@vscode/chat-lib 0.4.1-1 → 0.4.1-11

package/dist/src/_internal/extension/prompt/node/chatMLFetcher.js

@@ -3,45 +3,12 @@
  *  Copyright (c) Microsoft Corporation. All rights reserved.
  *  Licensed under the MIT License. See License.txt in the project root for license information.
  *--------------------------------------------------------------------------------------------*/
-var __createBinding = (this && this.__createBinding) || (Object.create ? (function(o, m, k, k2) {
-    if (k2 === undefined) k2 = k;
-    var desc = Object.getOwnPropertyDescriptor(m, k);
-    if (!desc || ("get" in desc ? !m.__esModule : desc.writable || desc.configurable)) {
-      desc = { enumerable: true, get: function() { return m[k]; } };
-    }
-    Object.defineProperty(o, k2, desc);
-}) : (function(o, m, k, k2) {
-    if (k2 === undefined) k2 = k;
-    o[k2] = m[k];
-}));
-var __setModuleDefault = (this && this.__setModuleDefault) || (Object.create ? (function(o, v) {
-    Object.defineProperty(o, "default", { enumerable: true, value: v });
-}) : function(o, v) {
-    o["default"] = v;
-});
 var __decorate = (this && this.__decorate) || function (decorators, target, key, desc) {
     var c = arguments.length, r = c < 3 ? target : desc === null ? desc = Object.getOwnPropertyDescriptor(target, key) : desc, d;
     if (typeof Reflect === "object" && typeof Reflect.decorate === "function") r = Reflect.decorate(decorators, target, key, desc);
     else for (var i = decorators.length - 1; i >= 0; i--) if (d = decorators[i]) r = (c < 3 ? d(r) : c > 3 ? d(target, key, r) : d(target, key)) || r;
     return c > 3 && r && Object.defineProperty(target, key, r), r;
 };
-var __importStar = (this && this.__importStar) || (function () {
-    var ownKeys = function(o) {
-        ownKeys = Object.getOwnPropertyNames || function (o) {
-            var ar = [];
-            for (var k in o) if (Object.prototype.hasOwnProperty.call(o, k)) ar[ar.length] = k;
-            return ar;
-        };
-        return ownKeys(o);
-    };
-    return function (mod) {
-        if (mod && mod.__esModule) return mod;
-        var result = {};
-        if (mod != null) for (var k = ownKeys(mod), i = 0; i < k.length; i++) if (k[i] !== "default") __createBinding(result, mod, k[i]);
-        __setModuleDefault(result, mod);
-        return result;
-    };
-})();
 var __param = (this && this.__param) || function (paramIndex, decorator) {
     return function (target, key) { decorator(target, key, paramIndex); }
 };
@@ -68,6 +35,7 @@ const fetcherService_1 = require("../../../platform/networking/common/fetcherSer
 const networking_1 = require("../../../platform/networking/common/networking");
 const openai_1 = require("../../../platform/networking/common/openai");
 const chatStream_1 = require("../../../platform/networking/node/chatStream");
+const chatWebSocketManager_1 = require("../../../platform/networking/node/chatWebSocketManager");
 const stream_1 = require("../../../platform/networking/node/stream");
 const fetch_2 = require("../../../platform/openai/node/fetch");
 const requestLogger_1 = require("../../../platform/requestLogger/node/requestLogger");
@@ -75,14 +43,14 @@ const nullExperimentationService_1 = require("../../../platform/telemetry/common
 const telemetry_1 = require("../../../platform/telemetry/common/telemetry");
 const telemetryData_1 = require("../../../platform/telemetry/common/telemetryData");
 const anomalyDetection_1 = require("../../../util/common/anomalyDetection");
-const crypto_1 = require("../../../util/common/crypto");
-const errorsUtil = __importStar(require("../../../util/common/errors"));
+const errors_1 = require("../../../util/common/errors");
 const async_1 = require("../../../util/vs/base/common/async");
-const errors_1 = require("../../../util/vs/base/common/errors");
+const errors_2 = require("../../../util/vs/base/common/errors");
 const event_1 = require("../../../util/vs/base/common/event");
 const lifecycle_1 = require("../../../util/vs/base/common/lifecycle");
 const strings_1 = require("../../../util/vs/base/common/strings");
 const uuid_1 = require("../../../util/vs/base/common/uuid");
+const instantiation_1 = require("../../../util/vs/platform/instantiation/common/instantiation");
