Merge pull request #2 from Khan/release/v1.40.5

Release/v1.40.5

Author: rjcorwin

.codecov.yml

@@ -0,0 +1,4 @@
+coverage:
+  ignore:
+    - "examples/**"
+    - "internal/test/**"

.github/workflows/pr.yml

@@ -16,7 +16,7 @@ jobs:
         go-version: '1.24'
     - name: Run vet
       run: |
-        go vet .
+        go vet -stdversion ./...
     - name: Run golangci-lint
       uses: golangci/golangci-lint-action@v7
       with:

chat.go

@@ -5,6 +5,8 @@ import (
 	"encoding/json"
 	"errors"
 	"net/http"
+
+	"github.com/sashabaranov/go-openai/jsonschema"
 )
 
 // Chat message role defined by the OpenAI API.
@@ -221,6 +223,31 @@ type ChatCompletionResponseFormatJSONSchema struct {
 	Strict      bool           `json:"strict"`
 }
 
+func (r *ChatCompletionResponseFormatJSONSchema) UnmarshalJSON(data []byte) error {
+	type rawJSONSchema struct {
+		Name        string          `json:"name"`
+		Description string          `json:"description,omitempty"`
+		Schema      json.RawMessage `json:"schema"`
+		Strict      bool            `json:"strict"`
+	}
+	var raw rawJSONSchema
+	if err := json.Unmarshal(data, &raw); err != nil {
+		return err
+	}
+	r.Name = raw.Name
+	r.Description = raw.Description
+	r.Strict = raw.Strict
+	if len(raw.Schema) > 0 && string(raw.Schema) != "null" {
+		var d jsonschema.Definition
+		err := json.Unmarshal(raw.Schema, &d)
+		if err != nil {
+			return err
+		}
+		r.Schema = &d
+	}
+	return nil
+}
+
 // ChatCompletionRequest represents a request structure for chat completion API.
 type ChatCompletionRequest struct {
 	Model    string                  `json:"model"`
@@ -280,6 +307,8 @@ type ChatCompletionRequest struct {
 	// Such as think mode for qwen3. "chat_template_kwargs": {"enable_thinking": false}
 	// https://qwen.readthedocs.io/en/latest/deployment/vllm.html#thinking-non-thinking-modes
 	ChatTemplateKwargs map[string]any `json:"chat_template_kwargs,omitempty"`
+	// Specifies the latency tier to use for processing the request.
+	ServiceTier ServiceTier `json:"service_tier,omitempty"`
 }
 
 type StreamOptions struct {
@@ -363,6 +392,15 @@ const (
 	FinishReasonNull          FinishReason = "null"
 )
 
+type ServiceTier string
+
+const (
+	ServiceTierAuto     ServiceTier = "auto"
+	ServiceTierDefault  ServiceTier = "default"
+	ServiceTierFlex     ServiceTier = "flex"
+	ServiceTierPriority ServiceTier = "priority"
+)
+
 func (r FinishReason) MarshalJSON() ([]byte, error) {
 	if r == FinishReasonNull || r == "" {
 		return []byte("null"), nil
@@ -395,6 +433,7 @@ type ChatCompletionResponse struct {
 	Usage               Usage                  `json:"usage"`
 	SystemFingerprint   string                 `json:"system_fingerprint"`
 	PromptFilterResults []PromptFilterResult   `json:"prompt_filter_results,omitempty"`
+	ServiceTier         ServiceTier            `json:"service_tier,omitempty"`
 
 	httpHeader
 }

chat_test.go

@@ -946,3 +946,142 @@ func TestFinishReason(t *testing.T) {
 		}
 	}
 }
+
+func TestChatCompletionResponseFormatJSONSchema_UnmarshalJSON(t *testing.T) {
+	type args struct {
+		data []byte
+	}
+	tests := []struct {
+		name    string
+		args    args
+		wantErr bool
+	}{
+		{
+			"",
+			args{
+				data: []byte(`{
+      "name":   "math_response",
+      "strict": true,
+      "schema": {
+        "type": "object",
+        "properties": {
+          "steps": {
+            "type": "array",
+            "items": {
+              "type": "object",
+              "properties": {
+                "explanation": { "type": "string" },
+                "output":      { "type": "string" }
+              },
+              "required": ["explanation","output"],
+              "additionalProperties": false
+            }
+          },
+          "final_answer": { "type": "string" }
+        },
+        "required": ["steps","final_answer"],
+        "additionalProperties": false
+      }
+  }`),
+			},
+			false,
+		},
+		{
+			"",
+			args{
+				data: []byte(`{
+      "name":   "math_response",
+      "strict": true,
+      "schema": null
+  }`),
+			},
+			false,
+		},
+		{
+			"",
+			args{
+				data: []byte(`[123,456]`),
+			},
+			true,
+		},
+		{
+			"",
+			args{
+				data: []byte(`{
+      "name":   "math_response",
+      "strict": true,
+      "schema": 123456
+  }`),
+			},
+			true,
+		},
+	}
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			var r openai.ChatCompletionResponseFormatJSONSchema
+			err := r.UnmarshalJSON(tt.args.data)
+			if (err != nil) != tt.wantErr {
+				t.Errorf("UnmarshalJSON() error = %v, wantErr %v", err, tt.wantErr)
+			}
+		})
+	}
+}
+
+func TestChatCompletionRequest_UnmarshalJSON(t *testing.T) {
+	type args struct {
+		bs []byte
+	}
+	tests := []struct {
+		name    string
+		args    args
+		wantErr bool
+	}{
+		{
+			"",
+			args{bs: []byte(`{
+  "model": "llama3-1b",
+  "messages": [
+    { "role": "system", "content": "You are a helpful math tutor." },
+    { "role": "user",   "content": "solve 8x + 31 = 2" }
+  ],
+  "response_format": {
+    "type": "json_schema",
+    "json_schema": {
+      "name":   "math_response",
+      "strict": true,
+      "schema": {
+        "type": "object",
+        "properties": {
+          "steps": {
+            "type": "array",
+            "items": {
+              "type": "object",
+              "properties": {
+                "explanation": { "type": "string" },
+                "output":      { "type": "string" }
+              },
+              "required": ["explanation","output"],
+              "additionalProperties": false
+            }
+          },
+          "final_answer": { "type": "string" }
+        },
+        "required": ["steps","final_answer"],
+        "additionalProperties": false
+      }
+    }
+  }
+}`)},
+			false,
+		},
+	}
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			var m openai.ChatCompletionRequest
+			err := json.Unmarshal(tt.args.bs, &m)
+			if err != nil {
+				t.Errorf("UnmarshalJSON() error = %v, wantErr %v", err, tt.wantErr)
+			}
+		})
+	}
+}

