Files
openclaw/src/telegram/bot-message-dispatch.ts
OpenCils 3fe4c19305 fix(telegram): prevent duplicate messages in DM draft streaming mode (#32118)
* fix(telegram): prevent duplicate messages in DM draft streaming mode

When using sendMessageDraft for DM streaming (streaming: 'partial'),
the draft bubble auto-converts to the final message. The code was
incorrectly falling through to sendPayload() after the draft was
finalized, causing a duplicate message.

This fix checks if we're in draft preview mode with hasStreamedMessage
and skips the sendPayload call, returning "preview-finalized" directly.

Key changes:
- Use hasStreamedMessage flag instead of previewRevision comparison
- Avoids double stopDraftLane calls by returning early
- Prevents duplicate messages when final text equals last streamed text

Root cause: In lane-delivery.ts, the final message handling logic
did not properly handle the DM draft flow where sendMessageDraft
creates a transient bubble that doesn't need a separate final send.

* fix(telegram): harden DM draft finalization path

* fix(telegram): require emitted draft preview for unchanged finals

* fix(telegram): require final draft text emission before finalize

* fix: update changelog for telegram draft finalization (#32118) (thanks @OpenCils)

---------

Co-authored-by: Ayaan Zaidi <zaidi@uplause.io>
2026-03-03 17:34:46 +05:30

733 lines
25 KiB
TypeScript

import type { Bot } from "grammy";
import { resolveAgentDir } from "../agents/agent-scope.js";
import {
findModelInCatalog,
loadModelCatalog,
modelSupportsVision,
} from "../agents/model-catalog.js";
import { resolveDefaultModelForAgent } from "../agents/model-selection.js";
import { resolveChunkMode } from "../auto-reply/chunk.js";
import { clearHistoryEntriesIfEnabled } from "../auto-reply/reply/history.js";
import { dispatchReplyWithBufferedBlockDispatcher } from "../auto-reply/reply/provider-dispatcher.js";
import type { ReplyPayload } from "../auto-reply/types.js";
import { removeAckReactionAfterReply } from "../channels/ack-reactions.js";
import { logAckFailure, logTypingFailure } from "../channels/logging.js";
import { createReplyPrefixOptions } from "../channels/reply-prefix.js";
import { createTypingCallbacks } from "../channels/typing.js";
import { resolveMarkdownTableMode } from "../config/markdown-tables.js";
import { loadSessionStore, resolveStorePath } from "../config/sessions.js";
import type { OpenClawConfig, ReplyToMode, TelegramAccountConfig } from "../config/types.js";
import { danger, logVerbose } from "../globals.js";
import { getAgentScopedMediaLocalRoots } from "../media/local-roots.js";
import type { RuntimeEnv } from "../runtime.js";
import type { TelegramMessageContext } from "./bot-message-context.js";
import type { TelegramBotOptions } from "./bot.js";
import { deliverReplies } from "./bot/delivery.js";
import type { TelegramStreamMode } from "./bot/types.js";
import type { TelegramInlineButtons } from "./button-types.js";
import { createTelegramDraftStream } from "./draft-stream.js";
import { renderTelegramHtmlText } from "./format.js";
import {
type ArchivedPreview,
createLaneDeliveryStateTracker,
createLaneTextDeliverer,
type DraftLaneState,
type LaneName,
} from "./lane-delivery.js";
import {
createTelegramReasoningStepState,
splitTelegramReasoningText,
} from "./reasoning-lane-coordinator.js";
import { editMessageTelegram } from "./send.js";
import { cacheSticker, describeStickerImage } from "./sticker-cache.js";
const EMPTY_RESPONSE_FALLBACK = "No response generated. Please try again.";
/** Minimum chars before sending first streaming message (improves push notification UX) */
const DRAFT_MIN_INITIAL_CHARS = 30;
async function resolveStickerVisionSupport(cfg: OpenClawConfig, agentId: string) {