 const openAIEndpoint_1 = require("../../byok/node/openAIEndpoint");
 const constants_1 = require("../../common/constants");
 const powerService_1 = require("../../power/common/powerService");
@@ -116,7 +84,7 @@ class AbstractChatMLFetcher extends lifecycle_1.Disposable {
 }
 exports.AbstractChatMLFetcher = AbstractChatMLFetcher;
 let ChatMLFetcherImpl = class ChatMLFetcherImpl extends AbstractChatMLFetcher {
-    constructor(_fetcherService, _telemetryService, _requestLogger, _logService, _authenticationService, _interactionService, _chatQuotaService, _capiClientService, options, _configurationService, _experimentationService, _powerService) {
+    constructor(_fetcherService, _telemetryService, _requestLogger, _logService, _authenticationService, _interactionService, _chatQuotaService, _capiClientService, options, _configurationService, _experimentationService, _powerService, _instantiationService, _webSocketManager) {
         super(options);
         this._fetcherService = _fetcherService;
         this._telemetryService = _telemetryService;
@@ -129,6 +97,8 @@ let ChatMLFetcherImpl = class ChatMLFetcherImpl extends AbstractChatMLFetcher {
         this._configurationService = _configurationService;
         this._experimentationService = _experimentationService;
         this._powerService = _powerService;
+        this._instantiationService = _instantiationService;
+        this._webSocketManager = _webSocketManager;
         /**
          * Delays (in ms) between connectivity check attempts before retrying a failed request.
          * Configurable for testing purposes.
@@ -139,13 +109,20 @@ let ChatMLFetcherImpl = class ChatMLFetcherImpl extends AbstractChatMLFetcher {
      * Note: the returned array of strings may be less than `n` (e.g., in case there were errors during streaming)
      */
     async fetchMany(opts, token) {
-        let { debugName, endpoint: chatEndpoint, finishedCb, location, messages, requestOptions, source, telemetryProperties, userInitiatedRequest } = opts;
+        let { debugName, endpoint: chatEndpoint, finishedCb, location, messages, requestOptions, source, telemetryProperties, userInitiatedRequest, requestKindOptions, conversationId, turnId } = opts;
         if (!telemetryProperties) {
             telemetryProperties = {};
         }
         if (!telemetryProperties.messageSource) {
             telemetryProperties.messageSource = debugName;
         }
+        // When WebSocket is enabled, default to using the stateful marker so
+        // only delta history is sent once the server has context from this turn.
+        const useWebSocket = !!(turnId && conversationId
+            && chatEndpoint.apiType === 'responses'
+            && this._configurationService.getExperimentBasedConfig(configurationService_1.ConfigKey.TeamInternal.ResponsesApiWebSocketEnabled, this._experimentationService));
+        const transport = useWebSocket ? 'websocket' : 'http';
+        const ignoreStatefulMarker = useWebSocket ? (opts.ignoreStatefulMarker ?? false) : opts.ignoreStatefulMarker;
         // TODO @lramos15 telemetry should not drive request ids
         const ourRequestId = telemetryProperties.requestId ?? telemetryProperties.messageId ?? (0, uuid_1.generateUuid)();
         const maxResponseTokens = chatEndpoint.maxOutputTokens;
@@ -159,6 +136,7 @@ let ChatMLFetcherImpl = class ChatMLFetcherImpl extends AbstractChatMLFetcher {
         const postOptions = this.preparePostOptions(requestOptions);
         const requestBody = chatEndpoint.createRequestBody({
             ...opts,
+            ignoreStatefulMarker,
             requestId: ourRequestId,
             postOptions
         });
@@ -173,7 +151,7 @@ let ChatMLFetcherImpl = class ChatMLFetcherImpl extends AbstractChatMLFetcher {
             ourRequestId,
             location: opts.location,
             body: requestBody,
-            ignoreStatefulMarker: opts.ignoreStatefulMarker,
+            ignoreStatefulMarker,
             isConversationRequest: opts.isConversationRequest,
             customMetadata: opts.customMetadata
         });
@@ -201,7 +179,7 @@ let ChatMLFetcherImpl = class ChatMLFetcherImpl extends AbstractChatMLFetcher {
             else {
                 const copilotToken = await this._authenticationService.getCopilotToken();
