Files
keyhunter/cmd/import_test.go
salvacybersec 9dbb0b87d4 feat(07-04): wire keyhunter import command with dedup and DB persist
- Replace import stub with cmd/import.go dispatching to pkg/importer
  (trufflehog, gitleaks, gitleaks-csv) via --format flag
- Reuse openDBWithKey helper so encryption + path resolution match scan/keys
- engineToStorage converts engine.Finding -> storage.Finding (Source -> SourcePath)
- Add pkg/storage.FindingExistsByKey for idempotent cross-import dedup
  keyed on (provider, masked key, source path, line number)
- cmd/import_test.go: selector table, field conversion, end-to-end trufflehog
  import with re-run duplicate assertion, unknown-format + missing-file errors
- pkg/storage queries_test: FindingExistsByKey hit and four miss cases

Delivers IMP-01/02/03 end-to-end.
2026-04-05 23:59:39 +03:00

157 lines
4.5 KiB
Go

package cmd
import (
"bytes"
"os"
"path/filepath"
"testing"
"github.com/salvacybersec/keyhunter/pkg/engine"
"github.com/salvacybersec/keyhunter/pkg/importer"
"github.com/salvacybersec/keyhunter/pkg/storage"
"github.com/spf13/viper"
"github.com/stretchr/testify/assert"
"github.com/stretchr/testify/require"
)
func TestSelectImporter(t *testing.T) {
cases := []struct {
format string
want string
wantErr bool
}{
{"trufflehog", "trufflehog", false},
{"gitleaks", "gitleaks", false},
{"gitleaks-csv", "gitleaks-csv", false},
{"bogus", "", true},
{"", "", true},
}
for _, tc := range cases {
t.Run(tc.format, func(t *testing.T) {
imp, err := selectImporter(tc.format)
if tc.wantErr {
assert.Error(t, err)
return
}
require.NoError(t, err)
assert.Equal(t, tc.want, imp.Name())
// Check concrete type matches expectation via interface dispatch.
switch tc.format {
case "trufflehog":
_, ok := imp.(importer.TruffleHogImporter)
assert.True(t, ok, "expected TruffleHogImporter")
case "gitleaks":
_, ok := imp.(importer.GitleaksImporter)
assert.True(t, ok, "expected GitleaksImporter")
case "gitleaks-csv":
_, ok := imp.(importer.GitleaksCSVImporter)
assert.True(t, ok, "expected GitleaksCSVImporter")
}
})
}
}
func TestEngineToStorage(t *testing.T) {
ef := engine.Finding{
ProviderName: "openai",
KeyValue: "sk-abcdefghijklmnop",
KeyMasked: "sk-abcde...mnop",
Confidence: "high",
Source: "a.yml",
SourceType: "import:trufflehog",
LineNumber: 5,
Verified: true,
VerifyStatus: "live",
VerifyHTTPCode: 200,
VerifyMetadata: map[string]string{"org": "acme"},
}
sf := engineToStorage(ef)
assert.Equal(t, "openai", sf.ProviderName)
assert.Equal(t, "sk-abcdefghijklmnop", sf.KeyValue)
assert.Equal(t, "sk-abcde...mnop", sf.KeyMasked)
assert.Equal(t, "high", sf.Confidence)
assert.Equal(t, "a.yml", sf.SourcePath, "engine.Source -> storage.SourcePath")
assert.Equal(t, "import:trufflehog", sf.SourceType)
assert.Equal(t, 5, sf.LineNumber)
assert.True(t, sf.Verified)
assert.Equal(t, "live", sf.VerifyStatus)
assert.Equal(t, 200, sf.VerifyHTTPCode)
assert.Equal(t, map[string]string{"org": "acme"}, sf.VerifyMetadata)
}
// seedImportDB spins up a temp SQLite database wired through viper +
// KEYHUNTER_PASSPHRASE so runImport -> openDBWithKey resolves to it.
func seedImportDB(t *testing.T) string {
t.Helper()
dir := t.TempDir()
dbPath := filepath.Join(dir, "import.db")
viper.Reset()
viper.Set("database.path", dbPath)
t.Setenv("KEYHUNTER_PASSPHRASE", "test-pass")
t.Cleanup(func() {
viper.Reset()
importFormat = ""
})
return dbPath
}
func TestRunImport_TruffleHogEndToEnd(t *testing.T) {
dbPath := seedImportDB(t)
// Use the canonical testdata shipped by pkg/importer.
importFormat = "trufflehog"
samplePath, err := filepath.Abs(filepath.Join("..", "pkg", "importer", "testdata", "trufflehog-sample.json"))
require.NoError(t, err)
_, err = os.Stat(samplePath)
require.NoError(t, err, "testdata trufflehog-sample.json must exist")
var out bytes.Buffer
importCmd.SetOut(&out)
importCmd.SetErr(&out)
// First import: all findings are new.
err = runImport(importCmd, []string{samplePath})
require.NoError(t, err)
first := out.String()
assert.Contains(t, first, "Imported 3 findings")
assert.Contains(t, first, "3 new")
assert.Contains(t, first, "0 duplicates")
// Confirm findings landed in the database.
db, err := storage.Open(dbPath)
require.NoError(t, err)
encKey, err := loadOrCreateEncKey(db, "test-pass")
require.NoError(t, err)
stored, err := db.ListFindings(encKey)
require.NoError(t, err)
assert.GreaterOrEqual(t, len(stored), 3)
require.NoError(t, db.Close())
// Second import of the same file: everything should now be a duplicate.
out.Reset()
err = runImport(importCmd, []string{samplePath})
require.NoError(t, err)
second := out.String()
assert.Contains(t, second, "Imported 3 findings")
assert.Contains(t, second, "0 new")
assert.Contains(t, second, "3 duplicates")
}
func TestRunImport_UnknownFormat(t *testing.T) {
_ = seedImportDB(t)
importFormat = "bogus"
err := runImport(importCmd, []string{"/nonexistent"})
require.Error(t, err)
assert.Contains(t, err.Error(), "unknown format")
}
func TestRunImport_MissingFile(t *testing.T) {
_ = seedImportDB(t)
importFormat = "trufflehog"
err := runImport(importCmd, []string{filepath.Join(t.TempDir(), "does-not-exist.json")})
require.Error(t, err)
assert.Contains(t, err.Error(), "opening")
}