webview-trpc-messaging

Реализует tRPC-основанное взаимодействие между хостом расширений VS Code и React-вебвью. Используйте при создании новых процедур вебвью (запросов, мутаций,…)

npx skills add https://github.com/microsoft/vscode-documentdb --skill webview-trpc-messaging

Webview tRPC Messaging

Type-safe RPC communication between the VS Code extension host (server) and React webviews (client) using tRPC.

Architecture Overview

React Webview (client)                    Extension Host (server)
─────────────────────                     ──────────────────────
useTrpcClient() hook                      WebviewController
  └─ createTRPCClient                       └─ setupTrpc()
       └─ vscodeLink ──── postMessage ────►     ├─ callerFactory(appRouter)
            (send/onReceive)              ◄─────┤   └─ procedure(input)
                                                └─ abort/subscription.stop

Key files (read as needed for implementation details):

FilePurpose
@microsoft/vscode-ext-webview (shared)tRPC init via initWebviewTrpc, publicProcedure, router, BaseRouterContext
@microsoft/vscode-ext-webview/host (telemetry)telemetryMiddlewareBody, ProcedureLogger, TelemetryRunner (consumer builds publicProcedureWithTelemetry)
src/webviews/_integration/trpc.tsConsumer tRPC instance: publicProcedureWithTelemetry, the DocumentDB TelemetryRunner, and the RpcEnrichment shape it contributes to ctx.actionContext
src/webviews/_integration/appRouter.tsRoot router + publicProcedureWithTelemetry wiring + DocumentDB BaseRouterContext
src/webviews/_integration/configuration.tsConsumer-owned knobs (telemetry namespace, bundle layout, dev-server host)
@microsoft/vscode-ext-webview/host (WebviewController)WebviewController + openWebview factory: WebviewPanel lifecycle, tRPC dispatcher (queries, mutations, subscriptions, abort)
src/webviews/_integration/openAppWebview.tsDocumentDB factory preset that pre-fills router + bundle layout (openAppWebview)
src/webviews/_integration/useTrpcClient.tsReact hook providing the tRPC client (pre-typed against AppRouter)
@microsoft/vscode-ext-webview/webview (vscodeLink)Custom tRPC link bridging postMessage transport

Creating a New Router

Each webview maintains its own router. Follow this pattern:

1. Define the router context

Extend BaseRouterContext with view-specific fields:

// src/webviews/documentdb/myView/myViewRouter.ts
import { type BaseRouterContext } from '../../_integration/appRouter';

export type RouterContext = BaseRouterContext & {
  clusterId: string;
  viewId: string;
  databaseName: string;
  // add view-specific fields
};

2. Define procedures

import { z } from 'zod';
import {
  publicProcedure,
  publicProcedureWithTelemetry,
  router,
  type WithTelemetry,
} from '../../_integration/appRouter';
import { type RouterContext } from './myViewRouter';

export const myViewRouter = router({
  // Query with telemetry (preferred for operations that touch external services)
  getData: publicProcedureWithTelemetry.input(z.object({ id: z.string() })).query(async ({ input, ctx }) => {
    const myCtx = ctx as WithTelemetry<RouterContext>;
    // Instrumented procedure: myCtx.actionContext (the full IActionContext) is present.
    myCtx.actionContext.telemetry.properties.itemId = input.id;
    // myCtx.signal is the AbortSignal for cancellation
    return { data: 'result' };
  }),

  // Mutation without telemetry (rare, use for fire-and-forget)
  doAction: publicProcedure.input(z.string()).mutation(({ input }) => {
    // lightweight operation
  }),
});

3. Register in appRouter

// src/webviews/_integration/appRouter.ts
import { myViewRouter } from '../../documentdb/myView/myViewRouter';

export const appRouter = router({
  common: commonRouter,
  mongoClusters: {
    documentView: documentViewRouter,
    collectionView: collectionViewRouter,
    myView: myViewRouter, // <-- add here
  },
});

4. Create the controller

Construction-only panels are opened with a factory function that builds the config + router context and calls the openAppWebview preset (which pre-fills the app router, caller factory, and bundle layout):

// src/webviews/documentdb/myView/myViewController.ts
import * as vscode from 'vscode';
import { API } from '../../../DocumentDBExperiences';
import { type AppWebviewController, openAppWebview } from '../../_integration/openAppWebview';
import { type RouterContext } from './myViewRouter';

export function openMyViewPanel(initialData: MyViewConfig): AppWebviewController<MyViewConfig> {
  const title = `${initialData.databaseName}`;

  const trpcContext: RouterContext = {
    dbExperience: API.DocumentDB,
    webviewName: 'myView',
    clusterId: initialData.clusterId,
    viewId: initialData.viewId,
    databaseName: initialData.databaseName,
  };

  return openAppWebview({
    title,
    webviewName: 'myView',
    config: initialData,
    context: trpcContext,
  });
}

The returned AppWebviewController handle exposes panel, onDisposed, revealToForeground, isDisposed, and dispose. Genuinely stateful panels may still extend WebviewController from @microsoft/vscode-ext-webview/host directly instead of using the factory.

