@llblab/pi-telegram 0.24.10 → 0.25.0

package/lib/outbound-markup.ts

@@ -16,16 +16,6 @@ interface TelegramTopLevelFenceState {
   length: number;
 }
 
-function isTelegramActionCommentContent(content: string): boolean {
-  const normalizedContent = content.replace(/^\s+/, "");
-  const [head = ""] = normalizedContent.split(/\r?\n/, 1);
-  return ["telegram_voice", "telegram_button"].some((command) => {
-    if (!head.startsWith(command)) return false;
-    const nextChar = head[command.length];
-    return nextChar === undefined || /\s|:/.test(nextChar);
-  });
-}
-
 function getMarkdownLineEnd(markdown: string, offset: number): number {
   const newlineIndex = markdown.indexOf("\n", offset);
   return newlineIndex === -1 ? markdown.length : newlineIndex + 1;
@@ -64,58 +54,6 @@ function isTopLevelClosingFence(
   );
 }
 
-function collectPairedTelegramVoiceActionBody(
-  markdown: string,
-  bodyStart: number,
-  commentContent: string,
-): { content: string; end: number } | undefined {
-  const normalizedContent = commentContent.trim();
-  if (
-    !normalizedContent.startsWith("telegram_voice") ||
-    !isTelegramActionCommentContent(commentContent)
-  ) {
-    return undefined;
-  }
-  let offset = bodyStart;
-  while (offset < markdown.length) {
-    const lineEnd = getMarkdownLineEnd(markdown, offset);
-    const line = getMarkdownLineText(markdown, offset, lineEnd);
-    if (line === "<!-- /telegram_voice -->") {
-      const body = markdown.slice(bodyStart, offset).trim();
-      if (!body) return undefined;
-      return {
-        content: `${commentContent.trimEnd()}\n${body}`,
-        end: lineEnd,
-      };
-    }
-    if (line.startsWith("<!--")) return undefined;
-    offset = lineEnd;
-  }
-  return undefined;
-}
-
-function collectInlineClosedTelegramActionBody(
-  markdown: string,
-  bodyStart: number,
-  commentContent: string,
-): { content: string; end: number } | undefined {
-  const bodyLineEnd = getMarkdownLineEnd(markdown, bodyStart);
-  const bodyLine = getMarkdownLineText(markdown, bodyStart, bodyLineEnd);
-  const closeLineEnd = getMarkdownLineEnd(markdown, bodyLineEnd);
-  const closeLine = getMarkdownLineText(markdown, bodyLineEnd, closeLineEnd);
-  const hasRecoverableBody =
-    isTelegramActionCommentContent(commentContent) &&
-    bodyLine.trim() !== "" &&
-    !bodyLine.startsWith("<!--") &&
-    !bodyLine.startsWith("-->") &&
-    closeLine === "-->";
-  if (!hasRecoverableBody) return undefined;
-  return {
-    content: `${commentContent.trimEnd()}\n${bodyLine}`,
-    end: bodyLineEnd + 3,
-  };
-}
-
 export function collectTopLevelHtmlComments(markdown: string): {
   comments: TelegramTopLevelHtmlComment[];
   openCommentStart?: number;
@@ -140,27 +78,9 @@ export function collectTopLevelHtmlComments(markdown: string): {
     if (line.startsWith("<!--")) {
       const closeIndex = markdown.indexOf("-->", offset + 4);
       if (closeIndex === -1) return { comments, openCommentStart: offset };
-      let end = closeIndex + 3;
-      let raw = markdown.slice(offset, end);
-      let content = raw.slice(4, -3);
-      const closeColumn = closeIndex - offset;
-      const closesOnOpeningLine = closeIndex < lineEnd;
-      const hasOnlyWhitespaceAfterClose =
-        line.slice(closeColumn + 3).trim() === "";
-      const pairedVoiceBody =
-        closesOnOpeningLine && hasOnlyWhitespaceAfterClose
-          ? collectPairedTelegramVoiceActionBody(markdown, lineEnd, content)
-          : undefined;
-      const inlineBody =
-        !pairedVoiceBody && closesOnOpeningLine && hasOnlyWhitespaceAfterClose
-          ? collectInlineClosedTelegramActionBody(markdown, lineEnd, content)
-          : undefined;
-      const recoveredBody = pairedVoiceBody ?? inlineBody;
-      if (recoveredBody) {
-        end = recoveredBody.end;
-        raw = markdown.slice(offset, end);
-        content = recoveredBody.content;
-      }
+      const end = closeIndex + 3;
+      const raw = markdown.slice(offset, end);
+      const content = raw.slice(4, -3);
       comments.push({ raw, content, start: offset, end });
       offset = getMarkdownLineEnd(markdown, end);
       continue;
@@ -233,18 +153,47 @@ export function parseTopLevelTelegramComment(
   };
 }
 
