write-tests

作者: microsoft

编写单元测试的指南。在创建测试以验证Python逻辑时使用此指南。

npx skills add https://github.com/microsoft/semantic-link-labs --skill write-tests

Writing Tests

This skill covers how to write tests for the Semantic Link Labs project.

When to Use This Skill

Use this skill when you need to:

  • Write tests for new functions
  • Add test coverage for existing code
  • Test error handling and edge cases
  • Validate input parsing and data transformations

Test Framework

ComponentDetails
Frameworkpytest
Locationtests/ directory
AssertionsStandard pytest assertions

Test File Structure

tests/
├── __init__.py
├── test_helper_functions.py
├── test_workspaces.py
├── test_admin.py
└── ...

Naming Conventions

  • Test files: test_*.py or *_test.py
  • Test functions: test_<function_name>_<scenario>
  • Test classes: Test<ClassName>

Writing Basic Tests

Simple Function Test

import pytest
import pandas as pd


def test_my_function_returns_dataframe():
    """Test that my_function returns a DataFrame."""
    from sempy_labs import my_function

    result = my_function()

    assert isinstance(result, pd.DataFrame)


def test_my_function_has_expected_columns():
    """Test that result has expected columns."""
    from sempy_labs import my_function

    result = my_function()

    expected_columns = ["Id", "Name", "Type"]
    for col in expected_columns:
        assert col in result.columns

Test with Parameters

def test_my_function_filters_by_type():
    """Test that my_function filters by item_type."""
    from sempy_labs import my_function

    result = my_function(item_type="Report")

    assert all(result["Type"] == "Report")

Testing Error Handling

Expected Exceptions

def test_my_function_raises_on_invalid_workspace():
    """Test that invalid workspace raises ValueError."""
    from sempy_labs import my_function

    with pytest.raises(ValueError, match="Invalid workspace"):
        my_function(workspace="NonExistent")


def test_my_function_raises_on_missing_parameter():
    """Test that missing required parameter raises error."""
    from sempy_labs import my_function

    with pytest.raises(TypeError):
        my_function()  # Missing required parameter

Exception Message Matching

def test_error_message_is_descriptive():
    """Test that error message contains helpful information."""
    from sempy_labs import my_function

    with pytest.raises(ValueError) as exc_info:
        my_function(invalid_param="bad")

    assert "invalid_param" in str(exc_info.value)
    assert "bad" in str(exc_info.value)

Using Fixtures

Basic Fixture

import pytest
import pandas as pd


@pytest.fixture
def sample_dataframe():
    """Create a sample DataFrame for testing."""
    return pd.DataFrame({
        "Id": ["1", "2", "3"],
        "Name": ["Item A", "Item B", "Item C"],
        "Type": ["Report", "Dataset", "Report"],
    })


def test_filter_function(sample_dataframe):
    """Test filtering function with sample data."""
    from sempy_labs._helper_functions import filter_items

    result = filter_items(sample_dataframe, type="Report")

    assert len(result) == 2
    assert all(result["Type"] == "Report")

Fixture with Parameters

@pytest.fixture(params=["Report", "Dataset", "Lakehouse"])
def item_type(request):
    """Parameterized fixture for item types."""
    return request.param


def test_with_different_item_types(item_type):
    """Test function with different item types."""
    from sempy_labs import my_function

    result = my_function(item_type=item_type)

    assert isinstance(result, pd.DataFrame)

Mocking External Dependencies

Mocking API Calls

from unittest.mock import patch, MagicMock


def test_function_with_mocked_api():
    """Test with mocked API response."""
    mock_response = MagicMock()
    mock_response.json.return_value = {
        "value": [
            {"id": "123", "name": "Test Item"}
        ]
    }
    mock_response.status_code = 200

    with patch('sempy_labs._helper_functions._base_api') as mock_api:
        mock_api.return_value = mock_response

        from sempy_labs import list_items
        result = list_items()

        assert len(result) == 1
        assert result.iloc[0]["Name"] == "Test Item"

Mocking Fabric Client

from unittest.mock import patch


def test_function_with_mocked_fabric():
    """Test with mocked sempy.fabric calls."""
    with patch('sempy.fabric.resolve_workspace_id') as mock_resolve:
        mock_resolve.return_value = "12345678-1234-1234-1234-123456789012"

        from sempy_labs._helper_functions import resolve_workspace_id
        result = resolve_workspace_id("My Workspace")

        mock_resolve.assert_called_once()

Parameterized Tests

Multiple Input Values

import pytest


@pytest.mark.parametrize("input_value,expected", [
    ("value1", "result1"),
    ("value2", "result2"),
    ("value3", "result3"),
])
def test_function_with_multiple_inputs(input_value, expected):
    """Test function with multiple input values."""
    from sempy_labs import my_function

    result = my_function(input_value)

    assert result == expected

Testing Edge Cases

@pytest.mark.parametrize("workspace", [
    None,           # Default workspace
    "My Workspace", # By name
    "12345678-1234-1234-1234-123456789012",  # By UUID string
])
def test_accepts_various_workspace_formats(workspace):
    """Test that function accepts various workspace formats."""
    from sempy_labs import my_function

