feat(#383): Streaming AI responses via Matrix message edits
Some checks failed
ci/woodpecker/push/api Pipeline failed

- Add MatrixStreamingService with editMessage, setTypingIndicator, streamResponse
- Rate-limited edits (500ms) for incremental streaming output
- Typing indicator management during generation
- Graceful error handling and fallback for non-streaming scenarios
- Add optional editMessage to IChatProvider interface
- Add getClient() accessor to MatrixService for streaming service
- Register MatrixStreamingService in BridgeModule
- Tests: 20 tests pass

Refs #383

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
This commit is contained in:
2026-02-15 02:34:36 -06:00
parent aa106a948a
commit 93cd31435b
6 changed files with 681 additions and 1 deletions

View File

@@ -2,6 +2,7 @@ import { Logger, Module } from "@nestjs/common";
import { DiscordService } from "./discord/discord.service"; import { DiscordService } from "./discord/discord.service";
import { MatrixService } from "./matrix/matrix.service"; import { MatrixService } from "./matrix/matrix.service";
import { MatrixRoomService } from "./matrix/matrix-room.service"; import { MatrixRoomService } from "./matrix/matrix-room.service";
import { MatrixStreamingService } from "./matrix/matrix-streaming.service";
import { CommandParserService } from "./parser/command-parser.service"; import { CommandParserService } from "./parser/command-parser.service";
import { StitcherModule } from "../stitcher/stitcher.module"; import { StitcherModule } from "../stitcher/stitcher.module";
import { CHAT_PROVIDERS } from "./bridge.constants"; import { CHAT_PROVIDERS } from "./bridge.constants";
@@ -31,6 +32,7 @@ const logger = new Logger("BridgeModule");
providers: [ providers: [
CommandParserService, CommandParserService,
MatrixRoomService, MatrixRoomService,
MatrixStreamingService,
DiscordService, DiscordService,
MatrixService, MatrixService,
{ {
@@ -57,6 +59,13 @@ const logger = new Logger("BridgeModule");
inject: [DiscordService, MatrixService], inject: [DiscordService, MatrixService],
}, },
], ],
exports: [DiscordService, MatrixService, MatrixRoomService, CommandParserService, CHAT_PROVIDERS], exports: [
DiscordService,
MatrixService,
MatrixRoomService,
MatrixStreamingService,
CommandParserService,
CHAT_PROVIDERS,
],
}) })
export class BridgeModule {} export class BridgeModule {}

View File

@@ -76,4 +76,17 @@ export interface IChatProvider {
* Parse a command from a message * Parse a command from a message
*/ */
parseCommand(message: ChatMessage): ChatCommand | null; parseCommand(message: ChatMessage): ChatCommand | null;
/**
* Edit an existing message in a channel.
*
* Optional method for providers that support message editing
* (e.g., Matrix via m.replace, Discord via message.edit).
* Used for streaming AI responses with incremental updates.
*
* @param channelId - The channel/room ID
* @param messageId - The original message/event ID to edit
* @param content - The updated message content
*/
editMessage?(channelId: string, messageId: string, content: string): Promise<void>;
} }

View File

@@ -1,2 +1,4 @@
export { MatrixService } from "./matrix.service"; export { MatrixService } from "./matrix.service";
export { MatrixRoomService } from "./matrix-room.service"; export { MatrixRoomService } from "./matrix-room.service";
export { MatrixStreamingService } from "./matrix-streaming.service";
export type { StreamResponseOptions } from "./matrix-streaming.service";

