Fix prettyText punctuation and long audio progress

src/handlers/onAudio.ts

@@ -40,6 +40,7 @@ export async function processAudio(
   const link = await ctx.telegram.getFileLink(voice.file_id);
   const oggPath = tmp.tmpNameSync({ postfix: ".ogg" });
   let mp3Path: string | null = null;
+  let progressTimer: NodeJS.Timeout | null = null;
 
   try {
     const response = await fetch(link.href);
@@ -50,6 +51,16 @@ export async function processAudio(
     await fs.promises.writeFile(oggPath, Buffer.from(arrayBuffer));
 
     mp3Path = await convertToMp3(oggPath);
+
+    progressTimer = setInterval(() => {
+      void sendTelegramMessage(
+        chatId,
+        "Распознавание продолжается...",
+        undefined,
+        ctx,
+      );
+    }, 60_000);
+
     const res = (await sendAudioWhisper({ mp3Path })) as WhisperResponse;
 
     if (res.error) {
@@ -110,6 +121,7 @@ export async function processAudio(
     );
   } finally {
     try {
+      if (progressTimer) clearInterval(progressTimer);
       if (fs.existsSync(oggPath)) fs.unlinkSync(oggPath);
       if (mp3Path && fs.existsSync(mp3Path)) fs.unlinkSync(mp3Path);
     } catch (cleanupError) {

src/utils/text.ts

@@ -22,37 +22,40 @@ export function splitBigMessage(text: string) {
       msg += line + "\n";
     }
   }
-  
+
   // Handle any remaining text in the buffer
   if (msg) {
     // Remove trailing newline if present
     const trimmedMsg = msg.endsWith("\n") ? msg.slice(0, -1) : msg;
     if (trimmedMsg) msgs.push(trimmedMsg);
   }
-  
+
   return msgs;
 }
 
 export function prettyText(text: string) {
-  const paragraphs = [];
-  const sentences = text.split(/[.?!][ \n]/g);
-  // console.log("sentences:", sentences.length);
-  let paragraph = '';
-  while (sentences.length > 0) {
-    paragraph += sentences.shift() + '. ';
-    paragraph = paragraph.replace(/\.\. $/, '. '); // remove ..
+  if (!text) return "";
+
+  const paragraphs: string[] = [];
+  const sentences = text.match(/[^.?!]+[.?!]?/g) || [];
+
+  let paragraph = "";
+  for (let sentence of sentences) {
+    sentence = sentence.trim();
+    if (!/[.?!]$/.test(sentence)) sentence += ".";
+
+    paragraph += sentence + " ";
+    paragraph = paragraph.replace(/\.\. $/, ". ");
 
     if (paragraph.length > 200) {
-      paragraphs.push(paragraph.trim() + '');
-      // console.log("zero paragraph:", paragraph);
-      paragraph = '';
+      paragraphs.push(paragraph.trim());
+      paragraph = "";
     }
   }
-  if (paragraph.length > 0) {
-    paragraphs.push(paragraph.trim() + '');
+
+  if (paragraph.trim()) {
+    paragraphs.push(paragraph.trim());
   }
-  // console.log("paragraphs", paragraphs);
 
-  const prettyText = paragraphs.join('\n\n');
-  return prettyText;
+  return paragraphs.join("\n\n");
 }

tests/handlers/onAudioProcess.test.ts

@@ -64,7 +64,7 @@ describe("processAudio", () => {
     await processAudio(ctx as Context, { file_id: "f" }, 1);
     expect(mockSendTelegramMessage).toHaveBeenCalledWith(
       1,
-      "hello",
+      "hello.",
       undefined,
       ctx,
     );

tests/utils/text.test.ts

@@ -13,41 +13,42 @@ describe("splitBigMessage", () => {
 
   it("truncates single oversized line", () => {
     const line = "x".repeat(4100);
-    const [msg] = splitBigMessage(line);
-    expect(msg.length).toBe(4096);
-    expect(msg.endsWith("...")).toBe(true);
+    const msgs = splitBigMessage(line);
+    expect(msgs.length).toBe(2);
+    expect(msgs[0].length).toBe(4096);
+    expect(msgs[1].length).toBe(4);
   });
 
   it("keeps blank lines", () => {
     const res = splitBigMessage("a\n\nb");
-    expect(res).toEqual(["a\n\nb\n"]);
+    expect(res).toEqual(["a\n\nb"]);
   });
 });
 
 describe("prettyText", () => {
   it("formats a simple sentence correctly", () => {
     const input = "Hello world.";
     const result = prettyText(input);
-    expect(result).toBe("Hello world. ");
+    expect(result).toBe("Hello world.");
   });
 
   it("combines short sentences into one paragraph", () => {
     const input = "First sentence. Second sentence. Third sentence.";
     const result = prettyText(input);
-    expect(result).toBe("First sentence. Second sentence. Third sentence. ");
+    expect(result).toBe("First sentence. Second sentence. Third sentence.");
   });
 
   it("splits long text into multiple paragraphs", () => {
-    const longText = 
+    const longText =
       "This is a long sentence that should be in the first paragraph. " +
       "This is another long sentence that should be in the first paragraph. " +
       "This is a very long sentence that should trigger a new paragraph because " +
       "it makes the total length exceed the 200 character limit. " +
       "This should be in the second paragraph.";
-    
+
     const result = prettyText(longText);
     const paragraphs = result.split("\n\n");
-    
+
     expect(paragraphs.length).toBe(2);
     expect(paragraphs[0].length).toBeGreaterThan(200);
     expect(paragraphs[0]).toContain("first paragraph.");
@@ -67,6 +68,6 @@ describe("prettyText", () => {
   });
 
   it("handles single word", () => {
-    expect(prettyText("Hello")).toBe("Hello. ");
+    expect(prettyText("Hello")).toBe("Hello.");
   });
 });