Merge pull request Wei-Shaw#1117 from alfadb/fix/empty-text-block-retry

fix: 修复空 text block 导致上游 400 错误未被重试捕获的问题

Author: Wei-Shaw

backend/internal/service/gateway_request.go

@@ -28,6 +28,12 @@ var (
 	patternEmptyContentSpaced = []byte(`"content": []`)
 	patternEmptyContentSp1    = []byte(`"content" : []`)
 	patternEmptyContentSp2    = []byte(`"content" :[]`)
+
+	// Fast-path patterns for empty text blocks: {"type":"text","text":""}
+	patternEmptyText       = []byte(`"text":""`)
+	patternEmptyTextSpaced = []byte(`"text": ""`)
+	patternEmptyTextSp1    = []byte(`"text" : ""`)
+	patternEmptyTextSp2    = []byte(`"text" :""`)
 )
 
 // SessionContext 粘性会话上下文，用于区分不同来源的请求。
@@ -233,15 +239,22 @@ func FilterThinkingBlocksForRetry(body []byte) []byte {
 		bytes.Contains(body, patternThinkingField) ||
 		bytes.Contains(body, patternThinkingFieldSpaced)
 
-	// Also check for empty content arrays that need fixing.
+	// Also check for empty content arrays and empty text blocks that need fixing.
 	// Note: This is a heuristic check; the actual empty content handling is done below.
 	hasEmptyContent := bytes.Contains(body, patternEmptyContent) ||
 		bytes.Contains(body, patternEmptyContentSpaced) ||
 		bytes.Contains(body, patternEmptyContentSp1) ||
 		bytes.Contains(body, patternEmptyContentSp2)
 
+	// Check for empty text blocks: {"type":"text","text":""}
+	// These cause upstream 400: "text content blocks must be non-empty"
+	hasEmptyTextBlock := bytes.Contains(body, patternEmptyText) ||
+		bytes.Contains(body, patternEmptyTextSpaced) ||
+		bytes.Contains(body, patternEmptyTextSp1) ||
+		bytes.Contains(body, patternEmptyTextSp2)
+
 	// Fast path: nothing to process
-	if !hasThinkingContent && !hasEmptyContent {
+	if !hasThinkingContent && !hasEmptyContent && !hasEmptyTextBlock {
 		return body
 	}
 
@@ -260,7 +273,7 @@ func FilterThinkingBlocksForRetry(body []byte) []byte {
 		bytes.Contains(body, patternTypeRedactedThinking) ||
 		bytes.Contains(body, patternTypeRedactedSpaced) ||
 		bytes.Contains(body, patternThinkingFieldSpaced)
-	if !hasEmptyContent && !containsThinkingBlocks {
+	if !hasEmptyContent && !hasEmptyTextBlock && !containsThinkingBlocks {
 		if topThinking := gjson.Get(jsonStr, "thinking"); topThinking.Exists() {
 			if out, err := sjson.DeleteBytes(body, "thinking"); err == nil {
 				out = removeThinkingDependentContextStrategies(out)
@@ -320,6 +333,16 @@ func FilterThinkingBlocksForRetry(body []byte) []byte {
 
 			blockType, _ := blockMap["type"].(string)
 
+			// Strip empty text blocks: {"type":"text","text":""}
+			// Upstream rejects these with 400: "text content blocks must be non-empty"
+			if blockType == "text" {
+				if txt, _ := blockMap["text"].(string); txt == "" {
+					modifiedThisMsg = true
+					ensureNewContent(bi)
+					continue
+				}
+			}
+
 			// Convert thinking blocks to text (preserve content) and drop redacted_thinking.
 			switch blockType {
 			case "thinking":

backend/internal/service/gateway_request_test.go

@@ -404,6 +404,51 @@ func TestFilterThinkingBlocksForRetry_EmptyContentGetsPlaceholder(t *testing.T)
 	require.NotEmpty(t, content0["text"])
 }
 
+func TestFilterThinkingBlocksForRetry_StripsEmptyTextBlocks(t *testing.T) {
+	// Empty text blocks cause upstream 400: "text content blocks must be non-empty"
+	input := []byte(`{
+		"messages":[
+			{"role":"user","content":[{"type":"text","text":"hello"},{"type":"text","text":""}]},
+			{"role":"assistant","content":[{"type":"text","text":""}]}
+		]
+	}`)
+
+	out := FilterThinkingBlocksForRetry(input)
+
+	var req map[string]any
+	require.NoError(t, json.Unmarshal(out, &req))
+	msgs, ok := req["messages"].([]any)
+	require.True(t, ok)
+
+	// First message: empty text block stripped, "hello" preserved
+	msg0 := msgs[0].(map[string]any)
+	content0 := msg0["content"].([]any)
+	require.Len(t, content0, 1)
+	require.Equal(t, "hello", content0[0].(map[string]any)["text"])
+
+	// Second message: only had empty text block → gets placeholder
+	msg1 := msgs[1].(map[string]any)
+	content1 := msg1["content"].([]any)
+	require.Len(t, content1, 1)
+	block1 := content1[0].(map[string]any)
+	require.Equal(t, "text", block1["type"])
+	require.NotEmpty(t, block1["text"])
+}
+
+func TestFilterThinkingBlocksForRetry_PreservesNonEmptyTextBlocks(t *testing.T) {
+	// Non-empty text blocks should pass through unchanged
+	input := []byte(`{
+		"messages":[
+			{"role":"user","content":[{"type":"text","text":"hello"},{"type":"text","text":"world"}]}
+		]
+	}`)
+
+	out := FilterThinkingBlocksForRetry(input)
+
+	// Fast path: no thinking content, no empty content, no empty text blocks → unchanged
+	require.Equal(t, input, out)
+}
+
 func TestFilterSignatureSensitiveBlocksForRetry_DowngradesTools(t *testing.T) {
 	input := []byte(`{
 		"thinking":{"type":"enabled","budget_tokens":1024},

backend/internal/service/gateway_service.go

@@ -6067,9 +6067,11 @@ func (s *GatewayService) isThinkingBlockSignatureError(respBody []byte) bool {
 		return true
 	}
 
-	// 检测空消息内容错误（可能是过滤 thinking blocks 后导致的）
+	// 检测空消息内容错误（可能是过滤 thinking blocks 后导致的，或客户端发送了空 text block）
 	// 例如: "all messages must have non-empty content"
-	if strings.Contains(msg, "non-empty content") || strings.Contains(msg, "empty content") {
+	//       "messages: text content blocks must be non-empty"
+	if strings.Contains(msg, "non-empty content") || strings.Contains(msg, "empty content") ||
+		strings.Contains(msg, "content blocks must be non-empty") {
 		logger.LegacyPrintf("service.gateway", "[SignatureCheck] Detected empty content error")
 		return true
 	}