Merge branch 'sashabaranov:master' into master

Author: fcce

.gitignore

@@ -16,4 +16,7 @@
 
 # Auth token for tests
 .openai-token
-.idea
+.idea
+
+# Generated by tests
+test.mp3

.golangci.yml

@@ -57,7 +57,7 @@ linters-settings:
         # Default: true
         skipRecvDeref: false
 
-  gomnd:
+  mnd:
     # List of function patterns to exclude from analysis.
     # Values always ignored: `time.Date`
     # Default: []
@@ -167,7 +167,7 @@ linters:
     - durationcheck # check for two durations multiplied together
     - errname # Checks that sentinel errors are prefixed with the Err and error types are suffixed with the Error.
     - errorlint # errorlint is a linter for that can be used to find code that will cause problems with the error wrapping scheme introduced in Go 1.13.
-    - execinquery # execinquery is a linter about query string checker in Query function which reads your Go src files and warning it finds
+    # Removed execinquery (deprecated). execinquery is a linter about query string checker in Query function which reads your Go src files and warning it finds
     - exhaustive # check exhaustiveness of enum switch statements
     - exportloopref # checks for pointers to enclosing loop variables
     - forbidigo # Forbids identifiers
@@ -180,14 +180,14 @@ linters:
     - gocyclo # Computes and checks the cyclomatic complexity of functions
     - godot # Check if comments end in a period
     - goimports # In addition to fixing imports, goimports also formats your code in the same style as gofmt.
-    - gomnd # An analyzer to detect magic numbers.
     - gomoddirectives # Manage the use of 'replace', 'retract', and 'excludes' directives in go.mod.
     - gomodguard # Allow and block list linter for direct Go module dependencies. This is different from depguard where there are different block types for example version constraints and module recommendations.
     - goprintffuncname # Checks that printf-like functions are named with f at the end
     - gosec # Inspects source code for security problems
     - lll # Reports long lines
     - makezero # Finds slice declarations with non-zero initial length
     # - nakedret # Finds naked returns in functions greater than a specified function length
+    - mnd # An analyzer to detect magic numbers.
     - nestif # Reports deeply nested if statements
     - nilerr # Finds the code that returns nil even if it checks that the error is not nil.
     - nilnil # Checks that there is no simultaneous return of nil error and an invalid value.

Makefile

@@ -5,7 +5,7 @@
 
 This library provides unofficial Go clients for [OpenAI API](https://platform.openai.com/). We support: 
 
-* ChatGPT
+* ChatGPT 4o, o1
 * GPT-3, GPT-4
 * DALL·E 2, DALL·E 3
 * Whisper

README.md

@@ -14,17 +14,20 @@ const (
 )
 
 type Assistant struct {
-	ID            string                 `json:"id"`
-	Object        string                 `json:"object"`
-	CreatedAt     int64                  `json:"created_at"`
-	Name          *string                `json:"name,omitempty"`
-	Description   *string                `json:"description,omitempty"`
-	Model         string                 `json:"model"`
-	Instructions  *string                `json:"instructions,omitempty"`
-	Tools         []AssistantTool        `json:"tools"`
-	FileIDs       []string               `json:"file_ids,omitempty"`
-	Metadata      map[string]any         `json:"metadata,omitempty"`
-	ToolResources *AssistantToolResource `json:"tool_resources,omitempty"`
+	ID             string                 `json:"id"`
+	Object         string                 `json:"object"`
+	CreatedAt      int64                  `json:"created_at"`
+	Name           *string                `json:"name,omitempty"`
+	Description    *string                `json:"description,omitempty"`
+	Model          string                 `json:"model"`
+	Instructions   *string                `json:"instructions,omitempty"`
+	Tools          []AssistantTool        `json:"tools"`
+	ToolResources  *AssistantToolResource `json:"tool_resources,omitempty"`
+	FileIDs        []string               `json:"file_ids,omitempty"` // Deprecated in v2
+	Metadata       map[string]any         `json:"metadata,omitempty"`
+	Temperature    *float32               `json:"temperature,omitempty"`
+	TopP           *float32               `json:"top_p,omitempty"`
+	ResponseFormat any                    `json:"response_format,omitempty"`
 
