openteam/server/internal/proxy/convert/convert_test.go

package convert

import (
	"encoding/json"
	"strings"
	"testing"
)

func mustJSON(t *testing.T, v any) string {
	t.Helper()
	b, err := json.Marshal(v)
	if err != nil {
		t.Fatalf("marshal: %v", err)
	}
	return string(b)
}

func TestChatToMessagesReq(t *testing.T) {
	in := `{
		"model":"claude-sonnet-5",
		"messages":[
			{"role":"system","content":"你是助手"},
			{"role":"user","content":"hi"},
			{"role":"assistant","content":"hello","tool_calls":[{"id":"call_1","function":{"name":"get_weather","arguments":"{\"city\":\"sz\"}"}}]},
			{"role":"tool","tool_call_id":"call_1","content":"sunny"}
		],
		"tools":[{"type":"function","function":{"name":"get_weather","description":"查天气","parameters":{"type":"object"}}}],
		"max_tokens":100,
		"stream":true
	}`
	out, err := ConvertRequest([]byte(in), ProtoChat, ProtoMessages)
	if err != nil {
		t.Fatalf("convert: %v", err)
	}
	var m map[string]any
	if err := json.Unmarshal(out, &m); err != nil {
		t.Fatalf("unmarshal out: %v\n%s", err, out)
	}
	if m["system"] != "你是助手" {
		t.Fatalf("system = %v", m["system"])
	}
	if m["max_tokens"] != float64(100) {
		t.Fatalf("max_tokens = %v", m["max_tokens"])
	}
	msgs := m["messages"].([]any)
	if len(msgs) != 3 {
		t.Fatalf("messages len = %d", len(msgs))
	}
	// assistant 含 tool_use 块
	assistant := msgs[1].(map[string]any)
	content := assistant["content"].([]any)
	foundToolUse := false
	for _, c := range content {
		cm := c.(map[string]any)
		if cm["type"] == "tool_use" {
			foundToolUse = true
			if cm["name"] != "get_weather" || cm["id"] != "call_1" {
				t.Fatalf("tool_use mismatch: %v", cm)
			}
		}
	}
	if !foundToolUse {
		t.Fatal("expected tool_use block")
	}
	// tool 消息 → user 消息的 tool_result 块
	tool := msgs[2].(map[string]any)
	if tool["role"] != "user" {
		t.Fatalf("tool message role = %v", tool["role"])
	}
}

func TestMessagesToChatReq(t *testing.T) {
	in := `{
		"model":"gpt-4o-mini",
		"system":"你是助手",
		"messages":[
			{"role":"user","content":"hi"},
			{"role":"assistant","content":[{"type":"text","text":"hello"},{"type":"tool_use","id":"call_1","name":"get_weather","input":{"city":"sz"}}]},
			{"role":"user","content":[{"type":"tool_result","tool_use_id":"call_1","content":"sunny"}]}
		],
		"tools":[{"name":"get_weather","description":"查天气","input_schema":{"type":"object"}}],
		"max_tokens":100,
		"stream":false
	}`
	out, err := ConvertRequest([]byte(in), ProtoMessages, ProtoChat)
	if err != nil {
		t.Fatalf("convert: %v", err)
	}
	var m map[string]any
	_ = json.Unmarshal(out, &m)
	msgs := m["messages"].([]any)
	// system + user + assistant + tool = 4 条
	if len(msgs) != 4 {
		t.Fatalf("messages len = %d: %s", len(msgs), out)
	}
	if msgs[0].(map[string]any)["role"] != "system" {
		t.Fatal("expected system message first")
	}
	assistant := msgs[2].(map[string]any)
	if tc := assistant["tool_calls"]; tc == nil {
		t.Fatalf("expected tool_calls in assistant: %s", out)
	}
	tool := msgs[3].(map[string]any)
	if tool["role"] != "tool" || tool["tool_call_id"] != "call_1" {
		t.Fatalf("tool message mismatch: %v", tool)
	}
}

func TestResponsesToChatReq(t *testing.T) {
	in := `{
		"model":"claude-sonnet-5",
		"instructions":"你是助手",
		"input":"hello",
		"tools":[{"type":"function","name":"get_weather","description":"查天气","parameters":{"type":"object"}}],
		"max_output_tokens":200,
		"stream":false
	}`
	out, err := ConvertRequest([]byte(in), ProtoResponses, ProtoChat)
	if err != nil {
		t.Fatalf("convert: %v", err)
	}
	var m map[string]any
	_ = json.Unmarshal(out, &m)
	msgs := m["messages"].([]any)
	if len(msgs) != 2 {
		t.Fatalf("messages len = %d: %s", len(msgs), out)
	}
	if msgs[0].(map[string]any)["role"] != "system" {
		t.Fatal("expected system from instructions")
	}
	if m["max_tokens"] != float64(200) {
		t.Fatalf("max_tokens = %v", m["max_tokens"])
	}
}