client.go

@@ -71,6 +71,17 @@ func NewOrgClient(authToken, org string) *Client {
 	return NewClientWithConfig(config)
 }
 
+// SetBaseURL updates the base URL for the client.
+// This allows changing the endpoint after client instantiation.
+func (c *Client) SetBaseURL(baseURL string) {
+	c.config.BaseURL = baseURL
+}
+
+// GetBaseURL returns the current base URL for the client.
+func (c *Client) GetBaseURL() string {
+	return c.config.BaseURL
+}
+
 type requestOptions struct {
 	body   any
 	header http.Header
@@ -84,6 +95,20 @@ func withBody(body any) requestOption {
 	}
 }
 
+func withExtraBody(extraBody map[string]any) requestOption {
+	return func(args *requestOptions) {
+		// Assert that args.body is a map[string]any.
+		bodyMap, ok := args.body.(map[string]any)
+		if ok {
+			// If it's a map[string]any then only add extraBody
+			// fields to args.body otherwise keep only fields in request struct.
+			for key, value := range extraBody {
+				bodyMap[key] = value
+			}
+		}
+	}
+}
+
 func withContentType(contentType string) requestOption {
 	return func(args *requestOptions) {
 		args.header.Set("Content-Type", contentType)

client_test.go

@@ -39,6 +39,35 @@ func TestClient(t *testing.T) {
 	}
 }
 
+func TestClient_SetBaseURL(t *testing.T) {
+	const mockToken = "mock token"
+	client := NewClient(mockToken)
+
+	// Test default base URL
+	defaultBaseURL := client.GetBaseURL()
+	if defaultBaseURL != openaiAPIURLv1 {
+		t.Errorf("Expected default BaseURL to be %q, got %q", openaiAPIURLv1, defaultBaseURL)
+	}
+
+	// Test setting a new base URL
+	const newBaseURL = "https://custom-api.example.com/v1"
+	client.SetBaseURL(newBaseURL)
+
+	// Verify the base URL was updated
+	updatedBaseURL := client.GetBaseURL()
+	if updatedBaseURL != newBaseURL {
+		t.Errorf("Expected BaseURL to be %q after SetBaseURL, got %q", newBaseURL, updatedBaseURL)
+	}
+
+	// Test that the updated base URL is used in fullURL method
+	suffix := "/chat/completions"
+	fullURL := client.fullURL(suffix)
+	expectedFullURL := newBaseURL + suffix
+	if fullURL != expectedFullURL {
+		t.Errorf("Expected fullURL to be %q, got %q", expectedFullURL, fullURL)
+	}
+}
+
 func TestSetCommonHeadersAnthropic(t *testing.T) {
 	config := DefaultAnthropicConfig("mock-token", "")
 	client := NewClientWithConfig(config)

completion.go

@@ -242,7 +242,7 @@ type CompletionResponse struct {
 	Created int64              `json:"created"`
 	Model   string             `json:"model"`
 	Choices []CompletionChoice `json:"choices"`
-	Usage   Usage              `json:"usage"`
+	Usage   *Usage             `json:"usage,omitempty"`
 
 	httpHeader
 }

completion_test.go

@@ -192,7 +192,7 @@ func handleCompletionEndpoint(w http.ResponseWriter, r *http.Request) {
 	}
 	inputTokens *= n
 	completionTokens := completionReq.MaxTokens * len(prompts) * n
-	res.Usage = openai.Usage{
+	res.Usage = &openai.Usage{
 		PromptTokens:     inputTokens,
 		CompletionTokens: completionTokens,
 		TotalTokens:      inputTokens + completionTokens,