View File

@@ -0,0 +1,408 @@
import { Test, TestingModule } from "@nestjs/testing";
import { MatrixStreamingService } from "./matrix-streaming.service";
import { MatrixService } from "./matrix.service";
import { vi, describe, it, expect, beforeEach, afterEach } from "vitest";
import type { StreamResponseOptions } from "./matrix-streaming.service";
// Mock matrix-bot-sdk to prevent native module loading
vi.mock("matrix-bot-sdk", () => {
return {
MatrixClient: class MockMatrixClient {},
SimpleFsStorageProvider: class MockStorageProvider {
constructor(_filename: string) {
// No-op for testing
}
},
AutojoinRoomsMixin: {
setupOnClient: vi.fn(),
},
};
});
// Mock MatrixClient
const mockClient = {
sendMessage: vi.fn().mockResolvedValue("$initial-event-id"),
sendEvent: vi.fn().mockResolvedValue("$edit-event-id"),
setTyping: vi.fn().mockResolvedValue(undefined),
};
// Mock MatrixService
const mockMatrixService = {
isConnected: vi.fn().mockReturnValue(true),
getClient: vi.fn().mockReturnValue(mockClient),
};
/**
* Helper: create an async iterable from an array of strings with optional delays
*/
async function* createTokenStream(
tokens: string[],
delayMs = 0
): AsyncGenerator<string, void, undefined> {
for (const token of tokens) {
if (delayMs > 0) {
await new Promise((resolve) => setTimeout(resolve, delayMs));
}
yield token;
}
}
/**
* Helper: create a token stream that throws an error mid-stream
*/
async function* createErrorStream(
tokens: string[],
errorAfter: number
): AsyncGenerator<string, void, undefined> {
let count = 0;
for (const token of tokens) {
if (count >= errorAfter) {
throw new Error("LLM provider connection lost");
}
yield token;
count++;
}
}
describe("MatrixStreamingService", () => {
let service: MatrixStreamingService;
beforeEach(async () => {
vi.useFakeTimers({ shouldAdvanceTime: true });
const module: TestingModule = await Test.createTestingModule({
providers: [
MatrixStreamingService,
{
provide: MatrixService,
useValue: mockMatrixService,
},
],
}).compile();
service = module.get<MatrixStreamingService>(MatrixStreamingService);
// Clear all mocks
vi.clearAllMocks();
// Re-apply default mock returns after clearing
mockMatrixService.isConnected.mockReturnValue(true);
mockMatrixService.getClient.mockReturnValue(mockClient);
mockClient.sendMessage.mockResolvedValue("$initial-event-id");
mockClient.sendEvent.mockResolvedValue("$edit-event-id");
mockClient.setTyping.mockResolvedValue(undefined);
});
afterEach(() => {
vi.useRealTimers();
});
describe("editMessage", () => {
it("should send a m.replace event to edit an existing message", async () => {
await service.editMessage("!room:example.com", "$original-event-id", "Updated content");
expect(mockClient.sendEvent).toHaveBeenCalledWith("!room:example.com", "m.room.message", {
"m.new_content": {
msgtype: "m.text",
body: "Updated content",
},
"m.relates_to": {
rel_type: "m.replace",
event_id: "$original-event-id",
},
// Fallback for clients that don't support edits
msgtype: "m.text",
body: "* Updated content",
});
});
it("should throw error when client is not connected", async () => {
mockMatrixService.isConnected.mockReturnValue(false);
await expect(
service.editMessage("!room:example.com", "$event-id", "content")
).rejects.toThrow("Matrix client is not connected");
});
it("should throw error when client is null", async () => {
mockMatrixService.getClient.mockReturnValue(null);
await expect(
service.editMessage("!room:example.com", "$event-id", "content")
).rejects.toThrow("Matrix client is not connected");
});
});
describe("setTypingIndicator", () => {
it("should call client.setTyping with true and timeout", async () => {
await service.setTypingIndicator("!room:example.com", true);
expect(mockClient.setTyping).toHaveBeenCalledWith("!room:example.com", true, 30000);
});
it("should call client.setTyping with false to clear indicator", async () => {
await service.setTypingIndicator("!room:example.com", false);
expect(mockClient.setTyping).toHaveBeenCalledWith("!room:example.com", false, undefined);
});