Important: The webviewName field passed to openAppWebview is the registry key (viewType, must match a key in WebviewRegistry, e.g. collectionView). The webviewName in the tRPC context is a telemetry label used in telemetry event names. These may be the same string but serve different purposes -- do not confuse them.

5. Register in WebviewRegistry

Add your React component to the registry. The key must match the webviewName passed to openAppWebview (viewType). The WebviewName type (exported from the same file) ensures compile-time validation of webview names.

// src/webviews/_integration/WebviewRegistry.ts
import { MyView } from '../../documentdb/myView/MyView';

export const WebviewRegistry = {
  collectionView: CollectionView,
  documentView: DocumentView,
  myViewName: MyView, // <-- add your entry
} as const;

export type WebviewName = keyof typeof WebviewRegistry;

Telemetry: publicProcedure vs publicProcedureWithTelemetry

BaseWhen to usectx.actionContext
publicProcedureFire-and-forget, no external calls, telemetry reported separatelyabsent (do not read it)
publicProcedureWithTelemetryDefault choice. Any procedure touching DB, network, or user-visible workGuaranteed, injected by the DocumentDB TelemetryRunner

publicProcedureWithTelemetry is publicProcedure.use(telemetryMiddlewareBody(documentDbTelemetryRunner, ...)) (built in trpc.ts). The framework's telemetryMiddlewareBody delegates to the DocumentDB TelemetryRunner, which wraps the call in callWithTelemetryAndErrorHandling, contributes the full IActionContext to ctx.actionContext, auto-generates a telemetry event named documentDB.rpc.{type}.{path}, and records errors, duration, and abort status.

actionContext is not a field on the base RouterContext — it is an additive enrichment. Narrow to WithTelemetry<RouterContext> (= RouterContext & { actionContext }) in an instrumented procedure to read it; a plain publicProcedure procedure narrows to bare RouterContext, so reading actionContext there is a compile error instead of a runtime undefined.

Access telemetry safely:

const myCtx = ctx as WithTelemetry<RouterContext>;
myCtx.actionContext.telemetry.properties.myCustomProp = 'value';
myCtx.actionContext.telemetry.measurements.itemCount = items.length;

AbortSignal Support

Every tRPC operation (query, mutation, subscription) receives its own AbortController. Cancellation flows:

Client (React)                              Server (Extension Host)
──────────────                              ──────────────────────
// Queries/Mutations:
ac = new AbortController()
trpcClient.myProc.query(input,
  { signal: ac.signal })
ac.abort()  →  sends 'abort' msg  →  abortController.abort()
                                        → ctx.signal.aborted = true

// Subscriptions:
sub = trpcClient.mySub.subscribe(...)
sub.unsubscribe()  →  'subscription.stop'  →  abortController.abort()

Using abort in procedures

// Pass signal to APIs that accept it (MongoDB driver, fetch, etc.)
getData: publicProcedureWithTelemetry
    .input(z.object({ filter: z.record(z.unknown()) }))
    .query(async ({ input, ctx }) => {
        const myCtx = ctx as RouterContext;

        // Option 1: Pass to driver (preferred)
        const cursor = collection.find(input.filter, { signal: myCtx.signal });

        // Option 2: Manual check in loops
        for (const item of items) {
            if (myCtx.signal?.aborted) return;
            await processItem(item);
        }
    }),

Client-side abort

const trpcClient = useTrpcClient();
const abortControllerRef = useRef<AbortController>();

const runQuery = async () => {
  abortControllerRef.current?.abort(); // cancel previous
  const ac = new AbortController();
  abortControllerRef.current = ac;

  const result = await trpcClient.mongoClusters.collectionView.myQuery.query(input, { signal: ac.signal });
};

When publicProcedureWithTelemetry detects an aborted signal, the DocumentDB TelemetryRunner sets telemetry.properties.aborted = 'true' and result = 'Canceled' automatically.