 	httpHeader
 }

assistant.go

@@ -211,7 +211,7 @@ func TestUploadBatchFileRequest_AddEmbedding(t *testing.T) {
 					Input: []string{"Hello", "World"},
 				},
 			},
-		}, []byte("{\"custom_id\":\"req-1\",\"body\":{\"input\":[\"Hello\",\"World\"],\"model\":\"gpt-3.5-turbo\",\"user\":\"\"},\"method\":\"POST\",\"url\":\"/v1/embeddings\"}\n{\"custom_id\":\"req-2\",\"body\":{\"input\":[\"Hello\",\"World\"],\"model\":\"text-embedding-ada-002\",\"user\":\"\"},\"method\":\"POST\",\"url\":\"/v1/embeddings\"}")}, //nolint:lll
+		}, []byte("{\"custom_id\":\"req-1\",\"body\":{\"input\":[\"Hello\",\"World\"],\"model\":\"gpt-3.5-turbo\"},\"method\":\"POST\",\"url\":\"/v1/embeddings\"}\n{\"custom_id\":\"req-2\",\"body\":{\"input\":[\"Hello\",\"World\"],\"model\":\"text-embedding-ada-002\"},\"method\":\"POST\",\"url\":\"/v1/embeddings\"}")}, //nolint:lll
 	}
 	for _, tt := range tests {
 		t.Run(tt.name, func(t *testing.T) {

batch_test.go

@@ -41,11 +41,23 @@ type Violence struct {
 	Severity string `json:"severity,omitempty"`
 }
 
+type JailBreak struct {
+	Filtered bool `json:"filtered"`
+	Detected bool `json:"detected"`
+}
+
+type Profanity struct {
+	Filtered bool `json:"filtered"`
+	Detected bool `json:"detected"`
+}
+
 type ContentFilterResults struct {
-	Hate     Hate     `json:"hate,omitempty"`
-	SelfHarm SelfHarm `json:"self_harm,omitempty"`
-	Sexual   Sexual   `json:"sexual,omitempty"`
-	Violence Violence `json:"violence,omitempty"`
+	Hate      Hate      `json:"hate,omitempty"`
+	SelfHarm  SelfHarm  `json:"self_harm,omitempty"`
+	Sexual    Sexual    `json:"sexual,omitempty"`
+	Violence  Violence  `json:"violence,omitempty"`
+	JailBreak JailBreak `json:"jailbreak,omitempty"`
+	Profanity Profanity `json:"profanity,omitempty"`
 }
 
 type PromptAnnotation struct {
@@ -82,6 +94,7 @@ type ChatMessagePart struct {
 type ChatCompletionMessage struct {
 	Role         string `json:"role"`
 	Content      string `json:"content"`
+	Refusal      string `json:"refusal,omitempty"`
 	MultiContent []ChatMessagePart
 
 	// This property isn't in the official documentation, but it's in
@@ -107,6 +120,7 @@ func (m ChatCompletionMessage) MarshalJSON() ([]byte, error) {
 		msg := struct {
 			Role         string            `json:"role"`
 			Content      string            `json:"-"`
+			Refusal      string            `json:"refusal,omitempty"`
 			MultiContent []ChatMessagePart `json:"content,omitempty"`
 			Name         string            `json:"name,omitempty"`
 			FunctionCall *FunctionCall     `json:"function_call,omitempty"`
@@ -115,9 +129,11 @@ func (m ChatCompletionMessage) MarshalJSON() ([]byte, error) {
 		}(m)
 		return json.Marshal(msg)
 	}
+
 	msg := struct {
 		Role         string            `json:"role"`
 		Content      string            `json:"content"`
+		Refusal      string            `json:"refusal,omitempty"`
 		MultiContent []ChatMessagePart `json:"-"`
 		Name         string            `json:"name,omitempty"`
 		FunctionCall *FunctionCall     `json:"function_call,omitempty"`
@@ -131,19 +147,22 @@ func (m *ChatCompletionMessage) UnmarshalJSON(bs []byte) error {
 	msg := struct {
 		Role         string `json:"role"`
 		Content      string `json:"content"`
+		Refusal      string `json:"refusal,omitempty"`
 		MultiContent []ChatMessagePart
 		Name         string        `json:"name,omitempty"`
 		FunctionCall *FunctionCall `json:"function_call,omitempty"`
 		ToolCalls    []ToolCall    `json:"tool_calls,omitempty"`
 		ToolCallID   string        `json:"tool_call_id,omitempty"`
 	}{}
+
 	if err := json.Unmarshal(bs, &msg); err == nil {
 		*m = ChatCompletionMessage(msg)
 		return nil
 	}
 	multiMsg := struct {
 		Role         string `json:"role"`
 		Content      string
+		Refusal      string            `json:"refusal,omitempty"`
 		MultiContent []ChatMessagePart `json:"content"`
 		Name         string            `json:"name,omitempty"`
 		FunctionCall *FunctionCall     `json:"function_call,omitempty"`
@@ -160,7 +179,7 @@ func (m *ChatCompletionMessage) UnmarshalJSON(bs []byte) error {
 type ToolCall struct {
 	// Index is not nil only in chat completion chunk object
 	Index    *int         `json:"index,omitempty"`
-	ID       string       `json:"id"`
+	ID       string       `json:"id,omitempty"`
 	Type     ToolType     `json:"type"`
 	Function FunctionCall `json:"function"`
 }
@@ -193,18 +212,25 @@ type ChatCompletionResponseFormatJSONSchema struct {
 