try {
const catalog = await loadModelCatalog({ config: cfg });
const defaultModel = resolveDefaultModelForAgent({ cfg, agentId });
const entry = findModelInCatalog(catalog, defaultModel.provider, defaultModel.model);
if (!entry) {
return false;
}
return modelSupportsVision(entry);
} catch {
return false;
}
}
export function pruneStickerMediaFromContext(
ctxPayload: {
MediaPath?: string;
MediaUrl?: string;
MediaType?: string;
MediaPaths?: string[];
MediaUrls?: string[];
MediaTypes?: string[];
},
opts?: { stickerMediaIncluded?: boolean },
) {
if (opts?.stickerMediaIncluded === false) {
return;
}
const nextMediaPaths = Array.isArray(ctxPayload.MediaPaths)
? ctxPayload.MediaPaths.slice(1)
: undefined;
const nextMediaUrls = Array.isArray(ctxPayload.MediaUrls)
? ctxPayload.MediaUrls.slice(1)
: undefined;
const nextMediaTypes = Array.isArray(ctxPayload.MediaTypes)
? ctxPayload.MediaTypes.slice(1)
: undefined;
ctxPayload.MediaPaths = nextMediaPaths && nextMediaPaths.length > 0 ? nextMediaPaths : undefined;
ctxPayload.MediaUrls = nextMediaUrls && nextMediaUrls.length > 0 ? nextMediaUrls : undefined;
ctxPayload.MediaTypes = nextMediaTypes && nextMediaTypes.length > 0 ? nextMediaTypes : undefined;
ctxPayload.MediaPath = ctxPayload.MediaPaths?.[0];
ctxPayload.MediaUrl = ctxPayload.MediaUrls?.[0] ?? ctxPayload.MediaPath;
ctxPayload.MediaType = ctxPayload.MediaTypes?.[0];
}
type DispatchTelegramMessageParams = {
context: TelegramMessageContext;
bot: Bot;
cfg: OpenClawConfig;
runtime: RuntimeEnv;
replyToMode: ReplyToMode;
streamMode: TelegramStreamMode;
textLimit: number;
telegramCfg: TelegramAccountConfig;
opts: Pick<TelegramBotOptions, "token">;
};
type TelegramReasoningLevel = "off" | "on" | "stream";
function resolveTelegramReasoningLevel(params: {
cfg: OpenClawConfig;
sessionKey?: string;
agentId: string;
}): TelegramReasoningLevel {
const { cfg, sessionKey, agentId } = params;
if (!sessionKey) {
return "off";
}
try {
const storePath = resolveStorePath(cfg.session?.store, { agentId });
const store = loadSessionStore(storePath, { skipCache: true });
const entry = store[sessionKey.toLowerCase()] ?? store[sessionKey];
const level = entry?.reasoningLevel;
if (level === "on" || level === "stream") {
return level;
}
} catch {
// Fall through to default.
}
return "off";
}
export const dispatchTelegramMessage = async ({
context,
bot,
cfg,
runtime,
replyToMode,
streamMode,
textLimit,
telegramCfg,
opts,
}: DispatchTelegramMessageParams) => {
const {
ctxPayload,
msg,
chatId,
isGroup,
threadSpec,
historyKey,
historyLimit,
groupHistories,
route,
skillFilter,
sendTyping,
sendRecordVoice,
ackReactionPromise,
reactionApi,
removeAckAfterReply,
statusReactionController,
} = context;
const draftMaxChars = Math.min(textLimit, 4096);
const tableMode = resolveMarkdownTableMode({
cfg,
channel: "telegram",
accountId: route.accountId,
});
const renderDraftPreview = (text: string) => ({
text: renderTelegramHtmlText(text, { tableMode }),
parseMode: "HTML" as const,
});
const accountBlockStreamingEnabled =
typeof telegramCfg.blockStreaming === "boolean"
? telegramCfg.blockStreaming
: cfg.agents?.defaults?.blockStreamingDefault === "on";
const resolvedReasoningLevel = resolveTelegramReasoningLevel({
cfg,
sessionKey: ctxPayload.SessionKey,
agentId: route.agentId,
});
const forceBlockStreamingForReasoning = resolvedReasoningLevel === "on";
const streamReasoningDraft = resolvedReasoningLevel === "stream";
const previewStreamingEnabled = streamMode !== "off";
const canStreamAnswerDraft =
previewStreamingEnabled && !accountBlockStreamingEnabled && !forceBlockStreamingForReasoning;
const canStreamReasoningDraft = canStreamAnswerDraft || streamReasoningDraft;
const draftReplyToMessageId =
replyToMode !== "off" && typeof msg.message_id === "number" ? msg.message_id : undefined;
