mirror of
https://gitcode.com/JianFeeeee/ModelRouter.git
synced 2026-11-12 12:17:23 +00:00
feat(types): 统一格式落地为内核实际使用的类型 —— 私有字段泄漏归零
Phase 2 收尾（入向）。ChatRequest 从"客户端发什么就是什么"改为
UnifiedRequest（最大集合）的别名，内核真正用上了那个超集类型。
改掉的具体缺陷：
1. 私有字段泄漏归零（v1 的核心问题）
v1 的 disable_thinking 是内核私有开关，却被序列化进交给适配器的 body，
迫使 10 个适配器各自写 req.disable_thinking = nil 来清理。
现在它是中立的 thinking 字段（{enabled, budget_tokens}），客户端与
适配器都认这个词：
- gateway/chat.go：thinkingFromClient() 把 disable_thinking 翻译成
thinking.enabled=false；冲突时中立字段优先（它能带预算，布尔不能）
- deepseek.lua：thinking.enabled=false → extra_body.thinking.type=disabled
- anthropic.lua：thinking.enabled=true + budget → 原生 {type,budget_tokens}
（enabled=false 时不传该字段——Anthropic 没有 type=disabled 这个写法）
- 其余 7 个适配器：把"清理内核私有字段"改成"清理中立字段"
2. Tools/ToolChoice 改 json.RawMessage
原为 []interface{}（OpenAI 原始形状 = 内核在假设某个协议）。
内核从不解释工具 schema——它原样交给适配器，所以原始形状不损失任何
内核侧统一性。ToolDef 结构化类型保留在包内供新适配器选用。
3. 删除 MarshalJSON 后门
v1 手工 splice ExtraBody 进 OpenAI 形状的 body；现在 ExtraBody 是
UnifiedRequest 上一个正正经经声明的字段。
渐进迁移：ChatRequest 保留为别名（不是新类型），49 处引用零改动编译，
语义变更不被机械改名淹没。
测试：新增 thinking_test.go（翻译链 4 例 + 泄漏回归 + tools 透传）；
vm_test.go 两处迁移到中立语义。golden 字节级未漂移。
This commit is contained in:
JianFeeeee
@ -53,9 +53,20 @@ type UnifiedRequest struct {
|
||||
StopSequences []string `json:"stop_sequences,omitempty"` // Anthropic
|
||||
|
||||
// ---- 工具调用 ----
|
||||
Tools []ToolDef `json:"tools,omitempty"`
|
||||
ToolChoice any `json:"tool_choice,omitempty"`
|
||||
ParallelToolCalls *bool `json:"parallel_tool_calls,omitempty"` // OpenAI
|
||||
// Tools 保留 json.RawMessage（宽松的原始形状）而不是结构化的 []ToolDef，
|
||||
// 原因是**渐进迁移**：内核现有的 12 个适配器与 gateway 都按 []interface{}
|
||||
// /json.RawMessage 形状序列化请求到适配器。若此处改成 []ToolDef，则
|
||||
// 所有适配器与调用点立即编译失败，必须同步改写 19 个非测试点 + 30 个测试。
|
||||
// 而 Tools 的真实消费者是适配器（transform_request），内核本身从不解析它，
|
||||
// 所以用原始形状不损失任何内核侧的统一性——它本就是"原样交给适配器的载荷"。
|
||||
//
|
||||
// ToolDef 类型保留在此包内供新适配器选用（maximal-set 的结构化语义），
|
||||
// 但内核序列化路径暂不强推它。
|
||||
Tools json.RawMessage `json:"tools,omitempty"`
|
||||
ToolChoice json.RawMessage `json:"tool_choice,omitempty"`
|
||||
|
||||
// ParallelToolCalls *bool 允许显式 false（OpenAI 语义）。
|
||||
ParallelToolCalls *bool `json:"parallel_tool_calls,omitempty"` // OpenAI
|
||||
|
||||
// ---- 结构化输出 ----
|
||||
ResponseFormat *ResponseFormat `json:"response_format,omitempty"` // OpenAI
|
||||
@ -69,6 +80,9 @@ type UnifiedRequest struct {
|
||||
System string `json:"system,omitempty"`
|
||||
|
||||
// ---- 推理/思考 ----
|
||||
// Thinking 取代 v1 的内核私有字段 DisableThinking：那个字段曾被序列化进
|
||||
// 交给适配器的 body，迫使 10 个适配器各自写 req.disable_thinking = nil。
|
||||
// 这里它是双方都认可的对称语义（客户端要什么 / 上游怎么映射）。
|
||||
// Thinking 是"要不要思考、思考预算多少"的语义；
|
||||
// 适配器各自映射（Anthropic thinking.budget_tokens、Gemini
|
||||
// thinkingConfig.thinkingBudget、DeepSeek extra_body.thinking.type）。
|
||||
@ -173,3 +187,14 @@ type UnifiedDelta struct {
|
||||
// Cost 上游自报费用。
|
||||
Cost string `json:"cost,omitempty"`
|
||||
}
|
||||
|
||||
// ChatRequest is the historical name for the in-bound request type.