-export function parseTelegramCommentAttributes(
-  input: string,
-): Record<string, string> {
+function parseCanonicalTelegramActionAttributes(
+  source: string,
+): Record<string, string> | undefined {
   const attributes: Record<string, string> = {};
-  for (const match of input.matchAll(
-    /([A-Za-z_][A-Za-z0-9_-]*)=(?:"([^"]*)"|'([^']*)'|(\S+))/g,
-  )) {
-    const key = match[1];
-    const value = (match[2] ?? match[3] ?? match[4] ?? "").trim();
-    if (value) attributes[key] = value;
+  const pattern = /\s*([A-Za-z_][A-Za-z0-9_-]*)="([^"]*)"/y;
+  let offset = 0;
+  while (offset < source.length) {
+    pattern.lastIndex = offset;
+    const match = pattern.exec(source);
+    if (!match) return undefined;
+    const value = match[2].trim();
+    if (value) attributes[match[1]] = value;
+    offset = pattern.lastIndex;
+  }
+  return Object.keys(attributes).length > 0 ? attributes : undefined;
+}
+
+export function parseTelegramActionPayload(
+  comment: TelegramTopLevelHtmlComment,
+  command: string,
+): Record<string, unknown> | undefined {
+  const parsed = parseTopLevelTelegramComment(comment, command);
+  if (!parsed) return undefined;
+  const source = [parsed.head, parsed.body]
+    .filter((part): part is string => part !== undefined)
+    .join("\n")
+    .trim()
+    .replace(/^:\s*/, "");
+  if (!source) return undefined;
+  if (source.startsWith("{")) {
+    try {
+      const value: unknown = JSON.parse(source);
+      return value !== null && typeof value === "object" && !Array.isArray(value)
+        ? (value as Record<string, unknown>)
+        : undefined;
+    } catch {
+      return undefined;
+    }
   }
-  return attributes;
+  if (parsed.body !== undefined) return undefined;
+  return parseCanonicalTelegramActionAttributes(source);
 }
 
 export function normalizeMarkdownAfterVoiceExtraction(
@@ -295,58 +244,14 @@ export interface TelegramVoiceReplyPlan {
   rate?: string;
 }
 
-function parseVoiceReplyAttributes(input: string): {
-  lang?: string;
-  rate?: string;
-  text?: string;
-} {
-  const attributes = parseTelegramCommentAttributes(input);
-  return {
-    ...(attributes.lang ? { lang: attributes.lang } : {}),
-    ...(attributes.rate ? { rate: attributes.rate } : {}),
-    ...(attributes.text ? { text: attributes.text } : {}),
-  };
-}
-
-function parseVoiceCommentBody(
-  head: string,
-  body: string | undefined,
-): {
-  attrs: string;
-  text: string;
-} {
-  const trimmedHead = head.trim();
-  if (body !== undefined) {
-    return { attrs: trimmedHead.replace(/^:/, "").trim(), text: body.trim() };
-  }
-  let colonIndex = -1;
-  let inQuote = false;
-  let quoteChar = "";
-  for (let i = 0; i < trimmedHead.length; i++) {
-    const char = trimmedHead[i];
-    if (inQuote) {
-      if (char === quoteChar) inQuote = false;
-    } else {
-      if (char === '"' || char === "'") {
-        inQuote = true;
-        quoteChar = char;
-      } else if (char === ":") {
-        colonIndex = i;
-        break;
-      }
-    }
-  }
-  if (colonIndex > 0) {
-    const attrsPart = trimmedHead.slice(0, colonIndex).trim();
-    const textPart = trimmedHead.slice(colonIndex + 1).trim();
-    const attrs = parseVoiceReplyAttributes(attrsPart);
-    return { attrs: attrsPart, text: textPart || attrs.text || "", ...attrs };
-  }
-  if (trimmedHead.startsWith(":")) {
-    return { attrs: "", text: trimmedHead.slice(1).trim() };
-  }
-  const attrs = parseVoiceReplyAttributes(trimmedHead);
-  return { attrs: trimmedHead, text: attrs.text ?? "" };
+function getTelegramActionString(
+  payload: Record<string, unknown>,
+  key: string,
+): string | undefined {
+  const value = payload[key];
+  if (typeof value !== "string") return undefined;
+  const trimmed = value.trim();
+  return trimmed || undefined;
 }
 