const draftMinInitialChars = DRAFT_MIN_INITIAL_CHARS;
const mediaLocalRoots = getAgentScopedMediaLocalRoots(cfg, route.agentId);
const archivedAnswerPreviews: ArchivedPreview[] = [];
const archivedReasoningPreviewIds: number[] = [];
const createDraftLane = (laneName: LaneName, enabled: boolean): DraftLaneState => {
const useMessagePreviewTransportForDmReasoning =
laneName === "reasoning" && threadSpec?.scope === "dm" && canStreamAnswerDraft;
const stream = enabled
? createTelegramDraftStream({
api: bot.api,
chatId,
maxChars: draftMaxChars,
thread: threadSpec,
previewTransport: useMessagePreviewTransportForDmReasoning ? "message" : "auto",
replyToMessageId: draftReplyToMessageId,
minInitialChars: draftMinInitialChars,
renderText: renderDraftPreview,
onSupersededPreview:
laneName === "answer" || laneName === "reasoning"
? (preview) => {
if (laneName === "reasoning") {
if (!archivedReasoningPreviewIds.includes(preview.messageId)) {
archivedReasoningPreviewIds.push(preview.messageId);
}
return;
}
archivedAnswerPreviews.push({
messageId: preview.messageId,
textSnapshot: preview.textSnapshot,
});
}
: undefined,
log: logVerbose,
warn: logVerbose,
})
: undefined;
return {
stream,
lastPartialText: "",
hasStreamedMessage: false,
previewRevisionBaseline: stream?.previewRevision?.() ?? 0,
};
};
const lanes: Record<LaneName, DraftLaneState> = {
answer: createDraftLane("answer", canStreamAnswerDraft),
reasoning: createDraftLane("reasoning", canStreamReasoningDraft),
};
const answerLane = lanes.answer;
const reasoningLane = lanes.reasoning;
let splitReasoningOnNextStream = false;
const reasoningStepState = createTelegramReasoningStepState();
type SplitLaneSegment = { lane: LaneName; text: string };
type SplitLaneSegmentsResult = {
segments: SplitLaneSegment[];
suppressedReasoningOnly: boolean;
};
const splitTextIntoLaneSegments = (text?: string): SplitLaneSegmentsResult => {
const split = splitTelegramReasoningText(text);
const segments: SplitLaneSegment[] = [];
const suppressReasoning = resolvedReasoningLevel === "off";
if (split.reasoningText && !suppressReasoning) {
segments.push({ lane: "reasoning", text: split.reasoningText });
}
if (split.answerText) {
segments.push({ lane: "answer", text: split.answerText });
}
return {
segments,
suppressedReasoningOnly:
Boolean(split.reasoningText) && suppressReasoning && !split.answerText,
};
};
const resetDraftLaneState = (lane: DraftLaneState) => {
lane.lastPartialText = "";
lane.hasStreamedMessage = false;
lane.previewRevisionBaseline = lane.stream?.previewRevision?.() ?? lane.previewRevisionBaseline;
};
const updateDraftFromPartial = (lane: DraftLaneState, text: string | undefined) => {
const laneStream = lane.stream;
if (!laneStream || !text) {
return;
}
if (text === lane.lastPartialText) {
return;
}
// Mark that we've received streaming content (for forceNewMessage decision).
lane.hasStreamedMessage = true;
// Some providers briefly emit a shorter prefix snapshot (for example
// "Sure." -> "Sure" -> "Sure."). Keep the longer preview to avoid
// visible punctuation flicker.
if (
lane.lastPartialText &&
lane.lastPartialText.startsWith(text) &&
text.length < lane.lastPartialText.length
) {
return;
}
lane.lastPartialText = text;
laneStream.update(text);
};
const ingestDraftLaneSegments = (text: string | undefined) => {
const split = splitTextIntoLaneSegments(text);
for (const segment of split.segments) {
if (segment.lane === "reasoning") {
reasoningStepState.noteReasoningHint();
reasoningStepState.noteReasoningDelivered();
}
updateDraftFromPartial(lanes[segment.lane], segment.text);
}
};
const flushDraftLane = async (lane: DraftLaneState) => {
if (!lane.stream) {
return;
}
await lane.stream.flush();
};
const disableBlockStreaming = !previewStreamingEnabled
