local-RAG-backend

"""設定のテスト""" import pytest from unittest.mock import patch import os from src.main.settings import load_config, AppConfig class TestSettings: """設定のテストクラス""" def test_load_config_全ての環境変数が設定されている場合_適切な設定が返されること( self, ): # ------------------------------ # 準備 (Arrange) # ------------------------------ env_vars = { "NEO4J_URI": "bolt://localhost:7687", "NEO4J_USER": "neo4j", "NEO4J_PASSWORD": "password", "LLM_MODEL_URL": "http://localhost:4000/v1", "LLM_MODEL_NAME": "claude-sonnet-4", "LLM_MODEL_KEY": "sk-1234", "EMBEDDING_MODEL_URL": "http://localhost:11434/v1", "EMBEDDING_MODEL_NAME": "kun432/cl-nagoya-ruri-large:latest", "EMBEDDING_MODEL_KEY": "dummy", "GROUP_ID": "test-group", "CHUNK_SIZE_MAX": "1500", "CHUNK_SIZE_MIN": "150", "CHUNK_OVERLAP": "50", } # ------------------------------ # 実行 (Act) # ------------------------------ with patch.dict(os.environ, env_vars, clear=True): config = load_config() # OPENAI_API_KEYが自動設定されることをpatch内で検証 openai_api_key = os.getenv("OPENAI_API_KEY") # ------------------------------ # 検証 (Assert) # ------------------------------ assert isinstance(config, AppConfig) # Neo4j設定 assert config.neo4j.uri == "bolt://localhost:7687" assert config.neo4j.user == "neo4j" assert config.neo4j.password == "password" # LLM設定 assert config.llm.url == "http://localhost:4000/v1" assert config.llm.name == "claude-sonnet-4" assert config.llm.key == "sk-1234" # 埋め込み設定 assert config.embedding.url == "http://localhost:11434/v1" assert config.embedding.name == "kun432/cl-nagoya-ruri-large:latest" assert config.embedding.key == "dummy" # チャンク設定 assert config.chunk.max_size == 1500 assert config.chunk.min_size == 150 assert config.chunk.overlap == 50 # グループID設定 assert config.group_id == "test-group" # OPENAI_API_KEYが自動設定されること assert openai_api_key == "sk-1234" def test_load_config_チャンク設定が省略されている場合_デフォルト値が使用されること( self, ): # ------------------------------ # 準備 (Arrange) # ------------------------------ env_vars = { "NEO4J_URI": "bolt://localhost:7687", "NEO4J_USER": "neo4j", "NEO4J_PASSWORD": "password", "LLM_MODEL_URL": "http://localhost:4000/v1", "LLM_MODEL_NAME": "claude-sonnet-4", "LLM_MODEL_KEY": "sk-1234", "EMBEDDING_MODEL_URL": "http://localhost:11434/v1", "EMBEDDING_MODEL_NAME": "kun432/cl-nagoya-ruri-large:latest", "EMBEDDING_MODEL_KEY": "dummy", "GROUP_ID": "test-group", # チャンク設定を省略 } # ------------------------------ # 実行 (Act) # ------------------------------ with patch.dict(os.environ, env_vars, clear=True): config = load_config() # ------------------------------ # 検証 (Assert) # ------------------------------ # デフォルト値が使用されること assert config.chunk.max_size == 1000 assert config.chunk.min_size == 100 assert config.chunk.overlap == 0 # グループID設定 assert config.group_id == "test-group" def test_load_config_必須環境変数が不足している場合_例外が発生すること(self): # ------------------------------ # 準備 (Arrange) # ------------------------------ # NEO4J_URIを省略 env_vars = { "NEO4J_USER": "neo4j", "NEO4J_PASSWORD": "password", "LLM_MODEL_URL": "http://localhost:4000/v1", "LLM_MODEL_NAME": "claude-sonnet-4", "LLM_MODEL_KEY": "sk-1234", "EMBEDDING_MODEL_URL": "http://localhost:11434/v1", "EMBEDDING_MODEL_NAME": "kun432/cl-nagoya-ruri-large:latest", "EMBEDDING_MODEL_KEY": "dummy", "GROUP_ID": "test-group", } # ------------------------------ # 実行 (Act) & 検証 (Assert) # ------------------------------ with patch.dict(os.environ, env_vars, clear=True): with pytest.raises(ValueError, match="必須の環境変数が設定されていません"): load_config() def test_load_config_複数の必須環境変数が不足している場合_すべてが報告されること( self, ): # ------------------------------ # 準備 (Arrange) # ------------------------------ # 複数の環境変数を省略 env_vars = { "NEO4J_USER": "neo4j", "LLM_MODEL_URL": "http://localhost:4000/v1", "EMBEDDING_MODEL_URL": "http://localhost:11434/v1", "GROUP_ID": "test-group", # NEO4J_URI, NEO4J_PASSWORD, LLM_MODEL_NAME, LLM_MODEL_KEY, # EMBEDDING_MODEL_NAME, EMBEDDING_MODEL_KEY を省略 } # ------------------------------ # 実行 (Act) & 検証 (Assert) # ------------------------------ with patch.dict(os.environ, env_vars, clear=True): with pytest.raises(ValueError) as exc_info: load_config() error_message = str(exc_info.value) assert "NEO4J_URI" in error_message assert "NEO4J_PASSWORD" in error_message assert "LLM_MODEL_NAME" in error_message assert "LLM_MODEL_KEY" in error_message assert "EMBEDDING_MODEL_NAME" in error_message assert "EMBEDDING_MODEL_KEY" in error_message def test_load_config_OPENAI_API_KEYが未設定の場合_LLM_MODEL_KEYと同じ値が設定されること( self, ): # ------------------------------ # 準備 (Arrange) # ------------------------------ env_vars = { "NEO4J_URI": "bolt://localhost:7687", "NEO4J_USER": "neo4j", "NEO4J_PASSWORD": "password", "LLM_MODEL_URL": "http://localhost:4000/v1", "LLM_MODEL_NAME": "claude-sonnet-4", "LLM_MODEL_KEY": "sk-test-key", "EMBEDDING_MODEL_URL": "http://localhost:11434/v1", "EMBEDDING_MODEL_NAME": "kun432/cl-nagoya-ruri-large:latest", "EMBEDDING_MODEL_KEY": "dummy", "GROUP_ID": "test-group", # OPENAI_API_KEY は設定しない } # ------------------------------ # 実行 (Act) # ------------------------------ with patch.dict(os.environ, env_vars, clear=True): load_config() # OPENAI_API_KEYが自動設定されることをpatch内で検証 openai_api_key = os.getenv("OPENAI_API_KEY") # ------------------------------ # 検証 (Assert) # ------------------------------ # OPENAI_API_KEYがLLM_MODEL_KEYと同じ値に設定されること assert openai_api_key == "sk-test-key" def test_load_config_OPENAI_API_KEYが既に設定されている場合_上書きされないこと( self, ): # ------------------------------ # 準備 (Arrange) # ------------------------------ env_vars = { "NEO4J_URI": "bolt://localhost:7687", "NEO4J_USER": "neo4j", "NEO4J_PASSWORD": "password", "LLM_MODEL_URL": "http://localhost:4000/v1", "LLM_MODEL_NAME": "claude-sonnet-4", "LLM_MODEL_KEY": "sk-llm-key", "EMBEDDING_MODEL_URL": "http://localhost:11434/v1", "EMBEDDING_MODEL_NAME": "kun432/cl-nagoya-ruri-large:latest", "EMBEDDING_MODEL_KEY": "dummy", "GROUP_ID": "test-group", "OPENAI_API_KEY": "sk-existing-key", # 既に設定されている } # ------------------------------ # 実行 (Act) # ------------------------------ with patch.dict(os.environ, env_vars, clear=True): load_config() # OPENAI_API_KEYが既存の値を維持することをpatch内で検証 openai_api_key = os.getenv("OPENAI_API_KEY") # ------------------------------ # 検証 (Assert) # ------------------------------ # OPENAI_API_KEYが既存の値を維持すること assert openai_api_key == "sk-existing-key"