func TestChatToResponsesReq(t *testing.T) {
	in := mustJSON(t, map[string]any{
		"model": "gpt-4o",
		"messages": []any{
			map[string]any{"role": "system", "content": "sys"},
			map[string]any{"role": "user", "content": "hi"},
		},
		"max_tokens": 300,
	})
	out, err := ConvertRequest([]byte(in), ProtoChat, ProtoResponses)
	if err != nil {
		t.Fatalf("convert: %v", err)
	}
	var m map[string]any
	_ = json.Unmarshal(out, &m)
	if m["instructions"] != "sys" {
		t.Fatalf("instructions = %v", m["instructions"])
	}
	if m["max_output_tokens"] != float64(300) {
		t.Fatalf("max_output_tokens = %v", m["max_output_tokens"])
	}
}

func TestMessagesToChatResp(t *testing.T) {
	in := `{"id":"msg_abc","type":"message","role":"assistant","model":"claude-sonnet-5",
		"content":[{"type":"text","text":"你好"},{"type":"tool_use","id":"call_1","name":"get_weather","input":{"city":"sz"}}],
		"stop_reason":"tool_use","usage":{"input_tokens":10,"output_tokens":5}}`
	out, err := ConvertResponse([]byte(in), ProtoMessages, ProtoChat)
	if err != nil {
		t.Fatalf("convert: %v", err)
	}
	var m map[string]any
	_ = json.Unmarshal(out, &m)
	choices := m["choices"].([]any)
	msg := choices[0].(map[string]any)["message"].(map[string]any)
	if msg["content"] != "你好" {
		t.Fatalf("content = %v", msg["content"])
	}
	if msg["tool_calls"] == nil {
		t.Fatal("expected tool_calls")
	}
	if choices[0].(map[string]any)["finish_reason"] != "tool_calls" {
		t.Fatalf("finish_reason = %v", choices[0].(map[string]any)["finish_reason"])
	}
}

func TestChatToMessagesResp(t *testing.T) {
	in := `{"id":"chatcmpl-xyz","object":"chat.completion","model":"gpt-4o",
		"choices":[{"index":0,"message":{"role":"assistant","content":"hi"},"finish_reason":"stop"}],
		"usage":{"prompt_tokens":3,"completion_tokens":2,"total_tokens":5}}`
	out, err := ConvertResponse([]byte(in), ProtoChat, ProtoMessages)
	if err != nil {
		t.Fatalf("convert: %v", err)
	}
	var m map[string]any
	_ = json.Unmarshal(out, &m)
	if m["stop_reason"] != "end_turn" {
		t.Fatalf("stop_reason = %v", m["stop_reason"])
	}
	content := m["content"].([]any)
	if content[0].(map[string]any)["text"] != "hi" {
		t.Fatalf("content = %v", content)
	}
	usage := m["usage"].(map[string]any)
	if usage["input_tokens"] != float64(3) || usage["output_tokens"] != float64(2) {
		t.Fatalf("usage = %v", usage)
	}
}

// ---------------------------------------------------------------------------
// 流式转换

func feedLines(t *testing.T, transformer func([]byte) []byte, lines []string) string {
	t.Helper()
	var sb strings.Builder
	for _, l := range lines {
		if out := transformer([]byte(l)); out != nil {
			sb.Write(out)
		}
	}
	return sb.String()
}

func TestStreamMessagesToChat(t *testing.T) {
	tf := newMessagesToChat().line
	out := feedLines(t, tf, []string{
		"event: message_start\n",
		`data: {"type":"message_start","message":{"id":"msg_1","model":"claude-sonnet-5"}}` + "\n\n",
		"event: content_block_delta\n",
		`data: {"type":"content_block_delta","index":0,"delta":{"type":"text_delta","text":"你好"}}` + "\n\n",
		"event: message_delta\n",
		`data: {"type":"message_delta","delta":{"stop_reason":"end_turn"},"usage":{"input_tokens":10,"output_tokens":5}}` + "\n\n",
		"event: message_stop\n",
		`data: {"type":"message_stop"}` + "\n\n",
	})
	if !strings.Contains(out, `"content":"你好"`) {
		t.Fatalf("missing content chunk: %s", out)
	}
	if !strings.Contains(out, `"finish_reason":"stop"`) {
		t.Fatalf("missing finish chunk: %s", out)
	}
	if !strings.Contains(out, `"usage"`) {
		t.Fatalf("missing usage chunk: %s", out)
	}
	if !strings.Contains(out, "data: [DONE]") {
		t.Fatalf("missing [DONE]: %s", out)
	}
}