? true
: forceBlockStreamingForReasoning
? false
: typeof telegramCfg.blockStreaming === "boolean"
? !telegramCfg.blockStreaming
: canStreamAnswerDraft
? true
: undefined;
const { onModelSelected, ...prefixOptions } = createReplyPrefixOptions({
cfg,
agentId: route.agentId,
channel: "telegram",
accountId: route.accountId,
});
const chunkMode = resolveChunkMode(cfg, "telegram", route.accountId);
// Handle uncached stickers: get a dedicated vision description before dispatch
// This ensures we cache a raw description rather than a conversational response
const sticker = ctxPayload.Sticker;
if (sticker?.fileId && sticker.fileUniqueId && ctxPayload.MediaPath) {
const agentDir = resolveAgentDir(cfg, route.agentId);
const stickerSupportsVision = await resolveStickerVisionSupport(cfg, route.agentId);
let description = sticker.cachedDescription ?? null;
if (!description) {
description = await describeStickerImage({
imagePath: ctxPayload.MediaPath,
cfg,
agentDir,
agentId: route.agentId,
});
}
if (description) {
// Format the description with sticker context
const stickerContext = [sticker.emoji, sticker.setName ? `from "${sticker.setName}"` : null]
.filter(Boolean)
.join(" ");
const formattedDesc = `[Sticker${stickerContext ? ` ${stickerContext}` : ""}] ${description}`;
sticker.cachedDescription = description;
if (!stickerSupportsVision) {
// Update context to use description instead of image
ctxPayload.Body = formattedDesc;
ctxPayload.BodyForAgent = formattedDesc;
// Drop only the sticker attachment; keep replied media context if present.
pruneStickerMediaFromContext(ctxPayload, {
stickerMediaIncluded: ctxPayload.StickerMediaIncluded,
});
}
// Cache the description for future encounters
if (sticker.fileId) {
cacheSticker({
fileId: sticker.fileId,
fileUniqueId: sticker.fileUniqueId,
emoji: sticker.emoji,
setName: sticker.setName,
description,
cachedAt: new Date().toISOString(),
receivedFrom: ctxPayload.From,
});
logVerbose(`telegram: cached sticker description for ${sticker.fileUniqueId}`);
} else {
logVerbose(`telegram: skipped sticker cache (missing fileId)`);
}
}
}
const replyQuoteText =
ctxPayload.ReplyToIsQuote && ctxPayload.ReplyToBody
? ctxPayload.ReplyToBody.trim() || undefined
: undefined;
const deliveryState = createLaneDeliveryStateTracker();
const finalizedPreviewByLane: Record<LaneName, boolean> = {
answer: false,
reasoning: false,
};
const clearGroupHistory = () => {
if (isGroup && historyKey) {
clearHistoryEntriesIfEnabled({ historyMap: groupHistories, historyKey, limit: historyLimit });
}
};
const deliveryBaseOptions = {
chatId: String(chatId),
accountId: route.accountId,
token: opts.token,
runtime,
bot,
mediaLocalRoots,
replyToMode,
textLimit,
thread: threadSpec,
tableMode,
chunkMode,
linkPreview: telegramCfg.linkPreview,
replyQuoteText,
};
const applyTextToPayload = (payload: ReplyPayload, text: string): ReplyPayload => {
if (payload.text === text) {
return payload;
}
return { ...payload, text };
};
const sendPayload = async (payload: ReplyPayload) => {
const result = await deliverReplies({
...deliveryBaseOptions,
replies: [payload],
onVoiceRecording: sendRecordVoice,
});
if (result.delivered) {
deliveryState.markDelivered();
}
return result.delivered;
};
const deliverLaneText = createLaneTextDeliverer({
lanes,
archivedAnswerPreviews,
finalizedPreviewByLane,
draftMaxChars,
applyTextToPayload,
sendPayload,
flushDraftLane,
stopDraftLane: async (lane) => {
await lane.stream?.stop();
},
editPreview: async ({ messageId, text, previewButtons }) => {
await editMessageTelegram(chatId, messageId, text, {
api: bot.api,
cfg,
accountId: route.accountId,
linkPreview: telegramCfg.linkPreview,
buttons: previewButtons,
});
},
deletePreviewMessage: async (messageId) => {