|
||||
//
|
||||
// Deprecated: it is an alias of UnifiedRequest. The rename happened because the
|
||||
// kernel's internal format stopped being "whatever the client sent" once the
|
||||
// maximal set landed — it is now the maximal set, and the client's own protocol
|
||||
// shape only survives until the adapter sees it.
|
||||
//
|
||||
// Kept as an alias (not a distinct type) so the 49 existing references keep
|
||||
// compiling while call sites migrate deliberately rather than in one sweep.
|
||||
type ChatRequest = UnifiedRequest
|
||||
|
||||
@ -34,8 +34,8 @@ func TestMaximalSetCoversEveryProtocol(t *testing.T) {
|
||||
Seed: i64(42),
|
||||
Stop: []string{"\n"},
|
||||
StopSequences: []string{"\n\n"},
|
||||
Tools: []ToolDef{{Name: "f", Type: "function"}},
|
||||
ToolChoice: "auto",
|
||||
Tools: json.RawMessage(`[{"type":"function","function":{"name":"f"}}]`),
|
||||
ToolChoice: json.RawMessage(`"auto"`),
|
||||
ParallelToolCalls: b(true),
|
||||
ResponseFormat: &ResponseFormat{Type: "json_object"},
|
||||
System: "you are helpful",
|
||||
|
||||
@ -17,42 +17,32 @@ import (
|
||||
// provider package (which pulls in the Lua runtime).
|
||||
var ErrBusy = errors.New("provider busy")
|
||||
|
||||
// ---- OpenAI wire request (gateway input) ----
|
||||
|
||||
type ChatRequest struct {
|
||||
Model string `json:"model"`
|
||||
Messages []ChatMessage `json:"messages"`
|
||||
Temperature *float64 `json:"temperature,omitempty"`
|
||||
MaxTokens int `json:"max_tokens,omitempty"`
|
||||
Stream bool `json:"stream,omitempty"`
|
||||
Tools []interface{} `json:"tools,omitempty"`
|
||||
ToolChoice interface{} `json:"tool_choice,omitempty"`
|
||||
DisableThinking bool `json:"disable_thinking"`
|
||||
ExtraBody map[string]interface{} `json:"-"`
|
||||
|
||||
// ClientSession 是客户端自带的会话标识（请求头 x-session-affinity 等）。
|
||||
// 不参与序列化：它只用于让适配器生成稳定的上游会话，不应泄漏给上游 body。
|
||||
ClientSession string `json:"-"`
|
||||
}
|
||||
|
||||
func (r *ChatRequest) MarshalJSON() ([]byte, error) {
|
||||
type Alias ChatRequest
|
||||
data, err := json.Marshal((*Alias)(r))
|
||||
if err != nil {
|
||||
return nil, err
|
||||
}
|
||||
if len(r.ExtraBody) == 0 {
|
||||
return data, nil
|
||||
}
|
||||
var raw map[string]interface{}
|
||||
if err := json.Unmarshal(data, &raw); err != nil {
|
||||
return nil, err
|
||||
}
|
||||
for k, v := range r.ExtraBody {
|
||||
raw[k] = v
|
||||
}
|
||||
return json.Marshal(raw)
|
||||
}
|
||||
// ---- in-bound request ----
|
||||
//
|
||||
// ChatRequest now lives in maxwire.go as an alias of UnifiedRequest: the
|
||||
// kernel's internal format is the maximal set (OpenAI / Anthropic / Gemini /
|
||||
// Ollama / DeepSeek fields unioned), not "whatever an OpenAI client sent". The
|
||||
// type stays named ChatRequest at every call site for now — renaming 19
|
||||
// non-test + 30 test references in one sweep would bury the semantic change in
|
||||
// mechanical noise, and an alias makes the migration free when it happens.
|
||||
//
|
||||
// The v1 struct that lived here is gone. Its fields and their fate:
|
||||
//
|
||||
// Model/Messages/Temperature/MaxTokens/Stream → UnifiedRequest, same tags
|
||||
// Tools/ToolChoice → json.RawMessage (raw passthrough
|
||||
// to the adapter; the kernel never
|
||||
// parses them)
|
||||
// DisableThinking (json:"disable_thinking") → REPLACED by Thinking. This was
|
||||
// the private field that leaked into
|
||||
// the body handed to adapters, forcing
|
||||
// 10 adapters to nil it out.
|
||||
// ExtraBody (hand-injected by MarshalJSON) → a real, declared field on
|
||||
// UnifiedRequest — no more
|
||||
// MarshalJSON back door
|
||||
// ClientSession (json:"-") → kept, still never serialized
|
||||
//
|
||||
// The custom MarshalJSON is gone with it: its only job was splicing ExtraBody
|
||||
// into an OpenAI-shaped body, which is no longer how the type serializes.
|
||||
|
||||
// ChatMessage supports both plain string content and multimodal arrays
|
||||
// (RawMessage preserves whatever the client sent for the adapter to process).
|
||||
|
||||
Reference in New Issue
Block a user