[WIP] Add comprehensive pytest test suite for SGR Deep Research

[hijera]

Add Comprehensive Pytest Test Suite

📊 Overview

This PR introduces a comprehensive test suite for the SGR Deep Research project, providing extensive coverage of all core components with 504 automated tests that execute in 1.75 seconds.

Why i'm doing it

Because this project need good test coverage

✨ What's Added

Test Coverage by Component

Component	Test File(s)	Tests	Coverage
Data Models	test_models.py	27	Core models: SourceData, SearchResult, AgentStatesEnum, ResearchContext
API Models	test_api_models.py	34	All API request/response models and enums
Configuration	test_settings.py	43	All Pydantic config models with validation
Prompts	test_prompts.py	13	PromptLoader with caching and fallbacks
Tavily Service	test_tavily_service.py	18	Search and extract with mocked API calls
Tools	test_web_search_tool.py, test_extract_page_tool.py, test_create_report_tool.py	50	Core research tools with context updates
Planning Tools	test_*_tool_extended.py	150	Extended tests for all planning/reasoning tools
BaseTool	test_base_tool_extended.py	40	Metaclass, inheritance, validation
Streaming	test_streaming.py	42	SSE format, OpenAI chunks, async generators
MCP	test_mcp.py	28	Dynamic tool creation, singleton pattern
BaseAgent	test_base_agent.py	27	Initialization, logging, abstract methods
API Endpoints	test_api_endpoints.py	20	Fastapi endpoints, schemas and responses
Agent Factory	test_agent_factory.py	17	Agent Factory and configs
SO Tools Calling Agent	test_sgr_so_tool_calling_agent.py	7	Agent with tool calls
Existing	test_tools.py	35	Original tool tests (preserved)
TOTAL	18 files	579	✅ All pass

🎯 Test Features

✅ Comprehensive Testing Approach

• Pydantic Validation: Boundary values, required fields, type checking
• Async Operations: Proper async/await handling with mocked async methods
• File I/O: Temporary directories for isolated file operations
• Edge Cases: Unicode, special characters, long strings, boundary values
• Mocking: External APIs (Tavily, OpenAI, FastMCP, httpx) for fast, reliable tests
• Inheritance: Multi-level inheritance, metaclasses, class variables
• JSON Serialization: Correct formatting with ensure_ascii=False

⚡ Performance

• 504 tests execute in 2.16 seconds
• All tests are isolated and independent
• No real API calls or network requests
• Uses mocks for external dependencies

🔒 Test Isolation

• Each test uses fresh instances
• Temporary directories for file operations
• Mocked external services prevent side effects
• No shared state between tests

🚀 How to Run Tests

Prerequisites

# Install dependencies (if not already installed)
pip install -e .

Run All Tests

# Option 1: Using pytest directly
pytest tests/ -v

# Option 2: Using Python module
python -m pytest tests/ -v

# Option 3: Quick summary
pytest tests/ -q

Run Specific Test Categories

# Models and configuration
pytest tests/test_models.py tests/test_api_models.py tests/test_settings.py -v

# Services
pytest tests/test_prompts.py tests/test_tavily_service.py -v

# Tools
pytest tests/test_*_tool*.py -v

# Streaming and MCP
pytest tests/test_streaming.py tests/test_mcp.py -v

# Agents
pytest tests/test_base_agent.py -v

Run with Coverage Report

pytest tests/ --cov=sgr_deep_research --cov-report=html

CI/CD Integration

Tests are designed to work with the existing CI/CD pipeline. The linter workflow will automatically validate code style on PR.

[hijera]

Обновил тесты, апдейтнул, 504 -> 579 тестов.

[hijera]

Поправил .gitignore , чтобы добавились новые тесты, и также закоммитил их

[hijera]

Выполнил pre-commit run -a и пофиксил под него.