await bot.api.deleteMessage(chatId, messageId);
},
log: logVerbose,
markDelivered: () => {
deliveryState.markDelivered();
},
});
let queuedFinal = false;
if (statusReactionController) {
void statusReactionController.setThinking();
}
const typingCallbacks = createTypingCallbacks({
start: sendTyping,
onStartError: (err) => {
logTypingFailure({
log: logVerbose,
channel: "telegram",
target: String(chatId),
error: err,
});
},
});
try {
({ queuedFinal } = await dispatchReplyWithBufferedBlockDispatcher({
ctx: ctxPayload,
cfg,
dispatcherOptions: {
...prefixOptions,
typingCallbacks,
deliver: async (payload, info) => {
const previewButtons = (
payload.channelData?.telegram as { buttons?: TelegramInlineButtons } | undefined
)?.buttons;
const split = splitTextIntoLaneSegments(payload.text);
const segments = split.segments;
const hasMedia = Boolean(payload.mediaUrl) || (payload.mediaUrls?.length ?? 0) > 0;
const flushBufferedFinalAnswer = async () => {
const buffered = reasoningStepState.takeBufferedFinalAnswer();
if (!buffered) {
return;
}
const bufferedButtons = (
buffered.payload.channelData?.telegram as
| { buttons?: TelegramInlineButtons }
| undefined
)?.buttons;
await deliverLaneText({
laneName: "answer",
text: buffered.text,
payload: buffered.payload,
infoKind: "final",
previewButtons: bufferedButtons,
});
reasoningStepState.resetForNextStep();
};
for (const segment of segments) {
if (
segment.lane === "answer" &&
info.kind === "final" &&
reasoningStepState.shouldBufferFinalAnswer()
) {
reasoningStepState.bufferFinalAnswer({ payload, text: segment.text });
continue;
}
if (segment.lane === "reasoning") {
reasoningStepState.noteReasoningHint();
}
const result = await deliverLaneText({
laneName: segment.lane,
text: segment.text,
payload,
infoKind: info.kind,
previewButtons,
allowPreviewUpdateForNonFinal: segment.lane === "reasoning",
});
if (segment.lane === "reasoning") {
if (result !== "skipped") {
reasoningStepState.noteReasoningDelivered();
await flushBufferedFinalAnswer();
}
continue;
}
if (info.kind === "final") {
if (reasoningLane.hasStreamedMessage) {
finalizedPreviewByLane.reasoning = true;
}
reasoningStepState.resetForNextStep();
}
}
if (segments.length > 0) {
return;
}
if (split.suppressedReasoningOnly) {
if (hasMedia) {
const payloadWithoutSuppressedReasoning =
typeof payload.text === "string" ? { ...payload, text: "" } : payload;
await sendPayload(payloadWithoutSuppressedReasoning);
}
if (info.kind === "final") {
await flushBufferedFinalAnswer();
}
return;
}
if (info.kind === "final") {
await answerLane.stream?.stop();
await reasoningLane.stream?.stop();
reasoningStepState.resetForNextStep();
}
const canSendAsIs =
hasMedia || (typeof payload.text === "string" && payload.text.length > 0);
if (!canSendAsIs) {
if (info.kind === "final") {
await flushBufferedFinalAnswer();
}
return;
}
await sendPayload(payload);
if (info.kind === "final") {
await flushBufferedFinalAnswer();
}
},
onSkip: (_payload, info) => {
if (info.reason !== "silent") {
deliveryState.markNonSilentSkip();
}
},
onError: (err, info) => {
deliveryState.markNonSilentFailure();
runtime.error?.(danger(`telegram ${info.kind} reply failed: ${String(err)}`));
},
},
replyOptions: {
skillFilter,
disableBlockStreaming,
onPartialReply:
answerLane.stream || reasoningLane.stream
? (payload) => ingestDraftLaneSegments(payload.text)
: undefined,
onReasoningStream: reasoningLane.stream
? (payload) => {
// Split between reasoning blocks only when the next reasoning
// stream starts. Splitting at reasoning-end can orphan the active
// preview and cause duplicate reasoning sends on reasoning final.
if (splitReasoningOnNextStream) {
reasoningLane.stream?.forceNewMessage();
resetDraftLaneState(reasoningLane);
splitReasoningOnNextStream = false;
}
