feat(phonic): handle user text input via generate_reply

[devin-ai-integration]

Summary

The Phonic realtime plugin currently only handles audio input. When a user types text (via LiveKit's chat/text stream), the framework calls update_chat_ctx with the user message followed by generate_reply. Previously, user-role messages in update_chat_ctx were silently ignored — typed input was dropped with no response.

This adds text input support:

# In update_chat_ctx — detect user messages, buffer text:
if isinstance(item, llm.ChatMessage) and item.role == "user":
    self._pending_user_text = item.text_content

# In _send_generate_reply — include buffered text in system_message:
if self._pending_user_text:
    system_message = f'The user sent: "{self._pending_user_text}". Please respond.'
    self._pending_user_text = None

The _pending_user_text bridges the gap between update_chat_ctx (where the framework delivers user text) and _send_generate_reply (where we can pass it to the downstream echo service via the system_message field of generate_reply).

Python counterpart of Phonic-Co/livekit-agents-js#5.

Link to Devin session: https://app.devin.ai/sessions/51df28f166ac4fff8cae41b88dc9bab2
Requested by: @qionghuang6

[devin-ai-integration]

🤖 Devin AI Engineer

I'll be helping with this pull request! Here's what you should know:

✅ I will automatically:

• Address comments on this PR. Add '(aside)' to your comment to have me ignore it.
• Look at CI failures and help fix them

Note: I can only respond to comments from users who have write access to this repository.

⚙️ Control Options:

• Disable automatic comment, CI, and merge conflict monitoring