Added tests and test suite - about 60% coverage

[Max Caldwell (maxtheman)]

Description:

Adds comprehensive test suite for the create_deep_agent function in the deepagents library. The tests verify the core functionality of agent creation with various configurations

Changes:

• Created tests/ directory with init.py for proper Python package structure

• Added test_create_deep_agent.py with 7 test cases - 60% coverage

• Added pytest configuration to pyproject.toml with test paths and discovery settings

• Added dev dependencies (pytest>=8.0.0, pytest-cov>=6.0.0) as optional dependencies in pyproject.toml

• Fixed license format in pyproject.toml from table to string to resolve setuptools deprecation warning

• Added CLAUDE.md with test running instructions and current project status

• Added Development section to README.md with testing and build instructions

  Dependencies:

  • pytest>=8.0.0 (dev dependency)
  • pytest-cov>=6.0.0 (dev dependency)

  Testing:

  • Verified all 7 tests pass successfully
  • Tested that package builds successfully without warnings
  • Verified built wheel can be installed and imported correctly
  • Confirmed test instructions in README work as documented

platform darwin -- Python 3.11.13, pytest-8.4.1, pluggy-1.6.0 -- /Users/max/Documents/deepagents/venv/bin/python
cachedir: .pytest_cache
rootdir: /Users/max/Documents/deepagents
configfile: pyproject.toml
testpaths: tests
plugins: anyio-4.9.0, langsmith-0.4.10, cov-6.2.1
collected 7 items

tests/test_create_deep_agent.py::TestCreateDeepAgent::test_create_deep_agent_happy_path PASSED                                                                                       [ 14%]
tests/test_create_deep_agent.py::TestCreateDeepAgent::test_create_deep_agent_with_subagents PASSED                                                                                   [ 28%]
tests/test_create_deep_agent.py::TestCreateDeepAgent::test_create_deep_agent_with_custom_model PASSED                                                                                [ 42%]
tests/test_create_deep_agent.py::TestCreateDeepAgent::test_create_deep_agent_with_custom_state_schema PASSED                                                                         [ 57%]
tests/test_create_deep_agent.py::TestCreateDeepAgent::test_create_deep_agent_with_multiple_tools PASSED                                                                              [ 71%]
tests/test_create_deep_agent.py::TestCreateDeepAgent::test_create_deep_agent_empty_tools PASSED                                                                                      [ 85%]
tests/test_create_deep_agent.py::TestCreateDeepAgent::test_create_deep_agent_complex_scenario PASSED                                                                                 [100%]

====================================================================================== tests coverage ======================================================================================
____________________________________________________________________ coverage: platform darwin, python 3.11.13-final-0 _____________________________________________________________________

Name                                                        Stmts   Miss  Cover
-------------------------------------------------------------------------------
venv/lib/python3.11/site-packages/deepagents/__init__.py        3      0   100%
venv/lib/python3.11/site-packages/deepagents/graph.py          20      0   100%
venv/lib/python3.11/site-packages/deepagents/model.py           3      1    67%
venv/lib/python3.11/site-packages/deepagents/prompts.py         5      0   100%
venv/lib/python3.11/site-packages/deepagents/state.py          16      5    69%
venv/lib/python3.11/site-packages/deepagents/sub_agent.py      37      7    81%
venv/lib/python3.11/site-packages/deepagents/tools.py          59     44    25%
-------------------------------------------------------------------------------
TOTAL                                                         143     57    60%

[Jason Madruga (ASTXRTYS)]

✅ VERIFICATION COMPLETED - The /tokens command accurately emits actual token counts for the agent with 99.4% accuracy. Comprehensive testing confirms the token counting system correctly accounts for all middleware-injected content and uses the model's official tokenizer. No changes required.

[vivek (vtrivedy)]

not needed, thanks for this. in general best to avoid committing and relying on claude.md directly for this repo