Files
openclaw/extensions/msteams/src/reply-dispatcher.ts
sudie-codes 99f76ec4c6 fix(msteams): keep streaming alive during long tool chains via typing indicator (#59731) (#64088)
* fix(msteams): keep streaming alive during long tool chains via periodic typing (#59731)

* test(msteams): align thread-session store mock with interface

* fix(msteams): treat failed streams as inactive

---------

Co-authored-by: Brad Groux <bradgroux@users.noreply.github.com>
Co-authored-by: Brad Groux <3053586+BradGroux@users.noreply.github.com>
2026-04-10 14:42:41 -05:00

347 lines
12 KiB
TypeScript

import { normalizeOptionalLowercaseString } from "openclaw/plugin-sdk/text-runtime";
import {
createChannelReplyPipeline,
logTypingFailure,
resolveChannelMediaMaxBytes,
type OpenClawConfig,
type MSTeamsReplyStyle,
type RuntimeEnv,
} from "../runtime-api.js";
import type { MSTeamsAccessTokenProvider } from "./attachments/types.js";
import type { StoredConversationReference } from "./conversation-store.js";
import {
classifyMSTeamsSendError,
formatMSTeamsSendErrorHint,
formatUnknownError,
} from "./errors.js";
import {
buildConversationReference,
type MSTeamsAdapter,
type MSTeamsRenderedMessage,
renderReplyPayloadsToMessages,
sendMSTeamsMessages,
} from "./messenger.js";
import type { MSTeamsMonitorLogger } from "./monitor-types.js";
import { createTeamsReplyStreamController } from "./reply-stream-controller.js";
import { withRevokedProxyFallback } from "./revoked-context.js";
import { getMSTeamsRuntime } from "./runtime.js";
import type { MSTeamsTurnContext } from "./sdk-types.js";
export { pickInformativeStatusText } from "./reply-stream-controller.js";
export function createMSTeamsReplyDispatcher(params: {
cfg: OpenClawConfig;
agentId: string;
sessionKey: string;
accountId?: string;
runtime: RuntimeEnv;
log: MSTeamsMonitorLogger;
adapter: MSTeamsAdapter;
appId: string;
conversationRef: StoredConversationReference;
context: MSTeamsTurnContext;
replyStyle: MSTeamsReplyStyle;
textLimit: number;
onSentMessageIds?: (ids: string[]) => void;
tokenProvider?: MSTeamsAccessTokenProvider;
sharePointSiteId?: string;
}) {
const core = getMSTeamsRuntime();
const msteamsCfg = params.cfg.channels?.msteams;
const conversationType = normalizeOptionalLowercaseString(
params.conversationRef.conversation?.conversationType,
);
const isTypingSupported = conversationType === "personal" || conversationType === "groupchat";
/**
* Keepalive cadence for the typing indicator while the bot is running
* (including long tool chains). Bot Framework 1:1 TurnContext proxies
* expire after ~30s of inactivity; sending a typing activity every 8s
* keeps the proxy alive so the post-tool reply can still land via the
* turn context. Sits in the middle of the 5-10s range recommended in
* #59731.
*/
const TYPING_KEEPALIVE_INTERVAL_MS = 8_000;
/**
* TTL ceiling for the typing keepalive loop. The default in
* createTypingCallbacks is 60s, which is too short for the Teams long tool
* chains described in #59731 (60s+ total runs are common). Give tool
* chains up to 10 minutes before auto-stopping the keepalive.
*/
const TYPING_KEEPALIVE_MAX_DURATION_MS = 10 * 60_000;
// Forward reference: sendTypingIndicator is built before the stream
// controller exists, but the keepalive tick needs to check stream state so
// we don't overlay "..." typing on the visible streaming card. The ref is
// wired once the stream controller is constructed below.
const streamActiveRef: { current: () => boolean } = { current: () => false };
const rawSendTypingIndicator = async () => {
await withRevokedProxyFallback({
run: async () => {
await params.context.sendActivity({ type: "typing" });
},
onRevoked: async () => {
