Go: fix SSRF (#17641)

### Summary

Check the URL to prevent SSRF attack.

---------

Signed-off-by: Jin Hai <haijin.chn@gmail.com>

internal/parser/parser/markdown_parser_test.go

@@ -203,6 +203,7 @@ func TestResolveMarkdownImage_NoImage(t *testing.T) {
 }
 
 func TestResolveMarkdownImage_HTTPImage(t *testing.T) {
+	withSSRFBypass(t)
 	// httptest servers bind loopback, which the SSRF guard rejects by
 	// default. Allow loopback for this test so the HTTP fetch path is
 	// exercised (production keeps ssrfAllowLoopback == false).
@@ -236,6 +237,7 @@ func TestFetchImageAsBase64_RejectsCredentials(t *testing.T) {
 }
 
 func TestFetchImageAsBase64_InvalidURL(t *testing.T) {
+	withSSRFBypass(t)
 	ts := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, _ *http.Request) {
 		w.WriteHeader(http.StatusNotFound)
 	}))

internal/parser/parser/pdf_parser_docling.go

@@ -88,7 +88,7 @@ func parsePDFWithDocling(ctx context.Context, filename string, data []byte, pars
 	var lastErr error
 	for _, candidate := range payloads {
 		url := baseURL + candidate.endpoint
-		resp, err := models.PostJSONRequest(context.Background(), models.NewDriverHTTPClient(), url, auth, candidate.body())
+		resp, err := models.PostJSONRequest(context.Background(), models.NewDriverHTTPClient(false), url, auth, candidate.body())
 		if err != nil {
 			lastErr = fmt.Errorf("%s: %w", candidate.endpoint, err)
 			continue

internal/parser/parser/pdf_parser_docling_test.go

@@ -11,6 +11,7 @@ import (
 )
 
 func TestPDFParser_ParseWithResult_DoclingChunkedMarkdownIntegration(t *testing.T) {
+	withSSRFBypass(t)
 	var requestCount atomic.Int32
 
 	server := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
@@ -84,6 +85,7 @@ func TestPDFParser_ParseWithResult_DoclingChunkedMarkdownIntegration(t *testing.
 }
 
 func TestPDFParser_ParseWithResult_DoclingFallbackToStandardJSONIntegration(t *testing.T) {
+	withSSRFBypass(t)
 	var requestCount atomic.Int32
 
 	server := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {

internal/parser/parser/pdf_parser_mineru_test.go

@@ -11,6 +11,7 @@ import (
 )
 
 func TestPDFParser_ParseWithResult_MinerUMarkdownIntegration(t *testing.T) {
+	withSSRFBypass(t)
 	var submitCalled atomic.Bool
 	var resultCalled atomic.Bool
 
@@ -100,6 +101,7 @@ func TestPDFParser_ParseWithResult_MinerUMarkdownIntegration(t *testing.T) {
 }
 
 func TestPDFParser_ParseWithResult_MinerUJSONIntegration(t *testing.T) {
+	withSSRFBypass(t)
 	server := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
 		switch {
 		case r.Method == http.MethodPost && r.URL.Path == "/file_parse":

internal/parser/parser/pdf_parser_opendataloader.go

@@ -41,7 +41,7 @@ func parsePDFWithOpenDataLoader(ctx context.Context, filename string, data []byt
 	if apiKey != "" {
 		req.Header.Set("Authorization", "Bearer "+apiKey)
 	}
-	resp, err := models.NewDriverHTTPClient().Do(req)
+	resp, err := models.NewDriverHTTPClient(false).Do(req)
 	if err != nil {
 		return ParseResult{Err: fmt.Errorf("parser: OpenDataLoader submit: %w", err)}
 	}

internal/parser/parser/pdf_parser_opendataloader_test.go

@@ -10,6 +10,7 @@ import (
 )
 
 func TestPDFParser_ParseWithResult_OpenDataLoaderJSONIntegration(t *testing.T) {
+	withSSRFBypass(t)
 	server := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
 		if r.Method != http.MethodPost || r.URL.Path != "/file_parse" {
 			http.NotFound(w, r)
@@ -84,6 +85,7 @@ func TestPDFParser_ParseWithResult_OpenDataLoaderJSONIntegration(t *testing.T) {
 }
 
 func TestPDFParser_ParseWithResult_OpenDataLoaderMarkdownFallback(t *testing.T) {
+	withSSRFBypass(t)
 	server := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
 		if r.URL.Path != "/file_parse" {
 			http.NotFound(w, r)

internal/parser/parser/pdf_parser_paddleocr_test.go

@@ -11,6 +11,7 @@ import (
 )
 
 func TestPDFParser_ParseWithResult_PaddleOCRMarkdownIntegration(t *testing.T) {
+	withSSRFBypass(t)
 	var called atomic.Bool
 
 	server := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
@@ -82,6 +83,7 @@ func TestPDFParser_ParseWithResult_PaddleOCRMarkdownIntegration(t *testing.T) {
 }
 