 // ChatCompletionRequest represents a request structure for chat completion API.
 type ChatCompletionRequest struct {
-	Model            string                        `json:"model"`
-	Messages         []ChatCompletionMessage       `json:"messages"`
-	MaxTokens        int                           `json:"max_tokens,omitempty"`
-	Temperature      float32                       `json:"temperature,omitempty"`
-	TopP             float32                       `json:"top_p,omitempty"`
-	N                int                           `json:"n,omitempty"`
-	Stream           bool                          `json:"stream,omitempty"`
-	Stop             []string                      `json:"stop,omitempty"`
-	PresencePenalty  float32                       `json:"presence_penalty,omitempty"`
-	ResponseFormat   *ChatCompletionResponseFormat `json:"response_format,omitempty"`
-	Seed             *int                          `json:"seed,omitempty"`
-	FrequencyPenalty float32                       `json:"frequency_penalty,omitempty"`
+	Model    string                  `json:"model"`
+	Messages []ChatCompletionMessage `json:"messages"`
+	// MaxTokens The maximum number of tokens that can be generated in the chat completion.
+	// This value can be used to control costs for text generated via API.
+	// This value is now deprecated in favor of max_completion_tokens, and is not compatible with o1 series models.
+	// refs: https://platform.openai.com/docs/api-reference/chat/create#chat-create-max_tokens
+	MaxTokens int `json:"max_tokens,omitempty"`
+	// MaxCompletionTokens An upper bound for the number of tokens that can be generated for a completion,
+	// including visible output tokens and reasoning tokens https://platform.openai.com/docs/guides/reasoning
+	MaxCompletionTokens int                           `json:"max_completion_tokens,omitempty"`
+	Temperature         float32                       `json:"temperature,omitempty"`
+	TopP                float32                       `json:"top_p,omitempty"`
+	N                   int                           `json:"n,omitempty"`
+	Stream              bool                          `json:"stream,omitempty"`
+	Stop                []string                      `json:"stop,omitempty"`
+	PresencePenalty     float32                       `json:"presence_penalty,omitempty"`
+	ResponseFormat      *ChatCompletionResponseFormat `json:"response_format,omitempty"`
+	Seed                *int                          `json:"seed,omitempty"`
+	FrequencyPenalty    float32                       `json:"frequency_penalty,omitempty"`
 	// LogitBias is must be a token id string (specified by their token ID in the tokenizer), not a word string.
 	// incorrect: `"logit_bias":{"You": 6}`, correct: `"logit_bias":{"1639": 6}`
 	// refs: https://platform.openai.com/docs/api-reference/chat/create#chat/create-logit_bias
@@ -229,6 +255,11 @@ type ChatCompletionRequest struct {
 	StreamOptions *StreamOptions `json:"stream_options,omitempty"`
 	// Disable the default behavior of parallel tool calls by setting it: false.
 	ParallelToolCalls any `json:"parallel_tool_calls,omitempty"`
+	// Store can be set to true to store the output of this completion request for use in distillations and evals.
+	// https://platform.openai.com/docs/api-reference/chat/create#chat-create-store
+	Store bool `json:"store,omitempty"`
+	// Metadata to store with the completion.
+	Metadata map[string]string `json:"metadata,omitempty"`
 }
 