func TestStreamChatToMessages(t *testing.T) {
	tf := newChatToMessages().line
	out := feedLines(t, tf, []string{
		`data: {"id":"chatcmpl-1","object":"chat.completion.chunk","model":"gpt-4o","choices":[{"index":0,"delta":{"role":"assistant"},"finish_reason":null}]}` + "\n\n",
		`data: {"id":"chatcmpl-1","object":"chat.completion.chunk","model":"gpt-4o","choices":[{"index":0,"delta":{"content":"你好"},"finish_reason":null}]}` + "\n\n",
		`data: {"id":"chatcmpl-1","object":"chat.completion.chunk","model":"gpt-4o","choices":[{"index":0,"delta":{},"finish_reason":"stop"}]}` + "\n\n",
		`data: {"id":"chatcmpl-1","object":"chat.completion.chunk","model":"gpt-4o","choices":[],"usage":{"prompt_tokens":12,"completion_tokens":9}}` + "\n\n",
		"data: [DONE]\n\n",
	})
	if !strings.Contains(out, "event: message_start") {
		t.Fatalf("missing message_start: %s", out)
	}
	if !strings.Contains(out, `"text":"你好"`) || !strings.Contains(out, `"type":"text_delta"`) {
		t.Fatalf("missing content delta: %s", out)
	}
	if !strings.Contains(out, `"stop_reason":"end_turn"`) {
		t.Fatalf("missing message_delta: %s", out)
	}
	if !strings.Contains(out, "event: message_stop") {
		t.Fatalf("missing message_stop: %s", out)
	}
}

func TestStreamResponsesToMessages(t *testing.T) {
	tf := newResponsesToMessages().line
	out := feedLines(t, tf, []string{
		"event: response.created\n",
		`data: {"type":"response.created","response":{"id":"resp_1","model":"claude-sonnet-5"}}` + "\n\n",
		"event: response.output_text.delta\n",
		`data: {"type":"response.output_text.delta","delta":"hi"}` + "\n\n",
		"event: response.completed\n",
		`data: {"type":"response.completed","response":{"id":"resp_1","status":"completed","usage":{"input_tokens":7,"output_tokens":8}}}` + "\n\n",
	})
	if !strings.Contains(out, "event: message_start") {
		t.Fatalf("missing message_start: %s", out)
	}
	if !strings.Contains(out, `"text":"hi"`) {
		t.Fatalf("missing content: %s", out)
	}
	if !strings.Contains(out, "event: message_stop") {
		t.Fatalf("missing message_stop: %s", out)
	}
}

func TestStreamMessagesToResponses(t *testing.T) {
	tf := newMessagesToResponses().line
	out := feedLines(t, tf, []string{
		"event: message_start\n",
		`data: {"type":"message_start","message":{"id":"msg_1","model":"claude-sonnet-5"}}` + "\n\n",
		"event: content_block_delta\n",
		`data: {"type":"content_block_delta","index":0,"delta":{"type":"text_delta","text":"hi"}}` + "\n\n",
		"event: message_stop\n",
		`data: {"type":"message_stop"}` + "\n\n",
	})
	if !strings.Contains(out, "event: response.created") {
		t.Fatalf("missing response.created: %s", out)
	}
	if !strings.Contains(out, "event: response.output_text.delta") {
		t.Fatalf("missing output_text.delta: %s", out)
	}
	if !strings.Contains(out, "event: response.completed") {
		t.Fatalf("missing response.completed: %s", out)
	}
}

// ---------------------------------------------------------------------------
// 图片内容块转换（messages image / chat image_url / responses input_image 互转不丢图）

const testB64 = "iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAQAAAC1HAwCAAAAC0lEQVR42mNkYAAAAAYAAjCB0C8AAAAASUVORK5CYII="

// userContentBlocks 断言 Chat 用户消息 content 为块数组并返回。
func userContentBlocks(t *testing.T, out []byte) []any {
	t.Helper()
	var m map[string]any
	if err := json.Unmarshal(out, &m); err != nil {
		t.Fatalf("unmarshal: %v\n%s", err, out)
	}
	msgs := m["messages"].([]any)
	if len(msgs) != 1 {
		t.Fatalf("messages len = %d: %s", len(msgs), out)
	}
	content, ok := msgs[0].(map[string]any)["content"].([]any)
	if !ok {
		t.Fatalf("content not array: %s", out)
	}
	return content
}

