Build provider requests directly as payloads

docs/roadmap/implementation.md

@@ -97,6 +97,11 @@ where fixture migration remains. No production construction supplies the removed
 
 ## Stage 2: Encode One Root Wire Representation
 
+**Completed:** Requests now build the provider payload directly, with structured
+dispatch evidence derived from that payload. Go 1.27.3 `go test ./internal/llm
+./internal/generation` passed, including mapping, presence, payload freezing,
+session, structured-output, and encoding-failure tests.
+
 **Finding:** 2. **Files:** `internal/llm/openai_compatible_client.go` and request-mapping tests.
 
 1. Replace the root `domain request -> openAIChatRequest -> map` sequence with

internal/llm/openai_compatible_client.go

@@ -125,12 +125,7 @@ func (c *OpenAICompatibleClient) PrepareGeneration(req domain.GenerateRequest) (
 		return nil, fmt.Errorf("%w: invalid endpoint path: %v", ErrInvalidRequest, err)
 	}
 
-	wireReq, err := openAIChatRequestFromGenerateRequest(req)
-	if err != nil {
-		return nil, fmt.Errorf("%w: %v", ErrInvalidRequest, err)
-	}
-
-	wirePayload, err := openAIChatRequestPayload(wireReq)
+	wirePayload, err := openAIChatRequestPayload(req)
 	if err != nil {
 		return nil, fmt.Errorf("%w: %v", ErrInvalidRequest, err)
 	}
@@ -159,7 +154,7 @@ func (c *OpenAICompatibleClient) PrepareGeneration(req domain.GenerateRequest) (
 	return &PreparedGeneration{
 		endpoint: endpoint, payload: payload, apiKey: apiKey,
 		timeoutSeconds: req.Target.TimeoutSeconds, httpClient: httpClient,
-		structuredOutputRequested: wireReq.ResponseFormat != nil,
+		structuredOutputRequested: wirePayload["response_format"] != nil,
 	}, nil
 }
 
@@ -425,96 +420,52 @@ func openAIChatResponseTooLargeError() error {
 	)
 }
 
-func openAIChatRequestFromGenerateRequest(req domain.GenerateRequest) (openAIChatRequest, error) {
+func openAIChatRequestPayload(req domain.GenerateRequest) (map[string]any, error) {
 	model := strings.TrimSpace(req.Target.Model)
 	if model == "" {
-		return openAIChatRequest{}, errors.New("model is required")
+		return nil, errors.New("model is required")
 	}
 
-	wireReq := openAIChatRequest{
-		Model: model,
-	}
 	sessionID, err := domain.NormalizeSessionID(req.Prompt.SessionID)
 	if err != nil {
-		return openAIChatRequest{}, err
+		return nil, err
 	}
-	wireReq.SessionID = sessionID
 
-	wireReq.Messages = make([]openAIChatRequestMessage, 0, len(req.Prompt.Messages))
+	messages := make([]openAIChatRequestMessage, 0, len(req.Prompt.Messages))
 	for _, msg := range req.Prompt.Messages {
-		wireReq.Messages = append(wireReq.Messages, openAIChatRequestMessageFromRenderedMessage(msg))
+		messages = append(messages, openAIChatRequestMessageFromRenderedMessage(msg))
 	}
 
+	out := map[string]any{"model": model, "messages": messages}
+	if sessionID != "" {
+		out["session_id"] = sessionID
+	}
 	if req.Target.Temperature != 0 || req.TargetPresence.Temperature {
-		wireReq.Temperature = &req.Target.Temperature
+		out["temperature"] = req.Target.Temperature
 	}
 	if req.Target.MaxTokens != 0 || req.TargetPresence.MaxTokens {
-		wireReq.MaxTokens = &req.Target.MaxTokens
+		out["max_tokens"] = req.Target.MaxTokens
 	}
 	if req.Target.TopP != 0 || req.TargetPresence.TopP {
-		wireReq.TopP = &req.Target.TopP
+		out["top_p"] = req.Target.TopP
 	}
 	if strings.TrimSpace(req.Target.ServiceTier) != "" {
-		wireReq.ServiceTier = req.Target.ServiceTier
+		out["service_tier"] = req.Target.ServiceTier
 	}
 	if strings.TrimSpace(req.Target.ReasoningEffort) != "" {
-		wireReq.ReasoningEffort = req.Target.ReasoningEffort
-	}
-	if len(req.Target.ExtraParams) > 0 {
-		wireReq.ExtraParams = req.Target.ExtraParams
+		out["reasoning_effort"] = req.Target.ReasoningEffort
 	}
 	if req.StructuredOutput != nil {
 		responseFormat, err := toOpenAIResponseFormat(req.StructuredOutput)
 		if err != nil {
-			return openAIChatRequest{}, err
+			return nil, err
+		}
+		if responseFormat != nil {
+			out["response_format"] = responseFormat
 		}
-		wireReq.ResponseFormat = responseFormat
 	}
 
-	return wireReq, nil
-}
-
-type openAIChatRequest struct {
-	Model           string                     `json:"model"`
-	SessionID       string                     `json:"session_id,omitempty"`
-	Messages        []openAIChatRequestMessage `json:"messages"`
-	Temperature     *float64                   `json:"temperature,omitempty"`
-	MaxTokens       *int                       `json:"max_tokens,omitempty"`
-	TopP            *float64                   `json:"top_p,omitempty"`
-	ServiceTier     string                     `json:"service_tier,omitempty"`
-	ReasoningEffort string                     `json:"reasoning_effort,omitempty"`
-	ResponseFormat  *openAIResponseFormat      `json:"response_format,omitempty"`
-	ExtraParams     map[string]any             `json:"-"`
-}
-
-func openAIChatRequestPayload(req openAIChatRequest) (map[string]any, error) {
-	out := map[string]any{
-		"model":    req.Model,
-		"messages": req.Messages,
-	}
-	if req.SessionID != "" {
-		out["session_id"] = req.SessionID
-	}
-	if req.Temperature != nil {
-		out["temperature"] = *req.Temperature
-	}
-	if req.MaxTokens != nil {
-		out["max_tokens"] = *req.MaxTokens
-	}
-	if req.TopP != nil {
-		out["top_p"] = *req.TopP
-	}
-	if req.ServiceTier != "" {
-		out["service_tier"] = req.ServiceTier
-	}
-	if req.ReasoningEffort != "" {
-		out["reasoning_effort"] = req.ReasoningEffort
-	}
-	if req.ResponseFormat != nil {
-		out["response_format"] = req.ResponseFormat
-	}
-
-	for key, value := range req.ExtraParams {
+	for key, value := range req.Target.ExtraParams {
 		if key == "" {
 			return nil, errors.New("extra_params key must not be empty")
 		}