fix(telegram): caption-less media messages crash with ChatMessage ValidationError (content min_length=1) #10483

Summary

Same defect class as #10481 but on the Telegram channel (#9006). api/telegram_bot.py:191 builds:

chat_message = ChatMessage(content=unified_message.message, ...)

unified_message.message comes from telegram_adapter.py:56:

text = message.get("text", "") or message.get("caption", "")

For a caption-less photo / document / voice / video / audio message, text == "". ChatMessage.content is Field(..., min_length=1, ...) (api/schemas_chat.py:463), so constructing it raises pydantic ValidationError.

Impact

Every caption-less media message received via the Telegram bot raises ValidationError in handle_telegram_update, so the message is not routed to chat and no reply is sent — the common case (sending a photo/voice with no caption) is broken.

Fix (mirror #10481)

When unified_message.message is empty but metadata.get("has_file") is set, synthesize a placeholder content naming the attachment, e.g. f"[{file_type} attachment]", before constructing ChatMessage. The Telegram adapter already records file_type/mime_type in metadata (telegram_adapter.py:48-53).

Found while fixing the equivalent WhatsApp bug (#10481) during verification of umbrella #9928.