                 usernameToScrub = copilotToken.username;
-                const fetchResult = await this._fetchAndStreamChat(chatEndpoint, requestBody, baseTelemetry, streamRecorder.callback, requestOptions.secretKey, copilotToken, opts.location, ourRequestId, postOptions.n, token, userInitiatedRequest, telemetryProperties, opts.useFetcher, canRetryOnce);
+                const fetchResult = await this._fetchAndStreamChat(chatEndpoint, requestBody, baseTelemetry, streamRecorder.callback, requestOptions.secretKey, copilotToken, opts.location, ourRequestId, postOptions.n, token, userInitiatedRequest, useWebSocket, turnId, conversationId, telemetryProperties, opts.useFetcher, canRetryOnce, requestKindOptions);
                 response = fetchResult.result;
                 actualFetcher = fetchResult.fetcher;
                 actualBytesReceived = fetchResult.bytesReceived;
@@ -221,7 +199,7 @@ let ChatMLFetcherImpl = class ChatMLFetcherImpl extends AbstractChatMLFetcher {
             pendingLoggedChatRequest?.markTimeToFirstToken(timeToFirstToken);
             switch (response.type) {
                 case fetch_2.FetchResponseKind.Success: {
-                    const result = await this.processSuccessfulResponse(response, messages, requestBody, ourRequestId, maxResponseTokens, tokenCount, timeToFirstToken, streamRecorder, baseTelemetry, chatEndpoint, userInitiatedRequest, actualFetcher, actualBytesReceived, suspendEventSeen, resumeEventSeen);
+                    const result = await this.processSuccessfulResponse(response, messages, requestBody, ourRequestId, maxResponseTokens, tokenCount, timeToFirstToken, streamRecorder, baseTelemetry, chatEndpoint, userInitiatedRequest, transport, actualFetcher, actualBytesReceived, suspendEventSeen, resumeEventSeen);
                     // Handle FilteredRetry case with augmented messages
                     if (result.type === commonTypes_1.ChatFetchResponseType.FilteredRetry) {
                         if (opts.enableRetryOnFilter) {
@@ -277,6 +255,7 @@ let ChatMLFetcherImpl = class ChatMLFetcherImpl extends AbstractChatMLFetcher {
                         requestId: ourRequestId,
                         model: chatEndpoint.model,
                         apiType: chatEndpoint.apiType,
+                        transport,
                         associatedRequestId: telemetryProperties.associatedRequestId,
                         retryAfterError: telemetryProperties.retryAfterError,
                         retryAfterErrorGitHubRequestId: telemetryProperties.retryAfterErrorGitHubRequestId,
@@ -317,6 +296,7 @@ let ChatMLFetcherImpl = class ChatMLFetcherImpl extends AbstractChatMLFetcher {
                             tokenCount,
                             maxResponseTokens,
                             timeToError: timeToFirstToken,
+                            transport,
                             actualFetcher,
                             bytesReceived: actualBytesReceived,
                             baseTelemetry,
@@ -342,6 +322,7 @@ let ChatMLFetcherImpl = class ChatMLFetcherImpl extends AbstractChatMLFetcher {
                         maxResponseTokens,
                         timeToFirstToken,
                         isVisionRequest: this.filterImageMessages(messages),
+                        transport,
                         fetcher: actualFetcher,
                         bytesReceived: actualBytesReceived,
                         issuedTime: baseTelemetry.issuedTime,
@@ -377,6 +358,7 @@ let ChatMLFetcherImpl = class ChatMLFetcherImpl extends AbstractChatMLFetcher {
                         tokenCount,
                         maxResponseTokens,
                         timeToError,
+                        transport,
                         actualFetcher,
                         bytesReceived: err.bytesReceived,
                         baseTelemetry,
@@ -401,6 +383,7 @@ let ChatMLFetcherImpl = class ChatMLFetcherImpl extends AbstractChatMLFetcher {
                     requestId: ourRequestId,
                     model: chatEndpoint.model,
                     apiType: chatEndpoint.apiType,
+                    transport,
                     associatedRequestId: telemetryProperties.associatedRequestId,