 type StreamOptions struct {
@@ -324,19 +355,21 @@ type ChatCompletionChoice struct {
 	// function_call: The model decided to call a function
 	// content_filter: Omitted content due to a flag from our content filters
 	// null: API response still in progress or incomplete
-	FinishReason FinishReason `json:"finish_reason"`
-	LogProbs     *LogProbs    `json:"logprobs,omitempty"`
+	FinishReason         FinishReason         `json:"finish_reason"`
+	LogProbs             *LogProbs            `json:"logprobs,omitempty"`
+	ContentFilterResults ContentFilterResults `json:"content_filter_results,omitempty"`
 }
 
 // ChatCompletionResponse represents a response structure for chat completion API.
 type ChatCompletionResponse struct {
-	ID                string                 `json:"id"`
-	Object            string                 `json:"object"`
-	Created           int64                  `json:"created"`
-	Model             string                 `json:"model"`
-	Choices           []ChatCompletionChoice `json:"choices"`
-	Usage             Usage                  `json:"usage"`
-	SystemFingerprint string                 `json:"system_fingerprint"`
+	ID                  string                 `json:"id"`
+	Object              string                 `json:"object"`
+	Created             int64                  `json:"created"`
+	Model               string                 `json:"model"`
+	Choices             []ChatCompletionChoice `json:"choices"`
+	Usage               Usage                  `json:"usage"`
+	SystemFingerprint   string                 `json:"system_fingerprint"`
+	PromptFilterResults []PromptFilterResult   `json:"prompt_filter_results,omitempty"`
 
 	httpHeader
 }
@@ -357,6 +390,10 @@ func (c *Client) CreateChatCompletion(
 		return
 	}
 
+	if err = validateRequestForO1Models(request); err != nil {
+		return
+	}
+
 	req, err := c.newRequest(
 		ctx,
 		http.MethodPost,

chat.go

@@ -10,13 +10,33 @@ type ChatCompletionStreamChoiceDelta struct {
 	Role         string        `json:"role,omitempty"`
 	FunctionCall *FunctionCall `json:"function_call,omitempty"`
 	ToolCalls    []ToolCall    `json:"tool_calls,omitempty"`
+	Refusal      string        `json:"refusal,omitempty"`
+}
+
+type ChatCompletionStreamChoiceLogprobs struct {
+	Content []ChatCompletionTokenLogprob `json:"content,omitempty"`
+	Refusal []ChatCompletionTokenLogprob `json:"refusal,omitempty"`
+}
+
+type ChatCompletionTokenLogprob struct {
+	Token       string                                 `json:"token"`
+	Bytes       []int64                                `json:"bytes,omitempty"`
+	Logprob     float64                                `json:"logprob,omitempty"`
+	TopLogprobs []ChatCompletionTokenLogprobTopLogprob `json:"top_logprobs"`
+}
+
+type ChatCompletionTokenLogprobTopLogprob struct {
+	Token   string  `json:"token"`
+	Bytes   []int64 `json:"bytes"`
+	Logprob float64 `json:"logprob"`
 }
 
 type ChatCompletionStreamChoice struct {
-	Index                int                             `json:"index"`
-	Delta                ChatCompletionStreamChoiceDelta `json:"delta"`
-	FinishReason         FinishReason                    `json:"finish_reason"`
-	ContentFilterResults ContentFilterResults            `json:"content_filter_results,omitempty"`
+	Index                int                                 `json:"index"`
+	Delta                ChatCompletionStreamChoiceDelta     `json:"delta"`
+	Logprobs             *ChatCompletionStreamChoiceLogprobs `json:"logprobs,omitempty"`
+	FinishReason         FinishReason                        `json:"finish_reason"`
+	ContentFilterResults ContentFilterResults                `json:"content_filter_results,omitempty"`
 }
 
 type PromptFilterResult struct {
@@ -60,6 +80,10 @@ func (c *Client) CreateChatCompletionStream(
 	}
 
 	request.Stream = true
+	if err = validateRequestForO1Models(request); err != nil {
+		return
+	}
+
 	req, err := c.newRequest(
 		ctx,
 		http.MethodPost,