// blockURL 取 image_url 块的 url。
func blockURL(t *testing.T, b any) string {
	t.Helper()
	bm := b.(map[string]any)
	if bm["type"] != "image_url" {
		t.Fatalf("block type = %v", bm["type"])
	}
	iu := bm["image_url"].(map[string]any)
	url, _ := iu["url"].(string)
	return url
}

func TestMessagesToChatReqImage(t *testing.T) {
	in := mustJSON(t, map[string]any{
		"model": "gpt-4o-mini",
		"messages": []any{map[string]any{"role": "user", "content": []any{
			map[string]any{"type": "text", "text": "describe"},
			map[string]any{"type": "image", "source": map[string]any{"type": "base64", "media_type": "image/png", "data": testB64}},
			map[string]any{"type": "image", "source": map[string]any{"type": "url", "url": "https://example.com/a.png"}},
		}}},
	})
	out, err := ConvertRequest([]byte(in), ProtoMessages, ProtoChat)
	if err != nil {
		t.Fatalf("convert: %v", err)
	}
	blocks := userContentBlocks(t, out)
	if len(blocks) != 3 {
		t.Fatalf("blocks len = %d: %s", len(blocks), out)
	}
	if got := blockURL(t, blocks[1]); got != "data:image/png;base64,"+testB64 {
		t.Fatalf("base64 image url = %q", got)
	}
	if got := blockURL(t, blocks[2]); got != "https://example.com/a.png" {
		t.Fatalf("url image = %q", got)
	}
}

// 多图 + 图/文交错顺序：转换后块数量与顺序保持不变（文档推荐的 Image 1:/Image 2: 标签模式）。
func TestMessagesToChatReqMultiImage(t *testing.T) {
	img := func(url string) any {
		return map[string]any{"type": "image", "source": map[string]any{"type": "url", "url": url}}
	}
	in := mustJSON(t, map[string]any{
		"model": "gpt-4o-mini",
		"messages": []any{map[string]any{"role": "user", "content": []any{
			map[string]any{"type": "text", "text": "Image 1:"},
			img("https://example.com/1.png"),
			map[string]any{"type": "text", "text": "Image 2:"},
			img("https://example.com/2.png"),
			map[string]any{"type": "text", "text": "How are these images different?"},
		}}},
	})
	out, err := ConvertRequest([]byte(in), ProtoMessages, ProtoChat)
	if err != nil {
		t.Fatalf("convert: %v", err)
	}
	blocks := userContentBlocks(t, out)
	if len(blocks) != 5 {
		t.Fatalf("blocks len = %d: %s", len(blocks), out)
	}
	types := []string{}
	for _, b := range blocks {
		types = append(types, b.(map[string]any)["type"].(string))
	}
	want := []string{"text", "image_url", "text", "image_url", "text"}
	for i := range want {
		if types[i] != want[i] {
			t.Fatalf("block[%d] type = %s, want %s: %v", i, types[i], want[i], types)
		}
	}
	// 第二张图仍是第二张
	if got := blockURL(t, blocks[3]); got != "https://example.com/2.png" {
		t.Fatalf("image2 url = %q", got)
	}
}

func TestChatToMessagesReqImage(t *testing.T) {
	in := mustJSON(t, map[string]any{
		"model": "claude-sonnet-5",
		"messages": []any{map[string]any{"role": "user", "content": []any{
			map[string]any{"type": "text", "text": "describe"},
			map[string]any{"type": "image_url", "image_url": map[string]any{"url": "data:image/jpeg;base64," + testB64}},
			map[string]any{"type": "image_url", "image_url": map[string]any{"url": "https://example.com/b.png"}},
		}}},
	})
	out, err := ConvertRequest([]byte(in), ProtoChat, ProtoMessages)
	if err != nil {
		t.Fatalf("convert: %v", err)
	}
	var m map[string]any
	_ = json.Unmarshal(out, &m)
	msgs := m["messages"].([]any)
	content := msgs[0].(map[string]any)["content"].([]any)
	if len(content) != 3 {
		t.Fatalf("content len = %d: %s", len(content), out)
	}
	src1 := content[1].(map[string]any)["source"].(map[string]any)
	if src1["type"] != "base64" || src1["media_type"] != "image/jpeg" || src1["data"] != testB64 {
		t.Fatalf("base64 source = %v", src1)
	}
	src2 := content[2].(map[string]any)["source"].(map[string]any)
	if src2["type"] != "url" || src2["url"] != "https://example.com/b.png" {
		t.Fatalf("url source = %v", src2)
	}
}

