fix(gateway): route response history through ai seam

apps/aether-gateway/src/ai_serving/finalize/internal/stream_rewrite.rs

@@ -1,9 +1,8 @@
 use serde_json::Value;
 
-use aether_ai_formats::api::ResponseHistoryRecord;
-
 use crate::ai_serving::{
     maybe_build_ai_surface_stream_rewriter, AiSurfaceFinalizeError, AiSurfaceStreamRewriter,
+    ResponseHistoryRecord,
 };
 use crate::GatewayError;
 

apps/aether-gateway/src/ai_serving/pure/mod.rs

@@ -57,7 +57,7 @@ pub(crate) use aether_ai_formats::api::{
     find_kiro_real_thinking_end_tag, find_kiro_real_thinking_end_tag_at_buffer_end,
     find_kiro_real_thinking_start_tag, forbid_upstream_streaming_for_provider,
     force_upstream_streaming_for_provider, gemini_request_is_image_generation,
-    implicit_sync_finalize_report_kind, is_core_error_finalize_kind,
+    hydrate_response_history, implicit_sync_finalize_report_kind, is_core_error_finalize_kind,
     is_matching_stream_http_request, is_matching_stream_request, is_openai_image_stream_request,
     is_openai_responses_compact_format, is_openai_responses_family_format,
     is_openai_responses_format, kiro_crc32, map_claude_stop_reason,
@@ -88,10 +88,10 @@ pub(crate) use aether_ai_formats::api::{
     provider_adaptation_descriptor_for_provider_type,
     provider_adaptation_requires_eventstream_accept,
     provider_adaptation_should_unwrap_stream_envelope,
-    provider_private_response_allows_sync_finalize, request_candidate_api_format_preference,
-    request_candidate_api_formats, request_conversion_kind,
-    request_conversion_requires_enable_flag, request_path_implies_stream_request,
-    resolve_claude_stream_spec, resolve_claude_sync_spec,
+    provider_private_response_allows_sync_finalize, record_converted_response_history,
+    request_candidate_api_format_preference, request_candidate_api_formats,
+    request_conversion_kind, request_conversion_requires_enable_flag,
+    request_path_implies_stream_request, resolve_claude_stream_spec, resolve_claude_sync_spec,
     resolve_codex_responses_model_capabilities, resolve_execution_runtime_stream_plan_kind,
     resolve_execution_runtime_stream_plan_kind_with_client_surface,
     resolve_execution_runtime_sync_plan_kind,
@@ -105,14 +105,14 @@ pub(crate) use aether_ai_formats::api::{
     resolve_openai_responses_sync_spec, resolve_requested_gemini_image_model_for_request,
     resolve_requested_openai_image_model_for_request,
     resolve_upstream_is_stream_for_provider as resolve_format_upstream_is_stream_for_provider,
-    resolve_upstream_is_stream_from_endpoint_config, sanitize_request_path,
-    sanitize_request_path_and_query, sanitize_request_query_string,
-    stream_body_contains_error_event, supports_stream_execution_decision_kind,
-    supports_sync_execution_decision_kind, sync_chat_response_conversion_kind,
-    sync_cli_response_conversion_kind, transform_provider_private_stream_line,
-    validate_openai_provider_request_contract, value_as_u64, AiControlPlanRequest,
-    AiSurfaceFinalizeError, AiSurfaceStreamRewriter, CanonicalStreamFrame,
-    ChatGptWebImageRequestError, ClaudeClientEmitter, ClaudeProviderState,
+    resolve_upstream_is_stream_from_endpoint_config, response_history_is_loaded,
+    response_history_storage_key, sanitize_request_path, sanitize_request_path_and_query,
+    sanitize_request_query_string, stream_body_contains_error_event,
+    supports_stream_execution_decision_kind, supports_sync_execution_decision_kind,
+    sync_chat_response_conversion_kind, sync_cli_response_conversion_kind,
+    transform_provider_private_stream_line, validate_openai_provider_request_contract,
+    value_as_u64, AiControlPlanRequest, AiSurfaceFinalizeError, AiSurfaceStreamRewriter,
+    CanonicalStreamFrame, ChatGptWebImageRequestError, ClaudeClientEmitter, ClaudeProviderState,
     CodexResponsesModelCapabilities, ExecutionRuntimeAuthContext, FinalizeStreamRewriteMode,
     FormatContext, GeminiClientEmitter, GeminiImageRequestForOpenAi, GeminiProviderState,
     KiroToClaudeCliStreamState, LocalCoreSyncErrorKind, LocalGeminiFilesSpec, LocalOpenAiImageSpec,
@@ -124,10 +124,10 @@ pub(crate) use aether_ai_formats::api::{
     OpenAiImageRequestForGemini, OpenAiImageResponseFormat, OpenAiImageStreamState,
     OpenAiImageSyncFinalizeProduct, OpenAiProviderRequestFinalization,
     ProviderAdaptationDescriptor, ProviderAdaptationSurface, ProviderPrivateStreamNormalizer,
-    ReasoningEffort, RequestConversionKind, ServiceTier, StandardCrossFormatSyncProduct,
-    StandardSyncFinalizeNormalizedProduct, StreamingStandardFormatMatrix,
-    SyncChatResponseConversionKind, SyncCliResponseConversionKind, SyncToStreamBridgeOutcome,
-    ANTIGRAVITY_V1INTERNAL_ENVELOPE_NAME, CLAUDE_CHAT_STREAM_PLAN_KIND,
+    ReasoningEffort, RequestConversionKind, ResponseHistoryRecord, ServiceTier,
+    StandardCrossFormatSyncProduct, StandardSyncFinalizeNormalizedProduct,
+    StreamingStandardFormatMatrix, SyncChatResponseConversionKind, SyncCliResponseConversionKind,
+    SyncToStreamBridgeOutcome, ANTIGRAVITY_V1INTERNAL_ENVELOPE_NAME, CLAUDE_CHAT_STREAM_PLAN_KIND,
     CLAUDE_CHAT_STREAM_SUCCESS_REPORT_KIND, CLAUDE_CHAT_SYNC_ERROR_REPORT_KIND,
     CLAUDE_CHAT_SYNC_FINALIZE_REPORT_KIND, CLAUDE_CHAT_SYNC_PLAN_KIND,
     CLAUDE_CHAT_SYNC_SUCCESS_REPORT_KIND, CLAUDE_CLI_STREAM_PLAN_KIND,

apps/aether-gateway/src/ai_serving/response_history.rs

@@ -1,4 +1,4 @@
-use aether_ai_formats::api::{
+use crate::ai_serving::{
     hydrate_response_history, normalize_api_format_alias, record_converted_response_history,
     response_history_is_loaded, response_history_storage_key, ResponseHistoryRecord,
 };