 func TestPDFParser_ParseWithResult_PaddleOCRJSONIntegration(t *testing.T) {
+	withSSRFBypass(t)
 	server := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
 		if r.Method != http.MethodPost || r.URL.Path != "/layout-parsing" {
 			http.NotFound(w, r)

internal/parser/parser/pdf_parser_somark.go

@@ -84,7 +84,7 @@ func soMarkSubmit(baseURL, filename string, data []byte, parser *PDFParser, apiK
 		return "", fmt.Errorf("parser: SoMark request: %w", err)
 	}
 	req.Header.Set("Content-Type", writer.FormDataContentType())
-	resp, err := models.NewDriverHTTPClient().Do(req)
+	resp, err := models.NewDriverHTTPClient(false).Do(req)
 	if err != nil {
 		return "", fmt.Errorf("parser: SoMark submit: %w", err)
 	}
@@ -122,7 +122,7 @@ func soMarkPoll(baseURL, taskID, apiKey string) (map[string]any, error) {
 		return nil, fmt.Errorf("parser: SoMark poll request: %w", err)
 	}
 	req.Header.Set("Content-Type", "application/x-www-form-urlencoded")
-	resp, err := models.NewDriverHTTPClient().Do(req)
+	resp, err := models.NewDriverHTTPClient(false).Do(req)
 	if err != nil {
 		return nil, fmt.Errorf("parser: SoMark poll: %w", err)
 	}

internal/parser/parser/pdf_parser_somark_test.go

@@ -10,6 +10,7 @@ import (
 )
 
 func TestPDFParser_ParseWithResult_SoMarkJSONIntegration(t *testing.T) {
+	withSSRFBypass(t)
 	var submitSeen bool
 	var pollSeen bool
 	server := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
@@ -78,6 +79,7 @@ func TestPDFParser_ParseWithResult_SoMarkJSONIntegration(t *testing.T) {
 }
 
 func TestPDFParser_ParseWithResult_SoMarkMarkdownIntegration(t *testing.T) {
+	withSSRFBypass(t)
 	server := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
 		switch r.URL.Path {
 		case "/parse/async":
@@ -123,6 +125,7 @@ func TestSoMarkBlockToItem_DropsHeaderByDefault(t *testing.T) {
 }
 
 func TestSoMarkSubmitMultipartShape(t *testing.T) {
+	withSSRFBypass(t)
 	var form multipart.Form
 	server := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
 		_ = r.ParseMultipartForm(1 << 20)

internal/parser/parser/pdf_parser_tcadp.go

@@ -40,7 +40,7 @@ func parsePDFWithTCADP(filename string, data []byte, parser *PDFParser) ParseRes
 			"MarkdownImageResponseType": parser.TCADPMarkdownImageResponseType,
 		},
 	}
-	resp, err := models.PostJSONRequest(context.Background(), models.NewDriverHTTPClient(), strings.TrimRight(baseURL, "/")+"/reconstruct_document", bearer(apiKey), requestBody)
+	resp, err := models.PostJSONRequest(context.Background(), models.NewDriverHTTPClient(false), strings.TrimRight(baseURL, "/")+"/reconstruct_document", bearer(apiKey), requestBody)
 	if err != nil {
 		return ParseResult{Err: fmt.Errorf("parser: TCADP submit: %w", err)}
 	}
@@ -68,7 +68,7 @@ func parsePDFWithTCADP(filename string, data []byte, parser *PDFParser) ParseRes
 	if auth := bearer(apiKey); auth != "" {
 		downloadReq.Header.Set("Authorization", auth)
 	}
-	downloadResp, err := models.NewDriverHTTPClient().Do(downloadReq)
+	downloadResp, err := models.NewDriverHTTPClient(false).Do(downloadReq)
 	if err != nil {
 		return ParseResult{Err: fmt.Errorf("parser: TCADP download: %w", err)}
 	}

internal/parser/parser/pdf_parser_tcadp_test.go

@@ -10,6 +10,7 @@ import (
 )
 
 func TestPDFParser_ParseWithResult_TCADPJSONIntegration(t *testing.T) {
+	withSSRFBypass(t)
 	zipPayload := tcadpZipFixture(t)
 	var server *httptest.Server
 	server = httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
@@ -47,6 +48,7 @@ func TestPDFParser_ParseWithResult_TCADPJSONIntegration(t *testing.T) {
 }
 
 func TestPDFParser_ParseWithResult_TCADPMarkdownIntegration(t *testing.T) {
+	withSSRFBypass(t)
 	zipPayload := tcadpZipFixture(t)
 	var server *httptest.Server
 	server = httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {

internal/parser/parser/ppt_parser_cgo_test.go

@@ -135,6 +135,7 @@ type tcadpPresentationParser interface {
 }
 
 func testPresentationTCADPIntegration(t *testing.T, p tcadpPresentationParser, wantFileType, filename string) {
+	withSSRFBypass(t)
 	t.Helper()
 	zipPayload := tcadpZipFixture(t)
 	var gotFileType string
@@ -189,6 +190,7 @@ func testPresentationTCADPIntegration(t *testing.T, p tcadpPresentationParser, w
 // from the TCADP download endpoint is surfaced as an explicit error rather
 // than parsed as a (malformed) ZIP artifact.
 func TestPPTXParser_TCADPDownloadHTTPError(t *testing.T) {
+	withSSRFBypass(t)
 	var server *httptest.Server
 	server = httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
 		switch r.URL.Path {