 export function planTelegramVoiceReply(
@@ -356,26 +261,24 @@ export function planTelegramVoiceReply(
   let lang: string | undefined;
   let rate: string | undefined;
   const stripped = replaceTopLevelHtmlComments(markdown, (comment) => {
-    let command = parseTopLevelTelegramComment(comment, "telegram_voice");
-    if (!command) {
-      let content = comment.content.replace(/^\s+/, "").replace(/^!/, "");
-      if (content.startsWith("telegram_voice")) {
-        const headPart = content.slice("telegram_voice".length).trim();
-        command = { head: headPart, body: undefined };
-      }
-    }
-    if (!command) return "";
-    const parsed = parseVoiceCommentBody(command.head, command.body);
-    const attrs = parseVoiceReplyAttributes(parsed.attrs);
-    if (parsed.text) {
+    const command = parseTopLevelTelegramComment(comment, "telegram_voice");
+    if (!command) return comment.raw;
+    const payload = parseTelegramActionPayload(comment, "telegram_voice");
+    if (!payload) return "";
+    const text =
+      getTelegramActionString(payload, "text") ??
+      getTelegramActionString(payload, "value");
+    const itemLang = getTelegramActionString(payload, "lang");
+    const itemRate = getTelegramActionString(payload, "rate");
+    if (text) {
       voiceReplies.push({
-        text: parsed.text,
-        ...(attrs.lang ? { lang: attrs.lang } : {}),
-        ...(attrs.rate ? { rate: attrs.rate } : {}),
+        text,
+        ...(itemLang ? { lang: itemLang } : {}),
+        ...(itemRate ? { rate: itemRate } : {}),
       });
     }
-    if (attrs.lang) lang = attrs.lang;
-    if (attrs.rate) rate = attrs.rate;
+    if (itemLang) lang = itemLang;
+    if (itemRate) rate = itemRate;
     return "";
   });
   const voiceText = voiceReplies

package/lib/polling.ts

@@ -319,7 +319,6 @@ export interface TelegramThreadCapabilityRuntimeDeps<
   TContext,
 > extends TelegramThreadCapabilityReaderDeps {
   topicTargetStore: TelegramThreadCapabilityStore;
-  isBusConfigured: () => boolean;
   ownsLock: (ctx: TContext) => boolean;
   getPollingStartedWithTelegramBus: () => boolean;
   setPollingStartedWithTelegramBus: (started: boolean) => void;
@@ -394,7 +393,6 @@ export interface TelegramThreadAwarePollingDeps<
   TContext,
   TOwner,
 > extends TelegramStartupThreadCapabilityProbeDeps {
-  isBusConfigured: () => boolean;
   isBusRuntimeEnabled: () => boolean;
   isTopicModeUnavailableError: (error: unknown) => boolean;
   getPollingStartedWithTelegramBus: () => boolean;
@@ -419,7 +417,6 @@ export interface TelegramThreadCapabilityOrchestrationDeps<
 > extends TelegramThreadCapabilityReaderDeps {
   state: TelegramThreadCapabilityStateRuntime;
   topicTargetStore: TelegramThreadCapabilityStore;
-  isBusConfigured: () => boolean;
   isBusRuntimeEnabled: () => boolean;
   ownsLock: (ctx: TContext) => boolean;
   startClassicPolling: (ctx: TContext) => MaybePromise<void>;
@@ -475,7 +472,6 @@ export function createTelegramThreadCapabilityOrchestration<TContext, TOwner>(
     getAllowedUserId: deps.getAllowedUserId,
     callApi: deps.callApi,
     topicTargetStore: deps.topicTargetStore,
-    isBusConfigured: deps.isBusConfigured,
     ownsLock: deps.ownsLock,
     getPollingStartedWithTelegramBus: deps.state.isBusPollingStarted,
     setPollingStartedWithTelegramBus: deps.state.setBusPollingStarted,
@@ -498,7 +494,6 @@ export function createTelegramThreadCapabilityOrchestration<TContext, TOwner>(
       getAllowedUserId: deps.getAllowedUserId,
       callApi: deps.callApi,
       topicTargetStore: deps.topicTargetStore,
-      isBusConfigured: deps.isBusConfigured,
       isBusRuntimeEnabled: deps.isBusRuntimeEnabled,
       isTopicModeUnavailableError: deps.isTopicModeUnavailableError,
       getPollingStartedWithTelegramBus: deps.state.isBusPollingStarted,
@@ -590,7 +585,6 @@ export async function applyTelegramThreadCapability<TContext>(
   deps: TelegramThreadCapabilityRuntimeDeps<TContext>,
 ): Promise<void> {
   await deps.topicTargetStore.load();
-  if (!deps.isBusConfigured()) return;
   const nowMs = (deps.getNowMs ?? Date.now)();
   const previousBotState = deps.topicTargetStore.getBotState();
   if (!threadModeEnabled) {
@@ -691,17 +685,14 @@ export function createTelegramThreadAwarePollingPorts<TContext, TOwner>(
     ctx: TContext,
     options?: { forceFreshLeaderThread?: boolean },
   ): Promise<void> => {
-    if (deps.isBusConfigured()) {
-      await deps.topicTargetStore.load();
-      let startupThreadCapability: boolean | undefined;
-      try {
-        startupThreadCapability =
-          await probeTelegramStartupThreadCapability(deps);
-      } catch (error) {
-        deps.recordEvent("bus", error, { phase: "startup-thread-mode-probe" });
-      }
-      deps.setTopicModeUnavailable(startupThreadCapability !== true);
+    await deps.topicTargetStore.load();
+    let startupThreadCapability: boolean | undefined;
+    try {
+      startupThreadCapability = await probeTelegramStartupThreadCapability(deps);
+    } catch (error) {
+      deps.recordEvent("bus", error, { phase: "startup-thread-mode-probe" });
     }
+    deps.setTopicModeUnavailable(startupThreadCapability !== true);
     if (deps.isBusRuntimeEnabled()) {
       deps.setTopicModeUnavailable(false);
       try {
@@ -886,7 +877,6 @@ export function createTelegramThreadCapabilityMonitor<TContext>(
   return {
     start(ctx) {
       stop();
-      if (!deps.isBusConfigured()) return;
       interval = setInterval(() => {
         check(ctx);
       }, intervalMs);

