Merge pull request Wei-Shaw#1212 from alfadb/fix/filter-empty-text-blocks-nested

fix(gateway): 修复 tool_result 嵌套内容中空 text block 导致上游 400 错误

Author: Wei-Shaw

backend/internal/service/gateway_request.go

@@ -253,6 +253,118 @@ func sliceRawFromBody(body []byte, r gjson.Result) []byte {
 	return []byte(r.Raw)
 }
 
+// stripEmptyTextBlocksFromSlice removes empty text blocks from a content slice (including nested tool_result content).
+// Returns (cleaned slice, true) if any blocks were removed, or (original, false) if unchanged.
+func stripEmptyTextBlocksFromSlice(blocks []any) ([]any, bool) {
+	var result []any
+	changed := false
+	for i, block := range blocks {
+		blockMap, ok := block.(map[string]any)
+		if !ok {
+			if result != nil {
+				result = append(result, block)
+			}
+			continue
+		}
+		blockType, _ := blockMap["type"].(string)
+
+		// Strip empty text blocks
+		if blockType == "text" {
+			if txt, _ := blockMap["text"].(string); txt == "" {
+				if result == nil {
+					result = make([]any, 0, len(blocks))
+					result = append(result, blocks[:i]...)
+				}
+				changed = true
+				continue
+			}
+		}
+
+		// Recurse into tool_result nested content
+		if blockType == "tool_result" {
+			if nestedContent, ok := blockMap["content"].([]any); ok {
+				if cleaned, nestedChanged := stripEmptyTextBlocksFromSlice(nestedContent); nestedChanged {
+					if result == nil {
+						result = make([]any, 0, len(blocks))
+						result = append(result, blocks[:i]...)
+					}
+					changed = true
+					blockCopy := make(map[string]any, len(blockMap))
+					for k, v := range blockMap {
+						blockCopy[k] = v
+					}
+					blockCopy["content"] = cleaned
+					result = append(result, blockCopy)
+					continue
+				}
+			}
+		}
+
+		if result != nil {
+			result = append(result, block)
+		}
+	}
+	if !changed {
+		return blocks, false
+	}
+	return result, true
+}
+
+// StripEmptyTextBlocks removes empty text blocks from the request body (including nested tool_result content).
+// This is a lightweight pre-filter for the initial request path to prevent upstream 400 errors.
+// Returns the original body unchanged if no empty text blocks are found.
+func StripEmptyTextBlocks(body []byte) []byte {
+	// Fast path: check if body contains empty text patterns
+	hasEmptyTextBlock := bytes.Contains(body, patternEmptyText) ||
+		bytes.Contains(body, patternEmptyTextSpaced) ||
+		bytes.Contains(body, patternEmptyTextSp1) ||
+		bytes.Contains(body, patternEmptyTextSp2)
+	if !hasEmptyTextBlock {
+		return body
+	}
+
+	jsonStr := *(*string)(unsafe.Pointer(&body))
+	msgsRes := gjson.Get(jsonStr, "messages")
+	if !msgsRes.Exists() || !msgsRes.IsArray() {
+		return body
+	}
+
+	var messages []any
+	if err := json.Unmarshal(sliceRawFromBody(body, msgsRes), &messages); err != nil {
+		return body
+	}
+
+	modified := false
+	for _, msg := range messages {
+		msgMap, ok := msg.(map[string]any)
+		if !ok {
+			continue
+		}
+		content, ok := msgMap["content"].([]any)
+		if !ok {
+			continue
+		}
+		if cleaned, changed := stripEmptyTextBlocksFromSlice(content); changed {
+			modified = true
+			msgMap["content"] = cleaned
+		}
+	}
+
+	if !modified {
+		return body
+	}
+
+	msgsBytes, err := json.Marshal(messages)
+	if err != nil {
+		return body
+	}
+	out, err := sjson.SetRawBytes(body, "messages", msgsBytes)
+	if err != nil {
+		return body
+	}
+	return out
+}
+
 // FilterThinkingBlocks removes thinking blocks from request body
 // Returns filtered body or original body if filtering fails (fail-safe)
 // This prevents 400 errors from invalid thinking block signatures
@@ -426,6 +538,23 @@ func FilterThinkingBlocksForRetry(body []byte) []byte {
 				}
 			}
 