it("should throw error when client is not connected", async () => {
mockMatrixService.isConnected.mockReturnValue(false);
await expect(service.setTypingIndicator("!room:example.com", true)).rejects.toThrow(
"Matrix client is not connected"
);
});
});
describe("sendStreamingMessage", () => {
it("should send an initial message and return the event ID", async () => {
const eventId = await service.sendStreamingMessage("!room:example.com", "Thinking...");
expect(eventId).toBe("$initial-event-id");
expect(mockClient.sendMessage).toHaveBeenCalledWith("!room:example.com", {
msgtype: "m.text",
body: "Thinking...",
});
});
it("should send a thread message when threadId is provided", async () => {
const eventId = await service.sendStreamingMessage(
"!room:example.com",
"Thinking...",
"$thread-root-id"
);
expect(eventId).toBe("$initial-event-id");
expect(mockClient.sendMessage).toHaveBeenCalledWith("!room:example.com", {
msgtype: "m.text",
body: "Thinking...",
"m.relates_to": {
rel_type: "m.thread",
event_id: "$thread-root-id",
is_falling_back: true,
"m.in_reply_to": {
event_id: "$thread-root-id",
},
},
});
});
it("should throw error when client is not connected", async () => {
mockMatrixService.isConnected.mockReturnValue(false);
await expect(service.sendStreamingMessage("!room:example.com", "Test")).rejects.toThrow(
"Matrix client is not connected"
);
});
});
describe("streamResponse", () => {
it("should send initial 'Thinking...' message and start typing indicator", async () => {
vi.useRealTimers();
const tokens = ["Hello", " world"];
const stream = createTokenStream(tokens);
await service.streamResponse("!room:example.com", stream);
// Should have sent initial message
expect(mockClient.sendMessage).toHaveBeenCalledWith(
"!room:example.com",
expect.objectContaining({
msgtype: "m.text",
body: "Thinking...",
})
);
// Should have started typing indicator
expect(mockClient.setTyping).toHaveBeenCalledWith("!room:example.com", true, 30000);
});
it("should use custom initial message when provided", async () => {
vi.useRealTimers();
const tokens = ["Hi"];
const stream = createTokenStream(tokens);
const options: StreamResponseOptions = { initialMessage: "Processing..." };
await service.streamResponse("!room:example.com", stream, options);
expect(mockClient.sendMessage).toHaveBeenCalledWith(
"!room:example.com",
expect.objectContaining({
body: "Processing...",
})
);
});
it("should edit message with accumulated tokens on completion", async () => {
vi.useRealTimers();
const tokens = ["Hello", " ", "world", "!"];
const stream = createTokenStream(tokens);
await service.streamResponse("!room:example.com", stream);
// The final edit should contain the full accumulated text
const sendEventCalls = mockClient.sendEvent.mock.calls;
const lastEditCall = sendEventCalls[sendEventCalls.length - 1];
expect(lastEditCall).toBeDefined();
// eslint-disable-next-line @typescript-eslint/no-unsafe-member-access
expect(lastEditCall[2]["m.new_content"].body).toBe("Hello world!");
});
it("should clear typing indicator on completion", async () => {
vi.useRealTimers();
const tokens = ["Done"];
const stream = createTokenStream(tokens);
await service.streamResponse("!room:example.com", stream);
// Last setTyping call should be false
const typingCalls = mockClient.setTyping.mock.calls;
const lastTypingCall = typingCalls[typingCalls.length - 1];
expect(lastTypingCall).toEqual(["!room:example.com", false, undefined]);
});
it("should rate-limit edits to at most one every 500ms", async () => {
vi.useRealTimers();
// Send tokens with small delays - all within one 500ms window
const tokens = ["a", "b", "c", "d", "e"];
const stream = createTokenStream(tokens, 50); // 50ms between tokens = 250ms total
await service.streamResponse("!room:example.com", stream);
// With 250ms total streaming time (5 tokens * 50ms), all tokens arrive
// within one 500ms window. We expect at most 1 intermediate edit + 1 final edit,
// or just the final edit. The key point is that there should NOT be 5 separate edits.