package/lib/prompts.ts

@@ -138,13 +138,13 @@ How to answer Telegram turns:
 Assistant-authored Telegram actions:
 - \`telegram_voice\` and \`telegram_button\` are hidden top-level HTML comments, not Pi tools.
 - Put action comments at column zero, outside code, quotes, lists, and indented examples.
-- Voice forms: \`<!-- telegram_voice text="Short summary" -->\`, \`<!-- telegram_voice: Short summary -->\`, multiline \`<!-- telegram_voice lang=ru\nShort summary.\n-->\`, or paired \`<!-- telegram_voice lang=ru -->\nShort summary.\n<!-- /telegram_voice -->\`.
-- Keep the complete action at top level and include a non-empty voice payload.
+- Action payloads use either JSON or double-quoted attributes; the colon after the action name is optional and does not affect parsing. Voice accepts equivalent \`text\` or \`value\`: \`<!-- telegram_voice: {"value":"Short summary","lang":"en"} -->\` or \`<!-- telegram_voice text="Short summary" lang="en" -->\`.
+- Keep the complete action in one top-level comment and include non-empty \`text\` or \`value\`; encode line breaks inside JSON strings as \`\\n\`.
 - Keep voice text TTS-friendly; avoid raw Markdown, code, and tables in voice text.
 - Voice delivery generates and attaches OGG automatically; do not also call \`telegram_attach\` for the same audio.
 - Voice reply modes are compact: \`hidden\` emits no automatic context, \`mirror\` emits it for voice/audio input, and \`always\` emits it for every Telegram turn. Explicit \`telegram_voice\` remains available for an intentionally distinct spoken payload.
-- Button forms: \`<!-- telegram_button: OK -->\`, \`<!-- telegram_button label=Continue prompt="Continue with the current plan." -->\`, or multiline \`<!-- telegram_button label="Show risks" selected_style="danger"\nList the main risks first.\n-->\`.
-- Optional \`selected_style\` controls the button after queue admission: \`primary\` (default, blue), \`success\` (green), or \`danger\` (red). It never suppresses the prompt.
+- Button payloads use \`label\` plus \`prompt\`, or compact \`value\` when both are identical: \`<!-- telegram_button: {"label":"Continue","prompt":"Continue with the current plan."} -->\` or \`<!-- telegram_button value="Continue" -->\`.
+- Optional \`selected_style\` in either payload form controls the button after queue admission: \`primary\` (default, blue), \`success\` (green), or \`danger\` (red). It never suppresses the prompt.
 - If hidden button comments form the whole reply, the bridge supplies the standard \`☑️ **Choose an option:**\` heading automatically.
 
