From 048585dab172c3256147a3f31671ab2d670401cc Mon Sep 17 00:00:00 2001
From: Richard Gomez <rmgomez368@gmail.com>
Date: Fri, 1 Sep 2023 00:12:39 -0400
Subject: [PATCH] feat: optional detectors

---
 pkg/detectors/detectors.go       | 18 ++++++++++++++++++
 pkg/detectors/parseur/parseur.go | 11 +++++++++++
 pkg/engine/engine.go             |  7 +++++++
 pkg/sources/sources.go           |  8 ++++++++
 4 files changed, 44 insertions(+)

diff --git a/pkg/detectors/detectors.go b/pkg/detectors/detectors.go
index ba50c9fbc82d6..0618435e91fc8 100644
--- a/pkg/detectors/detectors.go
+++ b/pkg/detectors/detectors.go
@@ -5,6 +5,7 @@ import (
 	"crypto/rand"
 	"math/big"
 	"net/url"
+	"regexp"
 	"strings"
 	"unicode"
 
@@ -27,6 +28,23 @@ type Detector interface {
 	Type() detectorspb.DetectorType
 }
 
+// ConditionalDetector is an optional interface that a detector can implement to
+// skip chunks based on specific criteria.
+type ConditionalDetector interface {
+	// ScanChunk determines whether the detector should run.
+	ScanChunk(chunk sources.Chunk) bool
+}
+
+// FilenameConditions is a set of common conditions to be used by ConditionalDetector.
+// (Using anonymous structs is weird, but Go has no concept of static members... https://stackoverflow.com/a/55390104)
+var FilenameConditions = struct {
+	// LockFiles are a common source of false-positives.
+	// https://github.com/trufflesecurity/trufflehog/issues/1460
+	LockFiles *regexp.Regexp
+}{
+	LockFiles: regexp.MustCompile(`(^|/)(package(-lock)?\.json|yarn\.lock)$`),
+}
+
 // Versioner is an optional interface that a detector can implement to
 // differentiate instances of the same detector type.
 type Versioner interface {
diff --git a/pkg/detectors/parseur/parseur.go b/pkg/detectors/parseur/parseur.go
index 4cad1da9dfcdd..adfa2023fa10c 100644
--- a/pkg/detectors/parseur/parseur.go
+++ b/pkg/detectors/parseur/parseur.go
@@ -7,6 +7,8 @@ import (
 	"regexp"
 	"strings"
 
+	"github.com/trufflesecurity/trufflehog/v3/pkg/sources"
+
 	"github.com/trufflesecurity/trufflehog/v3/pkg/common"
 	"github.com/trufflesecurity/trufflehog/v3/pkg/detectors"
 	"github.com/trufflesecurity/trufflehog/v3/pkg/pb/detectorspb"
@@ -16,6 +18,7 @@ type Scanner struct{}
 
 // Ensure the Scanner satisfies the interface at compile time
 var _ detectors.Detector = (*Scanner)(nil)
+var _ detectors.ConditionalDetector = (*Scanner)(nil)
 
 var (
 	client = common.SaneHttpClient()
@@ -30,6 +33,14 @@ func (s Scanner) Keywords() []string {
 	return []string{"parseur"}
 }
 
+func (s Scanner) ScanChunk(chunk sources.Chunk) bool {
+	// TODO: Can |chunk.SourceMetadata| be nil?
+	if m, ok := chunk.SourceMetadata.GetData().(sources.GitSourceMetadata); ok {
+		return !detectors.FilenameConditions.LockFiles.MatchString(m.GetFile())
+	}
+	return true
+}
+
 // FromData will find and optionally verify Parseur 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)
diff --git a/pkg/engine/engine.go b/pkg/engine/engine.go
index 38f0da44371a4..e11461b3e238e 100644
--- a/pkg/engine/engine.go
+++ b/pkg/engine/engine.go
@@ -477,6 +477,13 @@ func (e *Engine) detectorWorker(ctx context.Context) {
 				e.ahoCorasickCore.PopulateMatchingDetectors(string(decoded.Chunk.Data), chunkSpecificDetectors)
 
 				for k, detector := range chunkSpecificDetectors {
+					d, ok := detector.(detectors.ConditionalDetector)
+					if ok && !d.ScanChunk(*chunk) {
+						ctx.Logger().V(4).Info("skipping detector for chunk", "detector", detector.Type().String(), "chunk", chunk)
+						delete(chunkSpecificDetectors, k)
+						continue
+					}
+
 					decoded.Chunk.Verify = e.verify
 					wgDetect.Add(1)
 					e.detectableChunksChan <- detectableChunk{
diff --git a/pkg/sources/sources.go b/pkg/sources/sources.go
index c6ab88f7e912a..a929261dd7410 100644
--- a/pkg/sources/sources.go
+++ b/pkg/sources/sources.go
@@ -38,6 +38,14 @@ type Chunk struct {
 	Verify bool
 }
 
+// GitSourceMetadata defines a common interface for Git-based source metadata.
+// For example, this should match Git, Azure, Bitbucket, GitHub, and Gitlab.
+type GitSourceMetadata interface {
+	GetRepository() string
+	GetCommit() string
+	GetFile() string
+}
+
 // ChunkingTarget specifies criteria for a targeted chunking process.
 // Instead of collecting data indiscriminately, this struct allows the caller
 // to specify particular subsets of data they're interested in. This becomes