ingestDraftLaneSegments(payload.text);
}
: undefined,
onAssistantMessageStart: answerLane.stream
? async () => {
reasoningStepState.resetForNextStep();
if (answerLane.hasStreamedMessage) {
const previewMessageId = answerLane.stream?.messageId();
// Only archive previews that still need a matching final text update.
// Once a preview has already been finalized, archiving it here causes
// cleanup to delete a user-visible final message on later media-only turns.
if (typeof previewMessageId === "number" && !finalizedPreviewByLane.answer) {
archivedAnswerPreviews.push({
messageId: previewMessageId,
textSnapshot: answerLane.lastPartialText,
});
}
answerLane.stream?.forceNewMessage();
}
resetDraftLaneState(answerLane);
// New assistant message boundary: this lane now tracks a fresh preview lifecycle.
finalizedPreviewByLane.answer = false;
}
: undefined,
onReasoningEnd: reasoningLane.stream
? () => {
// Split when/if a later reasoning block begins.
splitReasoningOnNextStream = reasoningLane.hasStreamedMessage;
}
: undefined,
onToolStart: statusReactionController
? async (payload) => {
await statusReactionController.setTool(payload.name);
}
: undefined,
onModelSelected,
},
}));
} finally {
// Must stop() first to flush debounced content before clear() wipes state.
const streamCleanupStates = new Map<
NonNullable<DraftLaneState["stream"]>,
{ shouldClear: boolean }
>();
const lanesToCleanup: Array<{ laneName: LaneName; lane: DraftLaneState }> = [
{ laneName: "answer", lane: answerLane },
{ laneName: "reasoning", lane: reasoningLane },
];
for (const laneState of lanesToCleanup) {
const stream = laneState.lane.stream;
if (!stream) {
continue;
}
const shouldClear = !finalizedPreviewByLane[laneState.laneName];
const existing = streamCleanupStates.get(stream);
if (!existing) {
streamCleanupStates.set(stream, { shouldClear });
continue;
}
existing.shouldClear = existing.shouldClear && shouldClear;
}
for (const [stream, cleanupState] of streamCleanupStates) {
await stream.stop();
if (cleanupState.shouldClear) {
await stream.clear();
}
}
for (const archivedPreview of archivedAnswerPreviews) {
try {
await bot.api.deleteMessage(chatId, archivedPreview.messageId);
} catch (err) {
logVerbose(
`telegram: archived answer preview cleanup failed (${archivedPreview.messageId}): ${String(err)}`,
);
}
}
for (const messageId of archivedReasoningPreviewIds) {
try {
await bot.api.deleteMessage(chatId, messageId);
} catch (err) {
logVerbose(
`telegram: archived reasoning preview cleanup failed (${messageId}): ${String(err)}`,
);
}
}
}
let sentFallback = false;
const deliverySummary = deliveryState.snapshot();
if (
!deliverySummary.delivered &&
(deliverySummary.skippedNonSilent > 0 || deliverySummary.failedNonSilent > 0)
) {
const result = await deliverReplies({
replies: [{ text: EMPTY_RESPONSE_FALLBACK }],
...deliveryBaseOptions,
});
sentFallback = result.delivered;
}
const hasFinalResponse = queuedFinal || sentFallback;
if (statusReactionController && !hasFinalResponse) {
void statusReactionController.setError().catch((err) => {
logVerbose(`telegram: status reaction error finalize failed: ${String(err)}`);
});
}
if (!hasFinalResponse) {
clearGroupHistory();
return;
}
if (statusReactionController) {
void statusReactionController.setDone().catch((err) => {
logVerbose(`telegram: status reaction finalize failed: ${String(err)}`);
});
} else {
removeAckReactionAfterReply({
removeAfterReply: removeAckAfterReply,
ackReactionPromise,
ackReactionValue: ackReactionPromise ? "ack" : null,
remove: () => reactionApi?.(chatId, msg.message_id ?? 0, []) ?? Promise.resolve(),
onError: (err) => {
if (!msg.message_id) {
return;
}
logAckFailure({
log: logVerbose,
channel: "telegram",
target: `${chatId}/${msg.message_id}`,
error: err,
});
},
});
}
clearGroupHistory();
};