feat: implement mvp with email-first login flow and langgraph architecture

tests/test_multi_turn_planner_researcher.py

@@ -0,0 +1,79 @@
+import pytest
+from unittest.mock import MagicMock, patch
+from langchain_core.messages import HumanMessage, AIMessage
+from ea_chatbot.graph.nodes.planner import planner_node
+from ea_chatbot.graph.nodes.researcher import researcher_node
+from ea_chatbot.graph.nodes.summarizer import summarizer_node
+from ea_chatbot.schemas import TaskPlanResponse
+
+@pytest.fixture
+def mock_state_with_history():
+    return {
+        "messages": [
+            HumanMessage(content="Show me the 2024 results for Florida"),
+            AIMessage(content="Here are the results for Florida in 2024...")
+        ],
+        "question": "What about in New Jersey?",
+        "analysis": {"data_required": ["2024 results", "New Jersey"], "unknowns": [], "ambiguities": [], "conditions": []},
+        "next_action": "plan",
+        "summary": "The user is asking about 2024 election results.",
+        "plan": "Plan steps...",
+        "code_output": "Code output..."
+    }
+
+@patch("ea_chatbot.graph.nodes.planner.get_llm_model")
+@patch("ea_chatbot.utils.database_inspection.get_data_summary")
+@patch("ea_chatbot.graph.nodes.planner.PLANNER_PROMPT")
+def test_planner_uses_history_and_summary(mock_prompt, mock_get_summary, mock_get_llm, mock_state_with_history):
+    mock_get_summary.return_value = "Data summary"
+    mock_llm_instance = MagicMock()
+    mock_get_llm.return_value = mock_llm_instance
+    mock_structured_llm = MagicMock()
+    mock_llm_instance.with_structured_output.return_value = mock_structured_llm
+    
+    mock_structured_llm.invoke.return_value = TaskPlanResponse(
+        goal="goal",
+        reflection="reflection",
+        context={
+            "initial_context": "context",
+            "assumptions": [],
+            "constraints": []
+        },
+        steps=["Step 1: test"]
+    )
+    
+    planner_node(mock_state_with_history)
+    
+    mock_prompt.format_messages.assert_called_once()
+    kwargs = mock_prompt.format_messages.call_args[1]
+    assert kwargs["question"] == "What about in New Jersey?"
+    assert kwargs["summary"] == mock_state_with_history["summary"]
+    assert len(kwargs["history"]) == 2
+
+@patch("ea_chatbot.graph.nodes.researcher.get_llm_model")
+@patch("ea_chatbot.graph.nodes.researcher.RESEARCHER_PROMPT")
+def test_researcher_uses_history_and_summary(mock_prompt, mock_get_llm, mock_state_with_history):
+    mock_llm_instance = MagicMock()
+    mock_get_llm.return_value = mock_llm_instance
+    
+    researcher_node(mock_state_with_history)
+    
+    mock_prompt.format_messages.assert_called_once()
+    kwargs = mock_prompt.format_messages.call_args[1]
+    assert kwargs["question"] == "What about in New Jersey?"
+    assert kwargs["summary"] == mock_state_with_history["summary"]
+    assert len(kwargs["history"]) == 2
+
+@patch("ea_chatbot.graph.nodes.summarizer.get_llm_model")
+@patch("ea_chatbot.graph.nodes.summarizer.SUMMARIZER_PROMPT")
+def test_summarizer_uses_history_and_summary(mock_prompt, mock_get_llm, mock_state_with_history):
+    mock_llm_instance = MagicMock()
+    mock_get_llm.return_value = mock_llm_instance
+    
+    summarizer_node(mock_state_with_history)
+    
+    mock_prompt.format_messages.assert_called_once()
+    kwargs = mock_prompt.format_messages.call_args[1]
+    assert kwargs["question"] == "What about in New Jersey?"
+    assert kwargs["summary"] == mock_state_with_history["summary"]
+    assert len(kwargs["history"]) == 2