const editCalls = mockClient.sendEvent.mock.calls.filter(
(call) => call[1] === "m.room.message"
);
// Should have fewer edits than tokens (rate limiting in effect)
expect(editCalls.length).toBeLessThanOrEqual(2);
// Should have at least the final edit
expect(editCalls.length).toBeGreaterThanOrEqual(1);
});
it("should handle errors gracefully and edit message with error notice", async () => {
vi.useRealTimers();
const stream = createErrorStream(["Hello", " ", "world"], 2);
await service.streamResponse("!room:example.com", stream);
// Should edit message with error content
const sendEventCalls = mockClient.sendEvent.mock.calls;
const lastEditCall = sendEventCalls[sendEventCalls.length - 1];
expect(lastEditCall).toBeDefined();
// eslint-disable-next-line @typescript-eslint/no-unsafe-member-access
const finalBody = lastEditCall[2]["m.new_content"].body as string;
expect(finalBody).toContain("error");
// Should clear typing on error
const typingCalls = mockClient.setTyping.mock.calls;
const lastTypingCall = typingCalls[typingCalls.length - 1];
expect(lastTypingCall).toEqual(["!room:example.com", false, undefined]);
});
it("should include token usage in final message when provided", async () => {
vi.useRealTimers();
const tokens = ["Hello"];
const stream = createTokenStream(tokens);
const options: StreamResponseOptions = {
showTokenUsage: true,
tokenUsage: { prompt: 10, completion: 5, total: 15 },
};
await service.streamResponse("!room:example.com", stream, options);
const sendEventCalls = mockClient.sendEvent.mock.calls;
const lastEditCall = sendEventCalls[sendEventCalls.length - 1];
expect(lastEditCall).toBeDefined();
// eslint-disable-next-line @typescript-eslint/no-unsafe-member-access
const finalBody = lastEditCall[2]["m.new_content"].body as string;
expect(finalBody).toContain("15");
});
it("should throw error when client is not connected", async () => {
mockMatrixService.isConnected.mockReturnValue(false);
const stream = createTokenStream(["test"]);
await expect(service.streamResponse("!room:example.com", stream)).rejects.toThrow(
"Matrix client is not connected"
);
});
it("should handle empty token stream", async () => {
vi.useRealTimers();
const stream = createTokenStream([]);
await service.streamResponse("!room:example.com", stream);
// Should still send initial message
expect(mockClient.sendMessage).toHaveBeenCalled();
// Should edit with empty/no-content message
const sendEventCalls = mockClient.sendEvent.mock.calls;
expect(sendEventCalls.length).toBeGreaterThanOrEqual(1);
// Should clear typing
const typingCalls = mockClient.setTyping.mock.calls;
const lastTypingCall = typingCalls[typingCalls.length - 1];
expect(lastTypingCall).toEqual(["!room:example.com", false, undefined]);
});
it("should support thread context in streamResponse", async () => {
vi.useRealTimers();
const tokens = ["Reply"];
const stream = createTokenStream(tokens);
const options: StreamResponseOptions = { threadId: "$thread-root" };
await service.streamResponse("!room:example.com", stream, options);
// Initial message should include thread relation
expect(mockClient.sendMessage).toHaveBeenCalledWith(
"!room:example.com",
expect.objectContaining({
"m.relates_to": expect.objectContaining({
rel_type: "m.thread",
event_id: "$thread-root",
}),
})
);
});
it("should perform multiple edits for long-running streams", async () => {
vi.useRealTimers();
// Create tokens with 200ms delays - total ~2000ms, should get multiple edit windows
const tokens = Array.from({ length: 10 }, (_, i) => `token${String(i)} `);
const stream = createTokenStream(tokens, 200);
await service.streamResponse("!room:example.com", stream);
// With 10 tokens at 200ms each = 2000ms total, at 500ms intervals
// we expect roughly 3-4 intermediate edits + 1 final = 4-5 total
const editCalls = mockClient.sendEvent.mock.calls.filter(
(call) => call[1] === "m.room.message"
);
// Should have multiple edits (at least 2) but far fewer than 10
expect(editCalls.length).toBeGreaterThanOrEqual(2);
expect(editCalls.length).toBeLessThanOrEqual(8);
});
});
});