                     retryAfterError: telemetryProperties.retryAfterError,
                     retryAfterErrorGitHubRequestId: telemetryProperties.retryAfterErrorGitHubRequestId,
@@ -433,6 +416,7 @@ let ChatMLFetcherImpl = class ChatMLFetcherImpl extends AbstractChatMLFetcher {
                     maxResponseTokens,
                     timeToFirstToken: timeToError,
                     isVisionRequest: this.filterImageMessages(messages),
+                    transport,
                     fetcher: actualFetcher,
                     bytesReceived: err.bytesReceived,
                     issuedTime: baseTelemetry.issuedTime,
@@ -500,7 +484,7 @@ let ChatMLFetcherImpl = class ChatMLFetcherImpl extends AbstractChatMLFetcher {
         return authHeaders;
     }
     async _retryAfterError(params) {
-        const { opts, processed, telemetryProperties, requestBody, tokenCount, maxResponseTokens, timeToError, actualFetcher, bytesReceived, baseTelemetry, streamRecorder, retryReason, debugNamePrefix, pendingLoggedChatRequest, token, usernameToScrub, suspendEventSeen, resumeEventSeen, } = params;
+        const { opts, processed, telemetryProperties, requestBody, tokenCount, maxResponseTokens, timeToError, transport, actualFetcher, bytesReceived, baseTelemetry, streamRecorder, retryReason, debugNamePrefix, pendingLoggedChatRequest, token, usernameToScrub, suspendEventSeen, resumeEventSeen, } = params;
         // net::ERR_NETWORK_CHANGED: https://github.com/microsoft/vscode/issues/260297
         const isNetworkChangedError = ['darwin', 'linux'].includes(process.platform) && processed.reason.indexOf('net::ERR_NETWORK_CHANGED') !== -1;
         // When Electron's network process crashes, all requests through it fail permanently.
@@ -527,6 +511,7 @@ let ChatMLFetcherImpl = class ChatMLFetcherImpl extends AbstractChatMLFetcher {
             maxResponseTokens,
             timeToFirstToken: timeToError,
             isVisionRequest: this.filterImageMessages(opts.messages),
+            transport,
             fetcher: actualFetcher,
             bytesReceived,
             issuedTime: baseTelemetry.issuedTime,
@@ -552,7 +537,7 @@ let ChatMLFetcherImpl = class ChatMLFetcherImpl extends AbstractChatMLFetcher {
         pendingLoggedChatRequest?.resolve(retryResult, streamRecorder.deltas);
         return { retryResult, connectivityTestError, connectivityTestErrorGitHubRequestId };
     }
-    async _fetchAndStreamChat(chatEndpointInfo, request, baseTelemetryData, finishedCb, secretKey, copilotToken, location, ourRequestId, nChoices, cancellationToken, userInitiatedRequest, telemetryProperties, useFetcher, canRetryOnce) {
+    async _fetchAndStreamChat(chatEndpointInfo, request, baseTelemetryData, finishedCb, secretKey, copilotToken, location, ourRequestId, nChoices, cancellationToken, userInitiatedRequest, useWebSocket, turnId, conversationId, telemetryProperties, useFetcher, canRetryOnce, requestKindOptions) {
         const isPowerSaveBlockerEnabled = this._configurationService.getExperimentBasedConfig(configurationService_1.ConfigKey.TeamInternal.ChatRequestPowerSaveBlocker, this._experimentationService);
         const blockerHandle = isPowerSaveBlockerEnabled && location !== commonTypes_1.ChatLocation.Other ? this._powerService.acquirePowerSaveBlocker() : undefined;
         let suspendEventSeen = false;
@@ -566,7 +551,7 @@ let ChatMLFetcherImpl = class ChatMLFetcherImpl extends AbstractChatMLFetcher {
             this._logService.info(`System resumed during streaming request ${ourRequestId} (${commonTypes_1.ChatLocation.toString(location)})`);
         });
         try {
-            const fetchResult = await this._doFetchAndStreamChat(chatEndpointInfo, request, baseTelemetryData, finishedCb, secretKey, copilotToken, location, ourRequestId, nChoices, cancellationToken, userInitiatedRequest, telemetryProperties, useFetcher, canRetryOnce);
+            const fetchResult = await this._doFetchAndStreamChat(chatEndpointInfo, request, baseTelemetryData, finishedCb, secretKey, copilotToken, location, ourRequestId, nChoices, cancellationToken, userInitiatedRequest, useWebSocket, turnId, conversationId, telemetryProperties, useFetcher, canRetryOnce, requestKindOptions);