internal/parser/parser/pptx_tcadp.go

@@ -45,7 +45,7 @@ func parsePresentationWithTCADP(ctx context.Context, filename string, data []byt
 			"MarkdownImageResponseType": markdownImageResponseType,
 		},
 	}
-	resp, err := models.PostJSONRequest(ctx, models.NewDriverHTTPClient(),
+	resp, err := models.PostJSONRequest(ctx, models.NewDriverHTTPClient(false),
 		strings.TrimRight(baseURL, "/")+"/reconstruct_document", bearer(apiKey), requestBody)
 	if err != nil {
 		return ParseResult{Err: fmt.Errorf("parser: TCADP submit: %w", err)}
@@ -75,7 +75,7 @@ func parsePresentationWithTCADP(ctx context.Context, filename string, data []byt
 	if auth := bearer(apiKey); auth != "" {
 		downloadReq.Header.Set("Authorization", auth)
 	}
-	downloadResp, err := models.NewDriverHTTPClient().Do(downloadReq)
+	downloadResp, err := models.NewDriverHTTPClient(false).Do(downloadReq)
 	if err != nil {
 		return ParseResult{Err: fmt.Errorf("parser: TCADP download: %w", err)}
 	}

internal/parser/parser/ssrf_test_helper_test.go

@@ -0,0 +1,19 @@
+package parser
+
+import (
+	"testing"
+
+	"ragflow/internal/utility"
+)
+
+// withSSRFBypass enables the test-only SSRF bypass for the duration of a test
+// and restores the previous value in t.Cleanup. Parser tests exercise parsing
+// logic against httptest servers bound to 127.0.0.1, which the strict SSRF
+// guard rejects. SSRF enforcement has its own dedicated coverage in
+// internal/utility/ssrf_test.go.
+func withSSRFBypass(t *testing.T) {
+	t.Helper()
+	prev := utility.AllowAnyHostForTest
+	utility.AllowAnyHostForTest = true
+	t.Cleanup(func() { utility.AllowAnyHostForTest = prev })
+}

internal/parser/parser/xls_tcadp.go

@@ -38,7 +38,7 @@ func parseSpreadsheetWithTCADP(filename string, data []byte, fileType string, tc
 			"MarkdownImageResponseType": markdownImageResponseType,
 		},
 	}
-	resp, err := models.PostJSONRequest(context.Background(), models.NewDriverHTTPClient(), strings.TrimRight(baseURL, "/")+"/reconstruct_document", bearer(apiKey), requestBody)
+	resp, err := models.PostJSONRequest(context.Background(), models.NewDriverHTTPClient(false), strings.TrimRight(baseURL, "/")+"/reconstruct_document", bearer(apiKey), requestBody)
 	if err != nil {
 		return ParseResult{Err: fmt.Errorf("parser: TCADP submit: %w", err)}
 	}
@@ -66,7 +66,7 @@ func parseSpreadsheetWithTCADP(filename string, data []byte, fileType string, tc
 	if auth := bearer(apiKey); auth != "" {
 		downloadReq.Header.Set("Authorization", auth)
 	}
-	downloadResp, err := models.NewDriverHTTPClient().Do(downloadReq)
+	downloadResp, err := models.NewDriverHTTPClient(false).Do(downloadReq)
 	if err != nil {
 		return ParseResult{Err: fmt.Errorf("parser: TCADP download: %w", err)}
 	}

internal/parser/parser/xls_tcadp_test.go

@@ -7,6 +7,7 @@ import (
 )
 
 func TestXLSXParser_ParseWithResult_TCADPJSONIntegration(t *testing.T) {
+	withSSRFBypass(t)
 	ctx := t.Context()
 	zipPayload := tcadpZipFixture(t)
 	var server *httptest.Server
@@ -45,6 +46,7 @@ func TestXLSXParser_ParseWithResult_TCADPJSONIntegration(t *testing.T) {
 }
 
 func TestXLSParser_ParseWithResult_TCADPJSONIntegration(t *testing.T) {
+	withSSRFBypass(t)
 	ctx := t.Context()
 	zipPayload := tcadpZipFixture(t)
 	var server *httptest.Server
@@ -80,6 +82,7 @@ func TestXLSParser_ParseWithResult_TCADPJSONIntegration(t *testing.T) {
 }
 
 func TestCSVParser_ParseWithResult_TCADPJSONIntegration(t *testing.T) {
+	withSSRFBypass(t)
 	ctx := t.Context()
 	zipPayload := tcadpZipFixture(t)
 	var server *httptest.Server
@@ -112,6 +115,7 @@ func TestCSVParser_ParseWithResult_TCADPJSONIntegration(t *testing.T) {
 }
 
 func TestXLSXParser_ParseWithResult_TCADPMarkdownIntegration(t *testing.T) {
+	withSSRFBypass(t)
 	ctx := t.Context()
 	zipPayload := tcadpZipFixture(t)
 	var server *httptest.Server