AlgoPytest is a Pytest plugin framework which hides away all of the complexity and repetitiveness that comes with testing Algorand Smart Contracts.
A lot of boilerplate code is necessary in order to setup a suitable test environment. AlgoPytest takes care of all that. It handles deploying the smart contract, creating and funding any necessary user accounts and then using those accounts to interact with the smart contract itself. Additionally, each test is a run in a freshly deployed smart contract, facilitating a clean slate which prevents any stateful interference from any previously run tests.
The most relevant information needed for getting off the ground are:
This project's Read-The-Docs page is especially useful, since it lists and describes all of the available methods and fixtures provided to you by AlgoPytest.
- Python 3 and
pip - Install the Algorand sandbox.
pip install algopytest-frameworkMake sure the sandbox is up and running. Preferably use a local network for your testing such as dev.
# Spin up the `dev` network
./sandbox up devSet any necessary environment variables.
# The path to your sandbox so that AlgoPytest may interact with the sandbox
export SANDBOX_DIR=/path/to/installation/of/sandbox/
# The address in your `sandbox` which was allocated the initial funds
export INITIAL_FUNDS_ACCOUNT=4BJAN3J32NDMZJWU3DPIPGCPBQIUTXL3UB2LEG5Z3CFPRJZOOZC2GH5DMQ- List of environment variables recognized by AlgoPytest: documentation.
Create a conftest.py file in your Pytest tests directory and define a fixture deploying a smart contract as so:
# File: conftest.py
import pytest
from algopytest import create_app
# Load the smart contracts from this project. The path to find these
# imports can be set by the environment variable `$PYTHONPATH`.
# e.g. `export PYTHONPATH=.../smart-contract-project/assets`
from approval_program import approval_program
from clear_program import clear_program
# NOTE: It is critical to `yield` the `app_id` so that the clean up is properly performed.
# The `owner` is an automatically available AlgoPytest defined fixture
@pytest.fixture
def smart_contract_id(owner):
with create_app(
owner,
approval_program=diploma_program(),
clear_program=clear_program(),
local_bytes=1,
local_ints=1,
global_bytes=1,
) as app_id:
yield app_idNow, any Pytest tests you write automatically have access to the standard AlgoPytest fixtures as well as your defined smart_contract_id. Additionally, you can import and utilize the various helper functions that ship with the framework.
- List of available fixtures: documentation
- Provided helper functions: documentation
A simple test to make sure that the creator of the smart contract can update the application is provided below. It uses the AlgoPytest fixture owner, your defined smart_contract_id fixture and the helper function update_app.
# File: test_behavior.py
def test_update_from_owner(owner, smart_contract_id):
update_app(owner, smart_contract_id)This AlgoPytest project includes demos of Algorand Smart Contract projects that utilize this package to implement their test suite. These demo projects give examples of how a real-world project may use AlgoPytest for its testing. They provide greater context for how to integrate AlgoPytest into your project. The tests and the initialization code of the demos may be found within their respective tests directory. Therein, you will see how the fixtures are used to extensively stress test the Smart Contract code and life-cycle.
For example, a semi-involved test in one of the demos, algo-diploma, showcases AlgoPytest utilizing the power of Pytest fixtures:
@pytest.fixture
def owner_in(owner, smart_contract_id):
"""Create an ``owner`` fixture that has already opted in to ``smart_contract_id``."""
opt_in_app(owner, smart_contract_id)
# The test runs here
yield owner
# Clean up by closing out of the application
close_out_app(owner, smart_contract_id)
@pytest.fixture
def user1_in(user1, smart_contract_id):
"""Create a ``user1`` fixture that has already opted in to ``smart_contract_id``."""
opt_in_app(user1, smart_contract_id)
# The test runs here
yield user1
# Clean up by closing out of the application
close_out_app(user1, smart_contract_id)
def test_issue_diploma(owner_in, user1_in, smart_contract_id):
diploma_metadata = "Damian Barabonkov :: MIT :: BSc Computer Science and Engineering :: 2020"
# The application arguments and account to be passed in to
# the smart contract as it expects
app_args = ['issue_diploma', diploma_metadata]
accounts = [user1_in.address]
# Issue the `diploma_metadata` to the recipient `user1`
call_app(owner_in, smart_contract_id, app_args=app_args, accounts=accounts)Original source may be found here.
Refer to the Pytest Documentation References below for more specific explanations of key Pytest topics. These topics are essential to understand in order to use AlgoPytest effectively.
Firstly, you must follow the Pytest directory structure. Essentially, all tests will be found within a tests directory in the root of your project.
Before being able to write Pytest tests for your Algorand Smart Contract, you will need to initialize some essential fixtures such as the smart contracts, signatures, and assets to be tested. Utility functions are provided by AlgoPytest which deploy these objects and return a reference to be used in a fixture. For a smart contract, this is most easily achieved by creating a conftest.py file and yielding the result of algopytest.create_app from within a fixture. It is critical to yield rather than return to not interfere with the AlgoPytest fixture clean up sequence.
For example:
# Sample project file structure
smart-contract-project/
├── assets
│ ├── approval_program.py
│ ├── clear_program.py
└── tests
├── conftest.py
└── test_behavior.py# File: conftest.py
import pytest
from algopytest import create_app
# Load the smart contracts from this project. The path to find these
# imports can be set by the environment variable `$PYTHONPATH`.
# e.g. `export PYTHONPATH=.../smart-contract-project/assets`
from approval_program import approval_program
from clear_program import clear_program
# NOTE: It is critical to `yield` the `app_id` so that the clean up is properly performed.
# The `owner` is an automatically available AlgoPytest defined fixture
@pytest.fixture
def smart_contract_id(owner):
with create_app(
owner,
approval_program=diploma_program(),
clear_program=clear_program(),
local_bytes=1,
local_ints=1,
global_bytes=1,
) as app_id:
yield app_idThe approval_program and clear_program must be Python functions which return a pyteal.Expr.
For example, the simplest clear program in this format is:
# File: clear_program.py
def clear_program():
"""A clear program that always approves."""
return Return(Int(1))The AlgoPytest package is written as a Pytest plugin. This allows AlgoPytest to automatically register fixtures without you needing to import anything. You may simply use the fixtures directly in the function signature.
These fixtures make testing Algorand Smart Contracts significantly easier for you as the developer. Fixtures such as the user ones: owner, user1, user2, etc. automatically create users with funded balances and cleans them up at the end of the test. Any smart contracts, signatures and assets yielded as fixtures are automatically deployed over the life of the test and then cleaned up. Without AlgoPytest, writing a test even as simple as checking whether the owner of an application can update their application requires non-negligible boilerplate code. Now, in order to write this test, all of the necessary boilerplate code is taken care of; you only have to focus only on the testing code at hand and nothing else.
# File: test_behavior.py
def test_update_from_owner(owner, smart_contract_id):
"""The `owner` and `smart_contract_id` are AlgoPytest fixtures.
This call will raise if there is any error."""
update_app(owner, smart_contract_id)- Pytest directory structure: documentation
- Pytest fixtures: documentation
- Pytest
conftest.py: documentation
git clone https://github.com/DamianB-BitFlipper/algopytest.git
cd algopytest
conda env create -f environment.yml
pre-commit install
pip install -e .Please submit a Pull Request for any suggested changes you would like to make.
This package and smart contract(s) in this codebase have NOT been professionally audited. Therefore, they should not be used as a production application.