             return { ...fetchResult, suspendEventSeen: suspendEventSeen || undefined, resumeEventSeen: resumeEventSeen || undefined };
         }
         catch (err) {
@@ -584,7 +569,7 @@ let ChatMLFetcherImpl = class ChatMLFetcherImpl extends AbstractChatMLFetcher {
             blockerHandle?.dispose();
         }
     }
-    async _doFetchAndStreamChat(chatEndpointInfo, request, baseTelemetryData, finishedCb, secretKey, copilotToken, location, ourRequestId, nChoices, cancellationToken, userInitiatedRequest, telemetryProperties, useFetcher, canRetryOnce) {
+    async _doFetchAndStreamChat(chatEndpointInfo, request, baseTelemetryData, finishedCb, secretKey, copilotToken, location, ourRequestId, nChoices, cancellationToken, userInitiatedRequest, useWebSocket, turnId, conversationId, telemetryProperties, useFetcher, canRetryOnce, requestKindOptions) {
         if (cancellationToken.isCancellationRequested) {
             return { result: { type: fetch_2.FetchResponseKind.Canceled, reason: 'before fetch request' } };
         }
@@ -606,9 +591,100 @@ let ChatMLFetcherImpl = class ChatMLFetcherImpl extends AbstractChatMLFetcher {
                 }
             };
         }
+        // WebSocket path: use persistent WebSocket connection for Responses API endpoints
+        if (useWebSocket && turnId && conversationId) {
+            return this._doFetchViaWebSocket(chatEndpointInfo, request, baseTelemetryData, finishedCb, secretKey, location, ourRequestId, turnId, conversationId, cancellationToken, telemetryProperties);
+        }
+        return this._doFetchViaHttp(chatEndpointInfo, request, baseTelemetryData, finishedCb, secretKey, location, ourRequestId, nChoices, cancellationToken, userInitiatedRequest, telemetryProperties, useFetcher, canRetryOnce, requestKindOptions);
+    }
+    /**
+     * Sends a chat request via a persistent WebSocket connection instead of HTTP POST.
+     * Events are the same Responses API streaming events, processed by OpenAIResponsesProcessor.
+     */
+    async _doFetchViaWebSocket(chatEndpointInfo, request, baseTelemetryData, finishedCb, secretKey, location, ourRequestId, turnId, conversationId, cancellationToken, telemetryProperties) {
+        const connection = await this._webSocketManager.getOrCreateConnection(conversationId, turnId, secretKey);
+        // Generate unique ID to link input and output messages
+        const modelCallId = (0, uuid_1.generateUuid)();
+        const telemetryData = telemetryData_1.TelemetryData.createAndMarkAsIssued({
+            endpoint: 'completions',
+            engineName: 'chat',
+            uiKind: commonTypes_1.ChatLocation.toString(location),
+            transport: 'websocket',
+            ...{ ...telemetryProperties, modelCallId },
+        }, {
+            maxTokenWindow: chatEndpointInfo.modelMaxPromptTokens
+        });
+        for (const [key, value] of Object.entries(request)) {
+            if (key === 'messages' || key === 'input') {
+                continue;
+            } // Skip messages (PII)
+            telemetryData.properties[`request.option.${key}`] = JSON.stringify(value) ?? 'undefined';
+        }
+        telemetryData.properties['headerRequestId'] = ourRequestId;
+        this._telemetryService.sendGHTelemetryEvent('request.sent', telemetryData.properties, telemetryData.measurements);
+        const requestStart = Date.now();
+        const handle = connection.sendRequest(request, cancellationToken);
+        const extendedBaseTelemetryData = baseTelemetryData.extendedBy({ modelCallId });
+        const processor = this._instantiationService.createInstance(responsesApi_1.OpenAIResponsesProcessor, extendedBaseTelemetryData, ourRequestId, '');
+        const chatCompletions = new async_1.AsyncIterableObject(async (emitter) => {
+            try {
+                await new Promise((resolve, reject) => {
+                    handle.onEvent(event => {
+                        const completion = processor.push(event, finishedCb);
+                        if (completion) {
+                            (0, responsesApi_1.sendCompletionOutputTelemetry)(this._telemetryService, this._logService, completion, extendedBaseTelemetryData);
+                            emitter.emitOne(completion);
