fix: use SystemMessage+HumanMessage for follow-up question generation (#1751)

* fix: use SystemMessage+HumanMessage for follow-up question generation (fixes #1697)

Some models (e.g. MiniMax-M2.7) require the system prompt and user
content to be passed as separate message objects rather than a single
combined string. Invoking with a plain string sends everything as a
HumanMessage, which causes these models to ignore the generation
instructions and fail to produce valid follow-up questions.

* test: verify model is invoked with SystemMessage and HumanMessage

backend/app/gateway/routers/suggestions.py

@@ -2,6 +2,7 @@ import json
 import logging
 
 from fastapi import APIRouter
+from langchain_core.messages import HumanMessage, SystemMessage
 from pydantic import BaseModel, Field
 
 from deerflow.models import create_chat_model
@@ -106,7 +107,7 @@ async def generate_suggestions(thread_id: str, request: SuggestionsRequest) -> S
     if not conversation:
         return SuggestionsResponse(suggestions=[])
 
-    prompt = (
+    system_instruction = (
         "You are generating follow-up questions to help the user continue the conversation.\n"
         f"Based on the conversation below, produce EXACTLY {n} short questions the user might ask next.\n"
         "Requirements:\n"
@@ -114,14 +115,13 @@ async def generate_suggestions(thread_id: str, request: SuggestionsRequest) -> S
         "- Questions must be written in the same language as the user.\n"
         "- Keep each question concise (ideally <= 20 words / <= 40 Chinese characters).\n"
         "- Do NOT include numbering, markdown, or any extra text.\n"
-        "- Output MUST be a JSON array of strings only.\n\n"
+        "- Output MUST be a JSON array of strings only.\n"
-        "Conversation:\n"
-        f"{conversation}\n"
     )
+    user_content = f"Conversation Context:\n{conversation}\n\nGenerate {n} follow-up questions"
 
     try:
         model = create_chat_model(name=request.model_name, thinking_enabled=False)
-        response = model.invoke(prompt)
+        response = model.invoke([SystemMessage(content=system_instruction), HumanMessage(content=user_content)])
         raw = _extract_response_text(response.content)
         suggestions = _parse_json_string_list(raw) or []
         cleaned = [s.replace("\n", " ").strip() for s in suggestions if s.strip()]

backend/tests/test_suggestions_router.py

@@ -1,6 +1,8 @@
 import asyncio
 from unittest.mock import MagicMock
 
+from langchain_core.messages import HumanMessage, SystemMessage
+
 from app.gateway.routers import suggestions
 
 
@@ -100,3 +102,26 @@ def test_generate_suggestions_returns_empty_on_model_error(monkeypatch):
     result = asyncio.run(suggestions.generate_suggestions("t1", req))
 
     assert result.suggestions == []
+
+
+def test_generate_suggestions_invokes_model_with_system_and_human_messages(monkeypatch):
+    req = suggestions.SuggestionsRequest(
+        messages=[
+            suggestions.SuggestionMessage(role="user", content="What is Python?"),
+            suggestions.SuggestionMessage(role="assistant", content="Python is a programming language."),
+        ],
+        n=2,
+        model_name=None,
+    )
+    fake_model = MagicMock()
+    fake_model.invoke.return_value = MagicMock(content='["Q1", "Q2"]')
+    monkeypatch.setattr(suggestions, "create_chat_model", lambda **kwargs: fake_model)
+
+    asyncio.run(suggestions.generate_suggestions("t1", req))
+
+    call_args = fake_model.invoke.call_args[0][0]
+    assert len(call_args) == 2
+    assert isinstance(call_args[0], SystemMessage)
+    assert isinstance(call_args[1], HumanMessage)
+    assert "follow-up questions" in call_args[0].content
+    assert "What is Python?" in call_args[1].content