const baseRef = buildConversationReference(params.conversationRef);
await params.adapter.continueConversation(
params.appId,
{ ...baseRef, activityId: undefined },
async (ctx) => {
await ctx.sendActivity({ type: "typing" });
},
);
},
onRevokedLog: () => {
params.log.debug?.("turn context revoked, sending typing via proactive messaging");
},
});
};
const sendTypingIndicator = isTypingSupported
? async () => {
// While the streaming card is actively being updated the user
// already sees a live indicator in the stream — don't overlay a
// plain "..." typing on top of it. Between segments (tool chain)
// the stream is finalized, so typing indicators are appropriate
// and they are what keep the TurnContext alive. See #59731.
if (streamActiveRef.current()) {
return;
}
await rawSendTypingIndicator();
}
: async () => {};
const { onModelSelected, typingCallbacks, ...replyPipeline } = createChannelReplyPipeline({
cfg: params.cfg,
agentId: params.agentId,
channel: "msteams",
accountId: params.accountId,
typing: {
start: sendTypingIndicator,
keepaliveIntervalMs: TYPING_KEEPALIVE_INTERVAL_MS,
maxDurationMs: TYPING_KEEPALIVE_MAX_DURATION_MS,
onStartError: (err) => {
logTypingFailure({
log: (message) => params.log.debug?.(message),
channel: "msteams",
action: "start",
error: err,
});
},
},
});
const chunkMode = core.channel.text.resolveChunkMode(params.cfg, "msteams");
const tableMode = core.channel.text.resolveMarkdownTableMode({
cfg: params.cfg,
channel: "msteams",
});
const mediaMaxBytes = resolveChannelMediaMaxBytes({
cfg: params.cfg,
resolveChannelLimitMb: ({ cfg }) => cfg.channels?.msteams?.mediaMaxMb,
});
const feedbackLoopEnabled = params.cfg.channels?.msteams?.feedbackEnabled !== false;
const streamController = createTeamsReplyStreamController({
conversationType,
context: params.context,
feedbackLoopEnabled,
log: params.log,
});
// Wire the forward-declared gate used by sendTypingIndicator.
streamActiveRef.current = () => streamController.isStreamActive();
const blockStreamingEnabled =
typeof msteamsCfg?.blockStreaming === "boolean" ? msteamsCfg.blockStreaming : false;
const typingIndicatorEnabled =
typeof msteamsCfg?.typingIndicator === "boolean" ? msteamsCfg.typingIndicator : true;
const pendingMessages: MSTeamsRenderedMessage[] = [];
const sendMessages = async (messages: MSTeamsRenderedMessage[]): Promise<string[]> => {
return sendMSTeamsMessages({
replyStyle: params.replyStyle,
adapter: params.adapter,
appId: params.appId,
conversationRef: params.conversationRef,
context: params.context,
messages,
retry: {},
onRetry: (event) => {
params.log.debug?.("retrying send", {
replyStyle: params.replyStyle,
...event,
});
},
tokenProvider: params.tokenProvider,
sharePointSiteId: params.sharePointSiteId,
mediaMaxBytes,
feedbackLoopEnabled,
});
};
const queueDeliveryFailureSystemEvent = (failure: {
failed: number;
total: number;
error: unknown;
}) => {
const classification = classifyMSTeamsSendError(failure.error);
const errorText = formatUnknownError(failure.error);
const failedAll = failure.failed >= failure.total;
const summary = failedAll
? "the previous reply was not delivered"
: `${failure.failed} of ${failure.total} message blocks were not delivered`;
const sentences = [
`Microsoft Teams delivery failed: ${summary}.`,
`The user may not have received ${failedAll ? "that reply" : "the full reply"}.`,
`Error: ${errorText}.`,
classification.statusCode != null ? `Status: ${classification.statusCode}.` : undefined,
classification.kind === "transient" || classification.kind === "throttled"
? "Retrying later may succeed."
: undefined,
].filter(Boolean);
core.system.enqueueSystemEvent(sentences.join(" "), {
sessionKey: params.sessionKey,