+                        }
+                    });
+                    handle.onError(error => {
+                        if ((0, errors_2.isCancellationError)(error)) {
+                            reject(error);
+                            return;
+                        }
+                        const warningTelemetry = telemetryData.extendedBy({ error: error.message });
+                        this._telemetryService.sendGHTelemetryEvent('request.shownWarning', warningTelemetry.properties, warningTelemetry.measurements);
+                        const totalTimeMs = Date.now() - requestStart;
+                        telemetryData.measurements.totalTimeMs = totalTimeMs;
+                        telemetryData.properties.error = error.message;
+                        this._logService.debug(`request.error: [websocket], took ${totalTimeMs} ms`);
+                        this._telemetryService.sendGHTelemetryEvent('request.error', telemetryData.properties, telemetryData.measurements);
+                        reject(error);
+                    });
+                    handle.onComplete(() => {
+                        const totalTimeMs = Date.now() - requestStart;
+                        telemetryData.measurements.totalTimeMs = totalTimeMs;
+                        this._logService.debug(`request.response: [websocket], took ${totalTimeMs} ms`);
+                        this._telemetryService.sendGHTelemetryEvent('request.response', telemetryData.properties, telemetryData.measurements);
+                        resolve();
+                    });
+                });
+            }
+            finally {
+                let messagesToLog = request.messages;
+                if ((!messagesToLog || messagesToLog.length === 0) && request.input) {
+                    try {
+                        const rawMessages = (0, responsesApi_1.responseApiInputToRawMessagesForLogging)(request);
+                        messagesToLog = (0, openai_1.rawMessageToCAPI)(rawMessages);
+                    }
+                    catch (e) {
+                        this._logService.error(`Failed to convert Response API input to messages for telemetry:`, e);
+                        messagesToLog = [];
+                    }
+                }
+                (0, chatStream_1.sendEngineMessagesTelemetry)(this._telemetryService, messagesToLog ?? [], telemetryData, false, this._logService);
+            }
+        });
+        return {
+            result: {
+                type: fetch_2.FetchResponseKind.Success,
+                chatCompletions,
+            }
+        };
+    }
+    async _doFetchViaHttp(chatEndpointInfo, request, baseTelemetryData, finishedCb, secretKey, location, ourRequestId, nChoices, cancellationToken, userInitiatedRequest, telemetryProperties, useFetcher, canRetryOnce, requestKindOptions) {
         // Generate unique ID to link input and output messages
         const modelCallId = (0, uuid_1.generateUuid)();
-        const response = await this._fetchWithInstrumentation(chatEndpointInfo, ourRequestId, request, secretKey, location, cancellationToken, userInitiatedRequest, { ...telemetryProperties, modelCallId }, useFetcher, canRetryOnce);
+        const response = await this._fetchWithInstrumentation(chatEndpointInfo, ourRequestId, request, secretKey, location, cancellationToken, userInitiatedRequest, { ...telemetryProperties, modelCallId }, useFetcher, canRetryOnce, requestKindOptions);
         if (cancellationToken.isCancellationRequested) {
             try {
                 // Destroy the stream so that the server is hopefully notified we don't want any more data
@@ -679,7 +755,7 @@ let ChatMLFetcherImpl = class ChatMLFetcherImpl extends AbstractChatMLFetcher {
             bytesReceived: response.bytesReceived
         };
     }
-    async _fetchWithInstrumentation(chatEndpoint, ourRequestId, request, secretKey, location, cancellationToken, userInitiatedRequest, telemetryProperties, useFetcher, canRetryOnce) {
+    async _fetchWithInstrumentation(chatEndpoint, ourRequestId, request, secretKey, location, cancellationToken, userInitiatedRequest, telemetryProperties, useFetcher, canRetryOnce, requestKindOptions) {
         // If request contains an image, we include this header.