 Local/TUI direct delivery:

package/lib/queue.ts

@@ -968,6 +968,7 @@ export interface TelegramAgentEndRuntimeDeps<
   isSessionActive?: () => boolean;
   isTurnTransportActive?: (turn: TTurn) => boolean;
   waitForTypingIdle?: () => Promise<void>;
+  waitForActivityIdle?: () => Promise<void>;
   updateStatus: () => void;
   dispatchNextQueuedTelegramTurn: () => void;
   scheduleActiveTurnDelivery?: (task: () => Promise<void>) => void;
@@ -1047,6 +1048,7 @@ export interface TelegramAgentEndHookRuntimeDeps<
   isSessionActive?: (ctx: TContext) => boolean;
   isTurnTransportActive?: (turn: TTurn) => boolean;
   waitForTypingIdle?: () => Promise<void>;
+  waitForActivityIdle?: () => Promise<void>;
   updateStatus: (ctx: TContext) => void;
   dispatchNextQueuedTelegramTurn: (ctx: TContext) => void;
   requestDeferredDispatchNextQueuedTelegramTurn: (
@@ -1188,6 +1190,7 @@ export function createTelegramAgentEndHook<
       isSessionActive: () => deps.isSessionActive?.(ctx) ?? true,
       isTurnTransportActive: deps.isTurnTransportActive,
       waitForTypingIdle: deps.waitForTypingIdle,
+      waitForActivityIdle: deps.waitForActivityIdle,
       updateStatus: () => deps.updateStatus(ctx),
       dispatchNextQueuedTelegramTurn: () => {
         deps.requestDeferredDispatchNextQueuedTelegramTurn(
@@ -1358,6 +1361,7 @@ export async function handleTelegramAgentEndRuntime<
     return;
   }
   const deliverActiveTurn = async () => {
+    await deps.waitForActivityIdle?.();
     if (!isDeliveryActive()) return;
     if (finalText) deps.setPreviewPendingText(finalText);
     if (!finalText && hasOutboundArtifacts)

package/lib/telegram-api.ts

@@ -274,6 +274,24 @@ type TelegramInputRichMessageCommon = {
   skip_entity_detection?: boolean;
 };
 
+export type TelegramRichText =
+  | string
+  | TelegramRichText[]
+  | { type: "bold" | "code"; text: TelegramRichText };
+
+export type TelegramInputRichBlock =
+  | { type: "pre"; text: TelegramRichText; language?: string }
+  | {
+      type: "details";
+      summary: TelegramRichText;
+      blocks: TelegramInputRichBlock[];
+      is_open?: true;
+    };
+
+export type TelegramInputRichDraftBlock =
+  | TelegramInputRichBlock
+  | { type: "thinking"; text: TelegramRichText };
+
 export type TelegramInputRichMessage = TelegramInputRichMessageCommon &
   (
     | {
@@ -288,6 +306,12 @@ export type TelegramInputRichMessage = TelegramInputRichMessageCommon &
         blocks?: never;
         media?: TelegramInputRichMessageMedia[];
       }
+    | {
+        blocks: TelegramInputRichBlock[];
+        markdown?: never;
+        html?: never;
+        media?: never;
+      }
   );
 
 export type TelegramSendRichMessageBody = Record<string, unknown> & {
@@ -318,7 +342,14 @@ export type TelegramSendMessageDraftBody = Record<string, unknown> & {
 export type TelegramSendRichMessageDraftBody = Record<string, unknown> & {
   chat_id: number;
   draft_id: number;
-  rich_message: TelegramInputRichMessage;
+  rich_message:
+    | TelegramInputRichMessage
+    | (TelegramInputRichMessageCommon & {
+        blocks: TelegramInputRichDraftBlock[];
+        markdown?: never;
+        html?: never;
+        media?: never;
+      });
   message_thread_id?: number;
 };
 
@@ -497,6 +528,11 @@ export interface TelegramBridgeApiRuntime {
   editMessageText: (
     body: TelegramEditMessageTextBody,
   ) => Promise<"edited" | "unchanged">;
+  editMessageReplyMarkup: (
+    chatId: number,
+    messageId: number,
+    replyMarkup: unknown,
+  ) => Promise<void>;
   answerCallbackQuery: (
     callbackQueryId: string,
     text?: string,
@@ -1569,6 +1605,13 @@ export function createTelegramBridgeApiRuntime(
         throw error;
       }
     },
+    editMessageReplyMarkup: async (chatId, messageId, replyMarkup) => {
+      await callRecorded("editMessageReplyMarkup", {
+        chat_id: chatId,
+        message_id: messageId,
+        reply_markup: replyMarkup,
+      });
+    },
     answerCallbackQuery: async (callbackQueryId, text) => {
       try {
         await deps.client.answerCallbackQuery(callbackQueryId, text);

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@llblab/pi-telegram",
-  "version": "0.24.10",
+  "version": "0.25.0",
   "private": false,
   "publishConfig": {
     "access": "public"