contextKey: `msteams:delivery-failure:${params.conversationRef.conversation?.id ?? "unknown"}`,
});
};
const flushPendingMessages = async () => {
if (pendingMessages.length === 0) {
return;
}
const toSend = pendingMessages.splice(0);
const total = toSend.length;
let ids: string[];
try {
ids = await sendMessages(toSend);
} catch (batchError) {
ids = [];
let failed = 0;
let lastFailedError: unknown = batchError;
for (const msg of toSend) {
try {
const msgIds = await sendMessages([msg]);
ids.push(...msgIds);
} catch (msgError) {
failed += 1;
lastFailedError = msgError;
params.log.debug?.("individual message send failed, continuing with remaining blocks");
}
}
if (failed > 0) {
params.log.warn?.(`failed to deliver ${failed} of ${total} message blocks`, {
failed,
total,
});
queueDeliveryFailureSystemEvent({
failed,
total,
error: lastFailedError,
});
}
}
if (ids.length > 0) {
params.onSentMessageIds?.(ids);
}
};
const {
dispatcher,
replyOptions,
markDispatchIdle: baseMarkDispatchIdle,
} = core.channel.reply.createReplyDispatcherWithTyping({
...replyPipeline,
humanDelay: core.channel.reply.resolveHumanDelayConfig(params.cfg, params.agentId),
onReplyStart: async () => {
await streamController.onReplyStart();
// Always start the typing keepalive loop when typing is enabled and
// supported by this conversation type. The sendTypingIndicator gate
// skips actual sends while the stream card is visually active, so
// during the first text segment the user only sees the streaming UI.
// Once the stream finalizes (between segments / during tool chains),
// the loop starts sending typing activities and keeps the Bot Framework
// TurnContext alive so the post-tool reply can still land. See #59731.
if (typingIndicatorEnabled) {
await typingCallbacks?.onReplyStart?.();
}
},
typingCallbacks,
deliver: async (payload) => {
const preparedPayload = streamController.preparePayload(payload);
if (!preparedPayload) {
return;
}
const messages = renderReplyPayloadsToMessages([preparedPayload], {
textChunkLimit: params.textLimit,
chunkText: true,
mediaMode: "split",
tableMode,
chunkMode,
});
pendingMessages.push(...messages);
// When block streaming is enabled, flush immediately so blocks are
// delivered progressively instead of batching until markDispatchIdle.
if (blockStreamingEnabled) {
await flushPendingMessages();
}
},
onError: (err, info) => {
const errMsg = formatUnknownError(err);
const classification = classifyMSTeamsSendError(err);
const hint = formatMSTeamsSendErrorHint(classification);
params.runtime.error?.(
`msteams ${info.kind} reply failed: ${errMsg}${hint ? ` (${hint})` : ""}`,
);
params.log.error("reply failed", {
kind: info.kind,
error: errMsg,
classification,
hint,
});
},
});
const markDispatchIdle = (): Promise<void> => {
return flushPendingMessages()
.catch((err) => {
const errMsg = formatUnknownError(err);
const classification = classifyMSTeamsSendError(err);
const hint = formatMSTeamsSendErrorHint(classification);
params.runtime.error?.(`msteams flush reply failed: ${errMsg}${hint ? ` (${hint})` : ""}`);
params.log.error("flush reply failed", {
error: errMsg,
classification,
hint,
});
})
.then(() => {
return streamController.finalize().catch((err) => {
params.log.debug?.("stream finalize failed", { error: formatUnknownError(err) });
});
})
.finally(() => {
baseMarkDispatchIdle();
});
};
return {
dispatcher,
replyOptions: {
...replyOptions,
...(streamController.hasStream()
? {
onPartialReply: (payload: { text?: string }) =>
streamController.onPartialReply(payload),
}
: {}),
disableBlockStreaming:
typeof msteamsCfg?.blockStreaming === "boolean" ? !msteamsCfg.blockStreaming : undefined,
onModelSelected,
},
markDispatchIdle,
};
}