Subscriptions

Subscriptions stream multiple values from server to client using async generators:

// Server (router)
streamData: publicProcedureWithTelemetry
    .input(z.object({ batchSize: z.number() }))
    .subscription(async function* ({ input, ctx }) {
        const myCtx = ctx as RouterContext;

        for (let i = 0; i < total; i += input.batchSize) {
            if (myCtx.signal?.aborted) return; // check before each yield
            const batch = await fetchBatch(i, input.batchSize);
            yield batch;
        }
    }),

// Client (React)
const sub = trpcClient.mongoClusters.myView.streamData.subscribe(
    { batchSize: 100 },
    {
        onData(batch) { /* handle each batch */ },
        onComplete() { /* all done */ },
        onError(err) { /* handle error */ },
    },
);

// To stop:
sub.unsubscribe();

Client-Side Hook Usage

import { useTrpcClient } from '../_integration/useTrpcClient';
import { useConfiguration } from '@microsoft/vscode-ext-webview/react';

export const MyComponent = () => {
  const trpcClient = useTrpcClient();
  const config = useConfiguration<MyViewConfig>();

  useEffect(() => {
    trpcClient.mongoClusters.myView.getData.query({ id: config.documentId }).then(setData);
  }, []);
};

useConfiguration<T>() retrieves the initial config passed to WebviewController constructor (serialized via encodeURIComponent(JSON.stringify(...))).

Common Pitfalls

  • Never use any in procedure context casts — narrow with ctx as WithTelemetry<RouterContext> when the procedure reads telemetry (ctx.actionContext.telemetry), or ctx as RouterContext otherwise
  • Always prefer publicProcedureWithTelemetry unless you have a specific reason not to
  • Always check myCtx.signal?.aborted in long-running loops — not checking causes wasted work after client cancels
  • Do not mutate the shared context object — WebviewController clones it per-operation already, but router code should treat ctx as read-only
  • Input validation uses zod — always define .input(z.object({...})) for type safety
  • The commonRouter handles cross-cutting concerns (error reporting, telemetry events, surveys, URL opening) — do not duplicate these in view-specific routers

Больше skills от microsoft

oss-growth
microsoft
Персона OSS-хакера роста
agent-framework-azure-ai-py
microsoft
Создание агентов Azure AI Foundry с использованием Microsoft Agent Framework Python SDK (agent-framework-azure-ai). Используйте при создании постоянных агентов с AzureAIAgentsProvider, применении размещенных инструментов (интерпретатор кода, поиск файлов, веб-поиск), интеграции MCP-серверов, управлении потоками бесед или реализации потоковых ответов. Охватывает функциональные инструменты, структурированные выходные данные и агентов с несколькими инструментами.
development
airunway-aks-setup
microsoft
Настройка AI Runway на AKS — от пустого кластера до работающей модели. Охватывает проверку кластера, установку контроллера, оценку GPU, настройку провайдера и первое развертывание. КОГДА: «настроить AI Runway», «подключить кластер AKS», «установить AI Runway», «airunway setup», «развернуть модель на AKS», «GPU-инференс на AKS», «настройка KAITO на AKS», «запуск LLM на AKS», «vLLM на AKS», «настройка обслуживания моделей на AKS», «контроллер AI Runway».
devops
appinsights-instrumentation
microsoft
Руководство по инструментированию веб-приложений с помощью Azure Application Insights. Содержит шаблоны телеметрии, настройку SDK и справочники по конфигурации. КОГДА: как инструментировать приложение, SDK App Insights, шаблоны телеметрии, что такое App Insights, руководство по Application Insights, примеры инструментирования, лучшие практики APM.
devops
applicationinsights-web-ts
microsoft
Инструментируйте браузерные/веб-приложения с помощью JavaScript SDK Application Insights (@microsoft/applicationinsights-web). Используйте для мониторинга реальных пользователей (RUM) — просмотры страниц, клики, зависимости AJAX/fetch, исключения, пользовательские события и трассировки агентов GenAI на стороне браузера, коррелируемые с бэкенд-трассировками OpenTelemetry. Охватывает скрипт загрузчика SDK и настройку npm, расширения фреймворков (React, React Native, Angular), Click Analytics, инициализаторы телеметрии и семантические конвенции OTel GenAI для спанов агента/инструмента/модели, генерируемых из браузера.
devops
azure-ai-anomalydetector-java
microsoft
Создавайте приложения для обнаружения аномалий с помощью Azure AI Anomaly Detector SDK для Java. Используйте при реализации одномерного/многомерного обнаружения аномалий, анализа временных рядов или мониторинга на основе ИИ.
development
azure-ai-language-conversations-py
microsoft
Реализация понимания разговорного языка (CLU) с использованием Python SDK azure-ai-language-conversations. Используйте при работе с ConversationAnalysisClient для анализа намерений и сущностей в разговоре, создании NLP-функций или интеграции языкового понимания в приложения.
development
azure-ai-ml-py
microsoft
Azure Machine Learning SDK v2 для Python. Используется для рабочих областей ML, заданий, моделей, наборов данных, вычислений и конвейеров. Триггеры: "azure-ai-ml", "MLClient", "workspace", "model registry", "training jobs", "datasets".
development