iop/apps/edge/internal/openai/chat_completion.go

package openai

import (
	"context"
	"go.uber.org/zap"
	edgeservice "iop/apps/edge/internal/service"
	"net/http"
	"strings"
	"time"
)

func shouldSynthesizeTextToolCalls(dispatch edgeservice.RunDispatch, textToolFallback bool) bool {
	return textToolFallback || strings.TrimSpace(dispatch.Adapter) == "cli"
}

func chatSubmitRunRequest(dispatch routeDispatch, req chatCompletionRequest, workspace, prompt string, input map[string]any, metadata map[string]string) edgeservice.SubmitRunRequest {
	return edgeservice.SubmitRunRequest{
		NodeRef:          dispatch.NodeRef,
		ModelGroupKey:    strings.TrimSpace(req.Model),
		ProviderID:       dispatch.ProviderID,
		UsageAttribution: dispatch.UsageAttribution,
		Adapter:          dispatch.Adapter,
		Target:           dispatch.Target,
		SessionID:        dispatch.SessionID,
		Workspace:        workspace,
		Prompt:           prompt,
		Input:            input,
		TimeoutSec:       dispatch.TimeoutSec,
		MaxQueue:         dispatch.MaxQueue,
		QueueTimeoutMS:   dispatch.QueueTimeoutMS,
		Metadata:         metadata,
		ProviderPool:     dispatch.ProviderPool,
	}
}

// completeChatCompletion serves a non-streaming chat completion. When the
// stream evidence gate runtime is enabled, the Core request runtime is the
// single owner of hold/validate/rebuild/re-admission and this surface only
// supplies the buffered event source and the JSON renderer. The legacy
// retry loop below stays reachable exclusively through the runtime-disabled
// compatibility branch.
func (s *Server) completeChatCompletion(w http.ResponseWriter, dc *chatDispatchContext, handle edgeservice.RunResult) {
	if s.streamGateEnabled() {
		s.runOpenAIBufferedChatStreamGate(w, nil, dc, handle, false)
		return
	}
	s.completeChatCompletionLegacy(w, dc, handle)
}

// completeChatCompletionLegacy is the runtime-disabled compatibility path. It
// is the only remaining caller of dc.retrySubmit on the non-stream surface and
// is reachable exclusively from the !streamGateEnabled() branch of
// completeChatCompletion.
func (s *Server) completeChatCompletionLegacy(w http.ResponseWriter, dc *chatDispatchContext, handle edgeservice.RunResult) {
	r := dc.r
	req := dc.req
	outputPolicy := dc.outputPolicy
	attempt := 1
	for {
		result, err := collectChatCompletionOutput(r.Context(), req, handle, outputPolicy)
		if err != nil {
			s.cancelRunOnHTTPGiveUp(handle.Dispatch(), err)
			handle.Close()
			dc.finishUsageRequest(usageStatusForError(err), responseModeNormalized)
			writeError(w, httpStatusForRunError(err), "run_error", err.Error())
			return
		}
		dc.recordUsageAttempt(handle.Dispatch(), result.responseMode, usageObservationFromOpenAIUsage(result.usage, result.reasoningLen))
		valErr := result.toolValidationErr
		if valErr == nil {
			valErr = validateToolCallResponse(dc.validation, result.toolCalls, result.toolCallOrigin)
		}
		if valErr != nil {
			failedRunID := handle.Dispatch().RunID
			if attempt < maxToolValidationAttempts {
				handle.Close()
				attempt++
				retryReq := dc.submitReq
				retryReq.Metadata = toolValidationAttemptMetadata(dc.submitReq.Metadata, attempt, failedRunID, valErr.Error())
				s.logger.Warn("openai chat completion tool validation retry",
					zap.String("run_id", failedRunID),
					zap.Int("attempt", attempt),
					zap.String("reason", valErr.Error()),
				)
				next, submitErr := dc.retrySubmit(r.Context(), retryReq)
				if submitErr != nil {
					dc.finishUsageRequest(usageStatusError, responseModeNormalized)
					writeError(w, http.StatusBadGateway, "tool_validation_retry_error", submitErr.Error())
					return
				}
				// Extract RunResult from either a direct RunResult or a
				// ProviderPoolDispatchResult (pool retry). For provider-pool
				// retries, only normalized results with a non-nil Run are
				// accepted; tunnel or nil Run results are rejected to prevent
				// nil handle panic downstream.
				if rr, ok := next.(edgeservice.RunResult); ok {
					handle = rr
				} else if ppd, ok := next.(*edgeservice.ProviderPoolDispatchResult); ok {
					if ppd.Path != edgeservice.ProviderPoolPathNormalized || ppd.Run == nil {
						if ppd.Tunnel != nil {
							ppd.Tunnel.Close()
						}
						dc.finishUsageRequest(usageStatusError, responseModeNormalized)
						writeError(w, http.StatusBadGateway, "tool_validation_retry_error", "provider-pool retry selected tunnel path")
						return
					}
					handle = ppd.Run
				} else {
					handle.Close()
					dc.finishUsageRequest(usageStatusError, responseModeNormalized)
					writeError(w, http.StatusInternalServerError, "run_error", "unexpected retry result type")
					return
				}
				continue
			}
			handle.Close()
			s.logger.Warn("openai chat completion tool validation failed",
				zap.String("run_id", failedRunID),
				zap.Int("attempt", attempt),
				zap.String("reason", valErr.Error()),
			)
			dc.finishUsageRequest(usageStatusError, responseModeNormalized)
			writeError(w, http.StatusBadGateway, "tool_validation_error", valErr.Error())
			return
		}
		handle.Close()
		dc.finishUsageRequest(usageStatusSuccess, result.responseMode)
		s.logChatCompletionOutput(handle.Dispatch(), result, outputPolicy)
		created := time.Now().Unix()
		writeJSON(w, http.StatusOK, chatCompletionResponse{
			ID:      "chatcmpl-" + handle.Dispatch().RunID,
			Object:  "chat.completion",
			Created: created,
			Model:   responseModel(req.Model, handle.Dispatch().Target),
			Choices: []chatCompletionChoice{{
				Index:        0,
				Message:      result.message,
				FinishReason: result.finishReason,
			}},
			Usage: result.usage,
		})
		return
	}
}