    # Should not raise
    result = my_function(workspace=workspace)
    assert result is not None

Testing DataFrame Results

Column Validation

def test_result_has_required_columns():
    """Test that result DataFrame has required columns."""
    from sempy_labs import list_items

    result = list_items()

    required_columns = ["Id", "Name", "Type"]
    for col in required_columns:
        assert col in result.columns, f"Missing column: {col}"

Data Type Validation

def test_result_column_types():
    """Test that result columns have correct types."""
    from sempy_labs import list_items

    result = list_items()

    assert result["Id"].dtype == "object"  # string
    assert result["Name"].dtype == "object"  # string

Empty Result Handling

def test_handles_empty_result():
    """Test that function handles empty results gracefully."""
    from sempy_labs import list_items

    result = list_items(item_type="NonExistentType")

    assert isinstance(result, pd.DataFrame)
    assert result.empty
    # Columns should still exist even if empty
    assert "Id" in result.columns

Test Organization

Grouping Related Tests

class TestWorkspaceFunctions:
    """Tests for workspace-related functions."""

    def test_list_workspaces(self):
        """Test listing workspaces."""
        pass

    def test_resolve_workspace_id(self):
        """Test resolving workspace ID."""
        pass

    def test_resolve_workspace_name(self):
        """Test resolving workspace name."""
        pass


class TestHelperFunctions:
    """Tests for helper utility functions."""

    def test_is_valid_uuid(self):
        """Test UUID validation."""
        pass

    def test_create_dataframe(self):
        """Test DataFrame creation helper."""
        pass

Best Practices

Do's

  • ✅ Use descriptive test names
  • ✅ Test one thing per test function
  • ✅ Include docstrings explaining what's being tested
  • ✅ Use fixtures for reusable setup
  • ✅ Test both success and failure cases
  • ✅ Mock external dependencies when needed

Don'ts

  • ❌ Don't test multiple behaviors in one test
  • ❌ Don't rely on external services in unit tests
  • ❌ Don't use hardcoded secrets or credentials
  • ❌ Don't write tests that depend on test execution order
  • ❌ Don't ignore flaky tests - fix them

Pre-Commit Test Checklist

Before committing new tests:

  1. Run the tests locally:

    pytest -sv tests/ -k my_new_test
    
  2. Verify tests pass consistently (run multiple times)

  3. Check test coverage for the new code

  4. Ensure tests are independent and don't rely on each other

来自 microsoft 的更多技能

oss-growth
microsoft
OSS增长黑客角色
agent-framework-azure-ai-py
microsoft
使用Microsoft Agent Framework Python SDK(agent-framework-azure-ai)构建Azure AI Foundry代理。在创建使用AzureAIAgentsProvider的持久化代理、使用托管工具(代码解释器、文件搜索、网络搜索)、集成MCP服务器、管理对话线程或实现流式响应时使用。涵盖函数工具、结构化输出和多工具代理。
development
airunway-aks-setup
microsoft
在AKS上设置AI Runway——从裸集群到运行模型。涵盖集群验证、控制器安装、GPU评估、提供商设置和首次部署。适用场景:“设置AI Runway”、“接入AKS集群”、“安装AI Runway”、“airunway设置”、“将模型部署到AKS”、“在AKS上进行GPU推理”、“在AKS上配置KAITO”、“在AKS上运行LLM”、“在AKS上使用vLLM”、“在AKS上设置模型服务”、“AI Runway控制器”。
devops
appinsights-instrumentation
microsoft
使用Azure Application Insights对Web应用进行插桩的指南。提供遥测模式、SDK设置和配置参考。适用场景:如何对应用进行插桩、App Insights SDK、遥测模式、什么是App Insights、Application Insights指南、插桩示例、APM最佳实践。
devops
applicationinsights-web-ts
microsoft
使用Application Insights JavaScript SDK(@microsoft/applicationinsights-web)为浏览器/Web应用添加检测。用于真实用户监控(RUM)——页面视图、点击、AJAX/fetch依赖项、异常、自定义事件,以及与后端OpenTelemetry追踪关联的浏览器端GenAI代理追踪。涵盖SDK加载器脚本和npm设置、框架扩展(React、React Native、Angular)、点击分析、遥测初始化器,以及从浏览器发出的代理/工具/模型跨度所遵循的OTel GenAI语义约定。
devops
azure-ai-anomalydetector-java
microsoft
使用适用于 Java 的 Azure AI 异常检测器 SDK 构建异常检测应用程序。在实现单变量/多变量异常检测、时间序列分析或 AI 驱动的监控时使用。
development
azure-ai-language-conversations-py
microsoft
使用azure-ai-language-conversations Python SDK实现对话语言理解(CLU)。当使用ConversationAnalysisClient分析对话意图和实体、构建NLP功能或将语言理解集成到应用程序中时使用。
development
azure-ai-ml-py
microsoft
Azure Machine Learning SDK v2 for Python。用于机器学习工作区、作业、模型、数据集、计算资源和管道。 触发词:“azure-ai-ml”、“MLClient”、“工作区”、“模型注册表”、“训练作业”、“数据集”。
development