+			// Recursively strip empty text blocks from tool_result nested content.
+			if blockType == "tool_result" {
+				if nestedContent, ok := blockMap["content"].([]any); ok {
+					if cleaned, changed := stripEmptyTextBlocksFromSlice(nestedContent); changed {
+						modifiedThisMsg = true
+						ensureNewContent(bi)
+						blockCopy := make(map[string]any, len(blockMap))
+						for k, v := range blockMap {
+							blockCopy[k] = v
+						}
+						blockCopy["content"] = cleaned
+						newContent = append(newContent, blockCopy)
+						continue
+					}
+				}
+			}
+
 			if newContent != nil {
 				newContent = append(newContent, block)
 			}

backend/internal/service/gateway_request_test.go

@@ -435,6 +435,122 @@ func TestFilterThinkingBlocksForRetry_StripsEmptyTextBlocks(t *testing.T) {
 	require.NotEmpty(t, block1["text"])
 }
 
+func TestFilterThinkingBlocksForRetry_StripsNestedEmptyTextInToolResult(t *testing.T) {
+	// Empty text blocks nested inside tool_result content should also be stripped
+	input := []byte(`{
+		"messages":[
+			{"role":"user","content":[
+				{"type":"tool_result","tool_use_id":"t1","content":[
+					{"type":"text","text":"valid result"},
+					{"type":"text","text":""}
+				]}
+			]}
+		]
+	}`)
+
+	out := FilterThinkingBlocksForRetry(input)
+
+	var req map[string]any
+	require.NoError(t, json.Unmarshal(out, &req))
+	msgs := req["messages"].([]any)
+	msg0 := msgs[0].(map[string]any)
+	content0 := msg0["content"].([]any)
+	require.Len(t, content0, 1)
+	toolResult := content0[0].(map[string]any)
+	require.Equal(t, "tool_result", toolResult["type"])
+	nestedContent := toolResult["content"].([]any)
+	require.Len(t, nestedContent, 1)
+	require.Equal(t, "valid result", nestedContent[0].(map[string]any)["text"])
+}
+
+func TestFilterThinkingBlocksForRetry_NestedAllEmptyGetsEmptySlice(t *testing.T) {
+	// If all nested content blocks in tool_result are empty text, content becomes empty slice
+	input := []byte(`{
+		"messages":[
+			{"role":"user","content":[
+				{"type":"tool_result","tool_use_id":"t1","content":[
+					{"type":"text","text":""}
+				]},
+				{"type":"text","text":"hello"}
+			]}
+		]
+	}`)
+
+	out := FilterThinkingBlocksForRetry(input)
+
+	var req map[string]any
+	require.NoError(t, json.Unmarshal(out, &req))
+	msgs := req["messages"].([]any)
+	msg0 := msgs[0].(map[string]any)
+	content0 := msg0["content"].([]any)
+	require.Len(t, content0, 2)
+	toolResult := content0[0].(map[string]any)
+	nestedContent := toolResult["content"].([]any)
+	require.Len(t, nestedContent, 0)
+}
+
+func TestStripEmptyTextBlocks(t *testing.T) {
+	t.Run("strips top-level empty text", func(t *testing.T) {
+		input := []byte(`{"messages":[{"role":"user","content":[{"type":"text","text":"hello"},{"type":"text","text":""}]}]}`)
+		out := StripEmptyTextBlocks(input)
+		var req map[string]any
+		require.NoError(t, json.Unmarshal(out, &req))
+		msgs := req["messages"].([]any)
+		content := msgs[0].(map[string]any)["content"].([]any)
+		require.Len(t, content, 1)
+		require.Equal(t, "hello", content[0].(map[string]any)["text"])
+	})
+
+	t.Run("strips nested empty text in tool_result", func(t *testing.T) {
+		input := []byte(`{"messages":[{"role":"user","content":[{"type":"tool_result","tool_use_id":"t1","content":[{"type":"text","text":"ok"},{"type":"text","text":""}]}]}]}`)
+		out := StripEmptyTextBlocks(input)
+		var req map[string]any
+		require.NoError(t, json.Unmarshal(out, &req))
+		msgs := req["messages"].([]any)
+		content := msgs[0].(map[string]any)["content"].([]any)
+		toolResult := content[0].(map[string]any)
+		nestedContent := toolResult["content"].([]any)
+		require.Len(t, nestedContent, 1)
+		require.Equal(t, "ok", nestedContent[0].(map[string]any)["text"])
+	})
+
+	t.Run("no-op when no empty text", func(t *testing.T) {
+		input := []byte(`{"messages":[{"role":"user","content":[{"type":"text","text":"hello"}]}]}`)
+		out := StripEmptyTextBlocks(input)
+		require.Equal(t, input, out)
+	})
+
+	t.Run("preserves non-map blocks in content", func(t *testing.T) {
+		// tool_result content can be a string; non-map blocks should pass through unchanged
+		input := []byte(`{"messages":[{"role":"user","content":[{"type":"tool_result","tool_use_id":"t1","content":"string content"},{"type":"text","text":""}]}]}`)
+		out := StripEmptyTextBlocks(input)
+		var req map[string]any
+		require.NoError(t, json.Unmarshal(out, &req))
+		msgs := req["messages"].([]any)
+		content := msgs[0].(map[string]any)["content"].([]any)
+		require.Len(t, content, 1)
+		toolResult := content[0].(map[string]any)
+		require.Equal(t, "tool_result", toolResult["type"])
+		require.Equal(t, "string content", toolResult["content"])
+	})
+
+	t.Run("handles deeply nested tool_result", func(t *testing.T) {
+		// Recursive: tool_result containing another tool_result with empty text
+		input := []byte(`{"messages":[{"role":"user","content":[{"type":"tool_result","tool_use_id":"t1","content":[{"type":"tool_result","tool_use_id":"t2","content":[{"type":"text","text":""},{"type":"text","text":"deep"}]}]}]}]}`)
+		out := StripEmptyTextBlocks(input)
+		var req map[string]any
+		require.NoError(t, json.Unmarshal(out, &req))
+		msgs := req["messages"].([]any)
+		content := msgs[0].(map[string]any)["content"].([]any)
+		outer := content[0].(map[string]any)
+		innerContent := outer["content"].([]any)
+		inner := innerContent[0].(map[string]any)
+		deepContent := inner["content"].([]any)
+		require.Len(t, deepContent, 1)
+		require.Equal(t, "deep", deepContent[0].(map[string]any)["text"])
+	})
+}
+
 func TestFilterThinkingBlocksForRetry_PreservesNonEmptyTextBlocks(t *testing.T) {
 	// Non-empty text blocks should pass through unchanged
 	input := []byte(`{

