nanobot/channels/telegram.py

@@ -206,25 +206,89 @@ class TelegramChannel(BaseChannel):
             if msg.media:
                 await self._send_with_media(chat_id, html_content, msg.media)
             else:
-                # Text-only message
+                # Text-only message - split if too long
-                await self._app.bot.send_message(
+                await self._send_text_chunks(chat_id, html_content, parse_mode="HTML")
-                    chat_id=chat_id,
-                    text=html_content,
-                    parse_mode="HTML"
-                )
         except ValueError:
             logger.error(f"Invalid chat_id: {msg.chat_id}")
         except Exception as e:
             # Fallback to plain text if HTML parsing fails
             logger.warning(f"HTML parse failed, falling back to plain text: {e}")
             try:
-                await self._app.bot.send_message(
+                await self._send_text_chunks(int(msg.chat_id), msg.content, parse_mode=None)
-                    chat_id=int(msg.chat_id),
-                    text=msg.content
-                )
             except Exception as e2:
                 logger.error(f"Error sending Telegram message: {e2}")
 
+    async def _send_text_chunks(
+        self,
+        chat_id: int,
+        text: str,
+        parse_mode: str | None = "HTML"
+    ) -> None:
+        """Split and send long messages at sentence boundaries.
+
+        Telegram has a 4096 character limit per message.
+        Per design doc: split at sentence boundaries, send as multiple messages.
+        """
+        MAX_LENGTH = 4096
+
+        if len(text) <= MAX_LENGTH:
+            # Single message
+            await self._app.bot.send_message(
+                chat_id=chat_id,
+                text=text,
+                parse_mode=parse_mode
+            )
+            return
+
+        # Split at sentence boundaries
+        import re
+        # Split on sentence endings: . ! ? followed by space/newline/end
+        sentences = re.split(r'([.!?]+(?:\s+|$))', text)
+
+        # Rejoin sentence with its punctuation
+        parts = []
+        for i in range(0, len(sentences) - 1, 2):
+            parts.append(sentences[i] + (sentences[i+1] if i+1 < len(sentences) else ''))
+        if len(sentences) % 2 == 1:  # Last part without punctuation
+            parts.append(sentences[-1])
+
+        # Group into chunks under MAX_LENGTH
+        chunks = []
+        current_chunk = ""
+
+        for part in parts:
+            # If single part exceeds limit, force split it
+            if len(part) > MAX_LENGTH:
+                if current_chunk:
+                    chunks.append(current_chunk)
+                    current_chunk = ""
+                # Hard split at MAX_LENGTH
+                for i in range(0, len(part), MAX_LENGTH):
+                    chunks.append(part[i:i + MAX_LENGTH])
+                continue
+
+            # Try adding part to current chunk
+            test_chunk = current_chunk + part
+            if len(test_chunk) > MAX_LENGTH:
+                # Save current chunk, start new one
+                if current_chunk:
+                    chunks.append(current_chunk)
+                current_chunk = part
+            else:
+                current_chunk = test_chunk
+
+        # Add final chunk
+        if current_chunk:
+            chunks.append(current_chunk)
+
+        # Send all chunks as separate messages
+        for chunk in chunks:
+            await self._app.bot.send_message(
+                chat_id=chat_id,
+                text=chunk.strip(),
+                parse_mode=parse_mode
+            )
+
     async def _send_with_media(self, chat_id: int, caption: str, media_paths: list[str]) -> None:
         """
         Send message with media attachments.