         const additionalHeaders = {
             'X-Interaction-Id': this._interactionService.interactionId,
@@ -692,6 +768,7 @@ let ChatMLFetcherImpl = class ChatMLFetcherImpl extends AbstractChatMLFetcher {
             endpoint: 'completions',
             engineName: 'chat',
             uiKind: commonTypes_1.ChatLocation.toString(location),
+            transport: 'http',
             ...telemetryProperties // This includes the modelCallId from fetchAndStreamChat
         }, {
             maxTokenWindow: chatEndpoint.modelMaxPromptTokens
@@ -710,7 +787,19 @@ let ChatMLFetcherImpl = class ChatMLFetcherImpl extends AbstractChatMLFetcher {
         const requestStart = Date.now();
         const intent = locationToIntent(location);
         // Wrap the Promise with success/error callbacks so we can log/measure it
-        return (0, networking_1.postRequest)(this._fetcherService, this._telemetryService, this._capiClientService, chatEndpoint, secretKey, await (0, crypto_1.createRequestHMAC)(process.env.HMAC_SECRET), intent, ourRequestId, request, additionalHeaders, cancellationToken, useFetcher, canRetryOnce, location).then(response => {
+        return this._instantiationService.invokeFunction(networking_1.postRequest, {
+            endpointOrUrl: chatEndpoint,
+            secretKey,
+            intent,
+            requestId: ourRequestId,
+            body: request,
+            additionalHeaders,
+            cancelToken: cancellationToken,
+            useFetcher,
+            canRetryOnce,
+            location,
+            requestKindOptions,
+        }).then(response => {
             const apim = response.headers.get('apim-request-id');
             if (apim) {
                 this._logService.debug(`APIM request id: ${apim}`);
@@ -966,7 +1055,7 @@ let ChatMLFetcherImpl = class ChatMLFetcherImpl extends AbstractChatMLFetcher {
             reason: `Request Failed: ${response.status} ${text}`
         };
     }
-    async processSuccessfulResponse(response, messages, requestBody, requestId, maxResponseTokens, promptTokenCount, timeToFirstToken, streamRecorder, baseTelemetry, chatEndpointInfo, userInitiatedRequest, fetcher, bytesReceived, suspendEventSeen, resumeEventSeen) {
+    async processSuccessfulResponse(response, messages, requestBody, requestId, maxResponseTokens, promptTokenCount, timeToFirstToken, streamRecorder, baseTelemetry, chatEndpointInfo, userInitiatedRequest, transport, fetcher, bytesReceived, suspendEventSeen, resumeEventSeen) {
         const completions = [];
         for await (const chatCompletion of response.chatCompletions) {
             chatMLFetcherTelemetry_1.ChatMLFetcherTelemetrySender.sendSuccessTelemetry(this._telemetryService, {
@@ -980,6 +1069,7 @@ let ChatMLFetcherImpl = class ChatMLFetcherImpl extends AbstractChatMLFetcher {
                 timeToFirstToken,
                 timeToFirstTokenEmitted: (baseTelemetry && streamRecorder.firstTokenEmittedTime) ? streamRecorder.firstTokenEmittedTime - baseTelemetry.issuedTime : -1,
                 hasImageMessages: this.filterImageMessages(messages),
+                transport,
                 fetcher,
                 bytesReceived,
                 suspendEventSeen,
@@ -1184,7 +1274,7 @@ let ChatMLFetcherImpl = class ChatMLFetcherImpl extends AbstractChatMLFetcher {
                 serverRequestId: gitHubRequestId,
             };
         }
-        if ((0, errors_1.isCancellationError)(err)) {
+        if ((0, errors_2.isCancellationError)(err)) {
             return {
                 type: commonTypes_1.ChatFetchResponseType.Canceled,
                 reason: 'Got a cancellation error',
@@ -1201,7 +1291,7 @@ let ChatMLFetcherImpl = class ChatMLFetcherImpl extends AbstractChatMLFetcher {
                 serverRequestId: gitHubRequestId,
             };
         }
-        this._logService.error(errorsUtil.fromUnknown(err), `Error on conversation request`);
+        this._logService.error(errors_1.ErrorUtils.fromUnknown(err), `Error on conversation request`);
         this._telemetryService.sendGHTelemetryException(err, 'Error on conversation request');
         const userMessage = fetcher.getUserMessageForFetcherError(err);
         const errorDetail = (0, logService_1.collectSingleLineErrorMessage)(err, true);
@@ -1257,7 +1347,9 @@ exports.ChatMLFetcherImpl = ChatMLFetcherImpl = __decorate([
     __param(8, conversationOptions_1.IConversationOptions),
     __param(9, configurationService_1.IConfigurationService),
     __param(10, nullExperimentationService_1.IExperimentationService),
-    __param(11, powerService_1.IPowerService)
+    __param(11, powerService_1.IPowerService),
+    __param(12, instantiation_1.IInstantiationService),
+    __param(13, chatWebSocketManager_1.IChatWebSocketManager)
 ], ChatMLFetcherImpl);
 /**
  * Validates a chat request payload to ensure it is valid