// logChatCompletionOutput emits the assembled non-stream completion log line.
// The legacy loop and the runtime-enabled JSON release sink share it so both
// surfaces report the same assembled content, reasoning, and tool calls.
func (s *Server) logChatCompletionOutput(dispatch edgeservice.RunDispatch, result chatCompletionOutput, outputPolicy strictOutputPolicy) {
	toolCallNames := extractToolCallNames(result.toolCalls)
	s.logger.Info("openai chat completion output",
		zap.String("run_id", dispatch.RunID),
		zap.Bool("strict_output", outputPolicy.Strict),
		zap.Bool("strict_stream_buffer", outputPolicy.StreamBuffer),
		zap.String("xml_completion_tool", outputPolicy.XMLCompletionTool),
		zap.Bool("normalized", result.normalized),
		zap.String("response_mode", result.responseMode),
		zap.Int("content_len", result.contentLen),
		zap.Int("reasoning_len", result.reasoningLen),
		zap.String("assembled_content", result.message.Content),
		zap.String("assembled_reasoning", result.message.ReasoningContent),
		zap.Strings("assembled_tool_calls", toolCallNames),
		zap.Int("assembled_tool_call_count", len(toolCallNames)),
	)
}

type chatCompletionOutput struct {
	message           chatMessage
	finishReason      string
	usage             *openAIUsage
	normalized        bool
	responseMode      string
	contentLen        int
	reasoningLen      int
	toolCalls         []any
	toolCallOrigin    string
	toolValidationErr error
}

func collectChatCompletionOutput(ctx context.Context, req chatCompletionRequest, handle edgeservice.RunResult, outputPolicy strictOutputPolicy) (chatCompletionOutput, error) {
	text, reasoning, finishReason, nativeToolCalls, usage, _, err := collectRunResult(ctx, handle.Stream(), handle.WaitTimeout())
	if err != nil {
		return chatCompletionOutput{}, err
	}
	text, reasoning, normalized := normalizeCompletionOutput(outputPolicy, text, reasoning, req.explicitlyIncludesReasoning())
	message := chatMessage{Role: "assistant", Content: text, ReasoningContent: reasoning}
	includeReasoning := req.includeReasoning()
	if !includeReasoning {
		message.ReasoningContent = ""
	}
	var toolCalls []any
	toolCallOrigin := ""
	var toolValidationErr error

	if len(nativeToolCalls) > 0 {
		if len(req.Tools) > 0 && strings.TrimSpace(message.Content) != "" {
			filter := &streamToolTextFilter{}
			message.Content = filter.Append(message.Content) + filter.FlushNonCandidate()
		}
		nativeToolCalls = normalizeNativeToolCallsForResponse(nativeToolCalls, req.Tools)
		message.ToolCalls = nativeToolCalls
		toolCalls = nativeToolCalls
		toolCallOrigin = toolCallOriginNative
		finishReason = "tool_calls"
	} else if len(req.Tools) > 0 {
		res := synthesizeToolCallsFromTextResult(text, req.Tools, handle.Dispatch().RunID)
		if res.validationErr != nil {
			toolValidationErr = res.validationErr
		} else if len(res.toolCalls) > 0 {
			message.Content = res.cleaned
			message.ReasoningContent = ""
			message.ToolCalls = res.toolCalls
			toolCalls = res.toolCalls
			toolCallOrigin = toolCallOriginText
			finishReason = "tool_calls"
		}
	}
	if len(message.ToolCalls) == 0 && strings.TrimSpace(message.Content) == "" && strings.TrimSpace(reasoning) != "" {
		if includeReasoning {
			message.Content = reasoningOnlyFallbackContent(reasoning, finishReason)
		} else {
			message.Content = hiddenReasoningFallbackContent(finishReason)
		}
	}

	return chatCompletionOutput{
		message:           message,
		finishReason:      finishReason,
		usage:             usage,
		normalized:        normalized,
		responseMode:      responseModeNormalized,
		contentLen:        len(message.Content),
		reasoningLen:      len(reasoning),
		toolCalls:         toolCalls,
		toolCallOrigin:    toolCallOrigin,
		toolValidationErr: toolValidationErr,
	}, nil
}

func reasoningOnlyFallbackContent(reasoning, finishReason string) string {
	content := strings.TrimSpace(reasoning)
	if content == "" {
		return ""
	}
	reason := strings.TrimSpace(finishReason)
	if reason == "" || reason == "stop" {
		return content
	}
	return content + "\n\n[IOP notice: model finished before producing final assistant content; finish_reason=" + reason + "]"
}

func hiddenReasoningFallbackContent(finishReason string) string {
	reason := strings.TrimSpace(finishReason)
	if reason == "" {
		reason = "unknown"
	}
	return "[IOP notice: model produced reasoning but no final assistant content; reasoning was hidden by include_reasoning=false; finish_reason=" + reason + "]"
}