tests/test_telegram_chunking.py

@@ -0,0 +1,170 @@
+"""Tests for Telegram message chunking.
+
+Per design doc: messages >4096 chars should split at sentence boundaries.
+"""
+
+import pytest
+from unittest.mock import AsyncMock, MagicMock
+
+from nanobot.bus.events import OutboundMessage
+from nanobot.channels.telegram import TelegramChannel
+
+
+@pytest.mark.asyncio
+async def test_short_message_not_chunked():
+    """Messages under 4096 chars should send as single message."""
+    config = MagicMock()
+    config.token = "test-token"
+    bus = MagicMock()
+
+    channel = TelegramChannel(config, bus)
+
+    # Mock the bot
+    sent_messages = []
+
+    class MockBot:
+        async def send_message(self, chat_id, text, parse_mode=None):
+            sent_messages.append({"chat_id": chat_id, "text": text})
+
+    class MockApp:
+        bot = MockBot()
+
+    channel._app = MockApp()
+
+    # Short message
+    msg = OutboundMessage(
+        channel="telegram",
+        chat_id="123",
+        content="Short message."
+    )
+
+    await channel.send(msg)
+
+    # Should send exactly 1 message
+    assert len(sent_messages) == 1
+    assert sent_messages[0]["text"] == "Short message."
+
+
+@pytest.mark.asyncio
+async def test_long_message_splits_at_sentences():
+    """Messages >4096 chars should split at sentence boundaries."""
+    config = MagicMock()
+    config.token = "test-token"
+    bus = MagicMock()
+
+    channel = TelegramChannel(config, bus)
+
+    # Mock the bot
+    sent_messages = []
+
+    class MockBot:
+        async def send_message(self, chat_id, text, parse_mode=None):
+            sent_messages.append({"chat_id": chat_id, "text": text})
+
+    class MockApp:
+        bot = MockBot()
+
+    channel._app = MockApp()
+
+    # Create a message longer than 4096 chars with clear sentence boundaries
+    # Each sentence is 200 chars, need 21+ sentences to exceed 4096
+    sentence = "A" * 195 + "end. "  # 200 chars including "end. "
+    long_content = sentence * 25  # 5000 chars total
+
+    msg = OutboundMessage(
+        channel="telegram",
+        chat_id="123",
+        content=long_content
+    )
+
+    await channel.send(msg)
+
+    # Should split into multiple messages
+    assert len(sent_messages) > 1
+
+    # Each message should be under 4096 chars
+    for sent in sent_messages:
+        assert len(sent["text"]) <= 4096
+
+    # All messages combined should equal original (with whitespace trimming)
+    combined = "".join(sent["text"] for sent in sent_messages)
+    assert combined.replace(" ", "") == long_content.replace(" ", "")
+
+
+@pytest.mark.asyncio
+async def test_message_at_exactly_4096_chars():
+    """Message at exactly 4096 chars should not chunk."""
+    config = MagicMock()
+    config.token = "test-token"
+    bus = MagicMock()
+
+    channel = TelegramChannel(config, bus)
+
+    # Mock the bot
+    sent_messages = []
+
+    class MockBot:
+        async def send_message(self, chat_id, text, parse_mode=None):
+            sent_messages.append({"chat_id": chat_id, "text": text})
+
+    class MockApp:
+        bot = MockBot()
+
+    channel._app = MockApp()
+
+    # Exactly 4096 chars
+    content = "A" * 4096
+
+    msg = OutboundMessage(
+        channel="telegram",
+        chat_id="123",
+        content=content
+    )
+
+    await channel.send(msg)
+
+    # Should send exactly 1 message
+    assert len(sent_messages) == 1
+
+
+@pytest.mark.asyncio
+async def test_message_preserves_sentence_boundaries():
+    """Chunks should split at sentence endings, not mid-sentence."""
+    config = MagicMock()
+    config.token = "test-token"
+    bus = MagicMock()
+
+    channel = TelegramChannel(config, bus)
+
+    # Mock the bot
+    sent_messages = []
+
+    class MockBot:
+        async def send_message(self, chat_id, text, parse_mode=None):
+            sent_messages.append({"chat_id": chat_id, "text": text})
+
+    class MockApp:
+        bot = MockBot()
+
+    channel._app = MockApp()
+
+    # Create content with clear sentence markers
+    # First part: just under 4096 chars
+    part1 = "First sentence. " * 250  # ~4000 chars
+    part2 = "Second sentence. "
+    content = part1 + part2
+
+    msg = OutboundMessage(
+        channel="telegram",
+        chat_id="123",
+        content=content
+    )
+
+    await channel.send(msg)
+
+    # Verify chunks don't break mid-sentence
+    for sent in sent_messages:
+        text = sent["text"].strip()
+        # Each chunk should end with sentence punctuation
+        if text:
+            assert text[-1] in ".!?"