View File

@@ -0,0 +1,236 @@
import { Injectable, Logger } from "@nestjs/common";
import type { MatrixClient } from "matrix-bot-sdk";
import { MatrixService } from "./matrix.service";
/**
* Options for the streamResponse method
*/
export interface StreamResponseOptions {
/** Custom initial message (defaults to "Thinking...") */
initialMessage?: string;
/** Thread root event ID for threaded responses */
threadId?: string;
/** Whether to show token usage in the final message */
showTokenUsage?: boolean;
/** Token usage stats to display in the final message */
tokenUsage?: { prompt: number; completion: number; total: number };
}
/**
* Matrix message content for m.room.message events
*/
interface MatrixMessageContent {
msgtype: string;
body: string;
"m.new_content"?: {
msgtype: string;
body: string;
};
"m.relates_to"?: {
rel_type: string;
event_id: string;
is_falling_back?: boolean;
"m.in_reply_to"?: {
event_id: string;
};
};
}
/** Minimum interval between message edits (milliseconds) */
const EDIT_INTERVAL_MS = 500;
/** Typing indicator timeout (milliseconds) */
const TYPING_TIMEOUT_MS = 30000;
/**
* Matrix Streaming Service
*
* Provides streaming AI response capabilities for Matrix rooms using
* incremental message edits. Tokens from an LLM are buffered and the
* response message is edited at rate-limited intervals, providing a
* smooth streaming experience without excessive API calls.
*
* Key features:
* - Rate-limited edits (max every 500ms)
* - Typing indicator management during generation
* - Graceful error handling with user-visible error notices
* - Thread support for contextual responses
* - LLM-agnostic design via AsyncIterable<string> token stream
*/
@Injectable()
export class MatrixStreamingService {
private readonly logger = new Logger(MatrixStreamingService.name);
constructor(private readonly matrixService: MatrixService) {}
/**
* Edit an existing Matrix message using the m.replace relation.
*
* Sends a new event that replaces the content of an existing message.
* Includes fallback content for clients that don't support edits.
*
* @param roomId - The Matrix room ID
* @param eventId - The original event ID to replace
* @param newContent - The updated message text
*/
async editMessage(roomId: string, eventId: string, newContent: string): Promise<void> {
const client = this.getClientOrThrow();
const editContent: MatrixMessageContent = {
"m.new_content": {
msgtype: "m.text",
body: newContent,
},
"m.relates_to": {
rel_type: "m.replace",
event_id: eventId,
},
// Fallback for clients that don't support edits
msgtype: "m.text",
body: `* ${newContent}`,
};
await client.sendEvent(roomId, "m.room.message", editContent);
}
/**
* Set the typing indicator for the bot in a room.
*
* @param roomId - The Matrix room ID
* @param typing - Whether the bot is typing
*/
async setTypingIndicator(roomId: string, typing: boolean): Promise<void> {
const client = this.getClientOrThrow();
await client.setTyping(roomId, typing, typing ? TYPING_TIMEOUT_MS : undefined);
}
/**
* Send an initial message for streaming, optionally in a thread.
*
* Returns the event ID of the sent message, which can be used for
* subsequent edits via editMessage.
*
* @param roomId - The Matrix room ID
* @param content - The initial message content
* @param threadId - Optional thread root event ID
* @returns The event ID of the sent message
*/
async sendStreamingMessage(roomId: string, content: string, threadId?: string): Promise<string> {
const client = this.getClientOrThrow();
const messageContent: MatrixMessageContent = {
