GTA Integration Tests

Requirements

• Python 3.11+
• Install dependencies: pip install pytest pytest-html requests pyyaml pyjwt cryptography

Configuration

Edit config.yaml with the actual addresses of GTA Server and Agent:

server:
  host: "127.0.0.1"
  port: 8080

agent:
  host: "127.0.0.1"
  port: 8090

# Minimum seconds between consecutive RESTful requests from a single client,
# to avoid the service rejecting requests fired too fast.
request_interval: 0.5

Port reachability is checked automatically on the first request. A misconfigured address will cause an immediate error.

Running Tests

Use run_tests.sh to run by level, default is p1:

./run_tests.sh       # equivalent to p1, runs p0 + p1
./run_tests.sh p0    # smoke tests only, fastest
./run_tests.sh p1    # p0 + p1, daily CI
./run_tests.sh p2    # p0 + p1 + p2, pre-release
./run_tests.sh p3    # full suite, includes edge cases

You can also invoke pytest directly, e.g. to run a single module:

pytest tests/service_challenge/
pytest tests/ -m "p0 or p1" -k "nonce"

The test report is written to report.html by default.

Test Levels

Mark	Description	When to run
p0	Smoke tests	Every commit
p1	Core regression	Daily CI
p2	Full regression	Before release
p3	Edge cases & error paths	Periodic execution

Directory Structure

integration-tests/
├── config.yaml                  # Environment config (service addresses)
├── config.py                    # Global config singleton
├── pytest.ini                   # Mark registration, log format
├── run_tests.sh                 # Test runner entry point
├── services/
│   ├── api_client.py            # Base HTTP client (BaseApiClient)
│   ├── gta_server.py            # GTA Server API wrappers
│   └── gta_agent.py             # GTA Agent API wrappers
└── tests/
    ├── base_test.py             # Base test class (provides self.server / self.agent)
    ├── service_challenge/       # Server-side challenge test cases
    └── collect_evidence/        # Agent evidence collection test cases

Adding Test Cases

Add to an existing module

Add a method to the relevant test class in the corresponding test_*.py file:

@pytest.mark.p1
def test_my_new_case(self):
    """P1: describe what this case verifies"""
    resp = self.server.get_challenge(attester_type=["tpm_boot"])
    assert resp.status_code == 200

Add a new module

1. Create a directory under tests/ and add __init__.py:

   tests/
   └── my_feature/
       ├── __init__.py
       └── test_my_feature.py
   

2. Write a test class that inherits BaseIntegrationTest:

   import pytest
   from tests.base_test import BaseIntegrationTest
   
   class TestMyFeature(BaseIntegrationTest):
       __test__ = True
   
       @pytest.mark.p0
       def test_basic(self):
           """P0: description"""
           resp = self.server.some_api()
           assert resp.status_code == 200
   

   __test__ = True is required. The base class sets __test__ = False to prevent pytest from collecting it directly.

3. If you need to call an endpoint not yet wrapped in GtaServerService / GtaAgentService, use the underlying HTTP methods directly:

   resp = self.server.post("/global-trust-authority/service/v1/some-path", json={...})
   resp = self.agent.get("/global-trust-authority/agent/v1/some-path")
   

4. If the endpoint will be reused across multiple tests, wrap it as a method in services/gta_server.py or services/gta_agent.py.