = %v, verification error = %v", tt.wantVerificationErr, got[i].VerificationError)
+				}
 			}
-			if diff := pretty.Compare(got, tt.want); diff != "" {
-				t.Errorf("Databrickstoken.FromData() %s diff: (-got +want)\n%s", tt.name, diff)
+			ignoreOpts := cmpopts.IgnoreFields(detectors.Result{}, "Raw", "RawV2", "VerificationError")
+			if diff := cmp.Diff(got, tt.want, ignoreOpts); diff != "" {
+				t.Errorf("DatabricksToken.FromData() %s diff: (-got +want)\n%s", tt.name, diff)
 			}
 		})
 	}