func TestResponsesToChatReqImage(t *testing.T) {
	in := mustJSON(t, map[string]any{
		"model": "claude-sonnet-5",
		"input": []any{map[string]any{"role": "user", "content": []any{
			map[string]any{"type": "input_text", "text": "describe"},
			map[string]any{"type": "input_image", "image_url": "data:image/png;base64," + testB64},
		}}},
	})
	out, err := ConvertRequest([]byte(in), ProtoResponses, ProtoChat)
	if err != nil {
		t.Fatalf("convert: %v", err)
	}
	blocks := userContentBlocks(t, out)
	if len(blocks) != 2 {
		t.Fatalf("blocks len = %d: %s", len(blocks), out)
	}
	if got := blockURL(t, blocks[1]); got != "data:image/png;base64,"+testB64 {
		t.Fatalf("image url = %q", got)
	}
}

func TestChatToResponsesReqImage(t *testing.T) {
	in := mustJSON(t, map[string]any{
		"model": "gpt-4o",
		"messages": []any{map[string]any{"role": "user", "content": []any{
			map[string]any{"type": "text", "text": "describe"},
			map[string]any{"type": "image_url", "image_url": map[string]any{"url": "https://example.com/c.png"}},
		}}},
	})
	out, err := ConvertRequest([]byte(in), ProtoChat, ProtoResponses)
	if err != nil {
		t.Fatalf("convert: %v", err)
	}
	var m map[string]any
	_ = json.Unmarshal(out, &m)
	var input []map[string]any
	// 单条 user 消息项
	switch v := m["input"].(type) {
	case map[string]any:
		input = []map[string]any{v}
	case []any:
		for _, it := range v {
			input = append(input, it.(map[string]any))
		}
	}
	content := input[0]["content"].([]any)
	if len(content) != 2 {
		t.Fatalf("content len = %d: %s", len(content), out)
	}
	img := content[1].(map[string]any)
	if img["type"] != "input_image" {
		t.Fatalf("block type = %v", img["type"])
	}
	if img["image_url"] != "https://example.com/c.png" {
		t.Fatalf("image_url = %v", img["image_url"])
	}
}

// 链式转换：responses→messages、messages→responses（经 chat 中转）也不丢图。
func TestResponsesToMessagesReqImage(t *testing.T) {
	in := mustJSON(t, map[string]any{
		"model": "claude-sonnet-5",
		"input": []any{map[string]any{"role": "user", "content": []any{
			map[string]any{"type": "input_text", "text": "describe"},
			map[string]any{"type": "input_image", "image_url": "data:image/png;base64," + testB64},
		}}},
	})
	out, err := ConvertRequest([]byte(in), ProtoResponses, ProtoMessages)
	if err != nil {
		t.Fatalf("convert: %v", err)
	}
	var m map[string]any
	_ = json.Unmarshal(out, &m)
	msgs := m["messages"].([]any)
	content := msgs[0].(map[string]any)["content"].([]any)
	if len(content) != 2 {
		t.Fatalf("content len = %d: %s", len(content), out)
	}
	src := content[1].(map[string]any)["source"].(map[string]any)
	if src["type"] != "base64" || src["data"] != testB64 {
		t.Fatalf("source = %v", src)
	}
}

func TestMessagesToResponsesReqImage(t *testing.T) {
	in := mustJSON(t, map[string]any{
		"model": "gpt-4o",
		"messages": []any{map[string]any{"role": "user", "content": []any{
			map[string]any{"type": "image", "source": map[string]any{"type": "url", "url": "https://example.com/d.png"}},
			map[string]any{"type": "text", "text": "describe"},
		}}},
	})
	out, err := ConvertRequest([]byte(in), ProtoMessages, ProtoResponses)
	if err != nil {
		t.Fatalf("convert: %v", err)
	}
	var m map[string]any
	_ = json.Unmarshal(out, &m)
	var content []any
	switch v := m["input"].(type) {
	case map[string]any:
		content = v["content"].([]any)
	case []any:
		content = v[0].(map[string]any)["content"].([]any)
	}
	if len(content) != 2 {
		t.Fatalf("content len = %d: %s", len(content), out)
	}
	img := content[0].(map[string]any)
	if img["type"] != "input_image" || img["image_url"] != "https://example.com/d.png" {
		t.Fatalf("input_image = %v", img)
	}
}