backend/internal/service/gateway_service.go

@@ -4119,6 +4119,9 @@ func (s *GatewayService) Forward(ctx context.Context, c *gin.Context, account *A
 	// 调试日志：记录即将转发的账号信息
 	logger.LegacyPrintf("service.gateway", "[Forward] Using account: ID=%d Name=%s Platform=%s Type=%s TLSFingerprint=%v Proxy=%s",
 		account.ID, account.Name, account.Platform, account.Type, account.IsTLSFingerprintEnabled(), proxyURL)
+	// Pre-filter: strip empty text blocks (including nested in tool_result) to prevent upstream 400.
+	body = StripEmptyTextBlocks(body)
+
 	// 重试间复用同一请求体，避免每次 string(body) 产生额外分配。
 	setOpsUpstreamRequestBody(c, body)
 
@@ -4609,6 +4612,9 @@ func (s *GatewayService) forwardAnthropicAPIKeyPassthroughWithInput(
 	if c != nil {
 		c.Set("anthropic_passthrough", true)
 	}
+	// Pre-filter: strip empty text blocks (including nested in tool_result) to prevent upstream 400.
+	input.Body = StripEmptyTextBlocks(input.Body)
+
 	// 重试间复用同一请求体，避免每次 string(body) 产生额外分配。
 	setOpsUpstreamRequestBody(c, input.Body)
 
@@ -7887,6 +7893,9 @@ func (s *GatewayService) ForwardCountTokens(ctx context.Context, c *gin.Context,
 	body := parsed.Body
 	reqModel := parsed.Model
 
+	// Pre-filter: strip empty text blocks to prevent upstream 400.
+	body = StripEmptyTextBlocks(body)
+
 	isClaudeCode := isClaudeCodeRequest(ctx, c, parsed)
 	shouldMimicClaudeCode := account.IsOAuth() && !isClaudeCode