msgtype: "m.text",
body: content,
};
if (threadId) {
messageContent["m.relates_to"] = {
rel_type: "m.thread",
event_id: threadId,
is_falling_back: true,
"m.in_reply_to": {
event_id: threadId,
},
};
}
const eventId: string = await client.sendMessage(roomId, messageContent);
return eventId;
}
/**
* Stream an AI response to a Matrix room using incremental message edits.
*
* This is the main streaming method. It:
* 1. Sends an initial "Thinking..." message
* 2. Starts the typing indicator
* 3. Buffers incoming tokens from the async iterable
* 4. Edits the message every 500ms with accumulated text
* 5. On completion: sends a final clean edit, clears typing
* 6. On error: edits message with error notice, clears typing
*
* @param roomId - The Matrix room ID
* @param tokenStream - AsyncIterable that yields string tokens
* @param options - Optional configuration for the stream
*/
async streamResponse(
roomId: string,
tokenStream: AsyncIterable<string>,
options?: StreamResponseOptions
): Promise<void> {
// Validate connection before starting
this.getClientOrThrow();
const initialMessage = options?.initialMessage ?? "Thinking...";
const threadId = options?.threadId;
// Step 1: Send initial message
const eventId = await this.sendStreamingMessage(roomId, initialMessage, threadId);
// Step 2: Start typing indicator
await this.setTypingIndicator(roomId, true);
// Step 3: Buffer and stream tokens
let accumulatedText = "";
let lastEditTime = 0;
let hasError = false;
try {
for await (const token of tokenStream) {
accumulatedText += token;
const now = Date.now();
const elapsed = now - lastEditTime;
if (elapsed >= EDIT_INTERVAL_MS && accumulatedText.length > 0) {
await this.editMessage(roomId, eventId, accumulatedText);
lastEditTime = now;
}
}
} catch (error: unknown) {
hasError = true;
const errorMessage = error instanceof Error ? error.message : "Unknown error occurred";
this.logger.error(`Stream error in room ${roomId}: ${errorMessage}`);
// Edit message to show error
const errorContent = accumulatedText
? `${accumulatedText}\n\n[Streaming error: ${errorMessage}]`
: `[Streaming error: ${errorMessage}]`;
await this.editMessage(roomId, eventId, errorContent);
} finally {
// Step 4: Clear typing indicator
await this.setTypingIndicator(roomId, false);
}
// Step 5: Final edit with clean output (if no error)
if (!hasError) {
let finalContent = accumulatedText || "(No response generated)";
if (options?.showTokenUsage && options.tokenUsage) {
const { prompt, completion, total } = options.tokenUsage;
finalContent += `\n\n---\nTokens: ${String(total)} (prompt: ${String(prompt)}, completion: ${String(completion)})`;
}
await this.editMessage(roomId, eventId, finalContent);
}
}
/**
* Get the Matrix client from the parent MatrixService, or throw if not connected.
*/
private getClientOrThrow(): MatrixClient {
if (!this.matrixService.isConnected()) {
throw new Error("Matrix client is not connected");
}
const client = this.matrixService.getClient();
if (!client) {
throw new Error("Matrix client is not connected");
}
return client;
}
}

View File

@@ -269,6 +269,18 @@ export class MatrixService implements IChatProvider {
return this.connected; return this.connected;
} }
/**
* Get the underlying MatrixClient instance.
*
* Used by MatrixStreamingService for low-level operations
* (message edits, typing indicators) that require direct client access.
*
* @returns The MatrixClient instance, or null if not connected
*/
getClient(): MatrixClient | null {
return this.client;
}
/** /**
* Send a message to a room * Send a message to a room
*/ */