conftest.py

Overview

The conftest.py file is a pytest configuration and fixture setup module for the InfiniFlow testing environment. It provides command-line options to customize test runs, handles user registration and authentication, manages API tokens, and configures tenant-specific settings and model integrations for testing workflows.

This file is essential for initializing the test environment, ensuring the necessary user context and system state are established before any tests are executed. It interacts with the InfiniFlow backend API to automate user lifecycle management, token generation, and model registration, facilitating smooth and repeatable test executions.

Detailed Explanation

Global Constants

• MARKER_EXPRESSIONS
  A dictionary mapping test level keys (p1, p2, p3) to pytest marker expressions. It controls which tests run based on their level of priority or scope:

  • p1: smoke tests

  • p2: core tests (default)

  • p3: full test suite

pytest_addoption(parser: pytest.Parser) -> None

Purpose:
Adds custom command-line options to pytest to control test behavior.

Parameters:

• parser: pytest's option parser object.

Behavior:

• Adds --level option to select the test level (p1, p2, or p3), defaulting to p2.

• Adds --client-type option to specify the client type used in tests (python_sdk, http, web), defaulting to http.

Usage Example:

pytest --level=p1 --client-type=python_sdk

pytest_configure(config: pytest.Config) -> None

Purpose:
Configures pytest before tests start based on the command-line options.

Parameters:

• config: pytest configuration object.

Behavior:

• Sets the marker expression for filtering tests according to the --level option.

• Prints the active test level if verbosity is enabled.

register() -> None

Purpose:
Registers a user for testing if not already registered.

Behavior:

• Sends a POST request to the /user/register endpoint with email, nickname ("qa"), and password.

• If registration fails for reasons other than "already registered", raises an exception with the error message.

login() -> str

Purpose:
Logs in the registered user and retrieves an authorization token.

Returns:

• auth (str): Authorization token string from the response header.

Behavior:

• Sends a POST request to /user/login with email and password.

• Raises an exception if login fails.

• Returns the Authorization header for authenticated requests.

auth pytest fixture

Scope: session

Purpose:
Provides an authenticated token for the test session.

Behavior:

• Calls register() to ensure user existence (catches and logs exceptions).

• Calls login() to fetch an authorization token.

• Returns the authorization token.

Usage Example:

def test_example(auth):
    # auth is an authorization token string
    pass

token pytest fixture

Scope: session

Purpose:
Generates a new system token by calling the backend after authentication.

Parameters:

• auth: Authorization token from the auth fixture.

Returns:

• A new token string extracted from the response JSON.

Behavior:

• Sends a POST request to /system/new_token with the authorization header.

• Raises an exception on failure.

get_my_llms(auth: str, name: str) -> bool

Purpose:
Checks if a specified LLM (Large Language Model) is registered for the current user.

Parameters:

• auth: Authorization token string.

• name: Name of the LLM model to check.

Returns:

• True if the model exists for the user, otherwise False.

Behavior:

• Sends a GET request to /llm/my_llms.

• Parses the response data to verify presence of the model name.

add_models(auth: str) -> None

Purpose:
Registers predefined models with the backend for the authenticated user if not already present.

Parameters:

• auth: Authorization token string.

Behavior:

• Defines a dictionary models_info with model names and their API keys.

• For each model, checks if already registered (get_my_llms).

• If not registered, sends a POST request to /llm/set_api_key.

• If registration fails, calls pytest.exit() to abort testing.

get_tenant_info(auth: str) -> str

Purpose:
Retrieves tenant information for the authenticated user.

Parameters:

• auth: Authorization token string.

Returns:

• Tenant ID (string).

Behavior:

• Sends a GET request to /user/tenant_info.

• Raises exception on failure.

set_tenant_info pytest fixture

Scope: session, autouse=True

Purpose:
Automatically prepares tenant information and links necessary LLM and service IDs for the test session.

Parameters:

• auth: Authorization token string.

Behavior:

• Calls add_models() to ensure models are registered.

• Calls get_tenant_info() to fetch tenant ID.

• Sends a POST request to /user/set_tenant_info with tenant and model details.

• On failure, exits pytest with an error.

Important Implementation Details

• The file uses the requests library for HTTP interactions with the InfiniFlow backend API.

• It leverages pytest's fixture and configuration hooks to integrate setup steps seamlessly.

• Exception handling is used to fail early on critical errors (e.g., registration, token retrieval).

• The add_models function dynamically checks and sets API keys for language models required for testing.

• The tenant setup fixture runs automatically, ensuring all tests run with proper tenant context.

• Test selection via --level leverages pytest marker expressions to filter tests by priority or scope.

Interactions with Other System Components

• Backend API: All user, token, tenant, and model management is done through REST API calls to the InfiniFlow backend specified by HOST_ADDRESS and VERSION.

• Configurations: Uses constants (EMAIL, PASSWORD, ZHIPU_AI_API_KEY, etc.) imported from a configs module, which must provide environment-specific credentials and keys.

• Test Suites: Provides fixtures and configuration options consumed by test modules to run tests under authenticated and tenant-aware contexts.

• Models: Manages registration and availability of AI models (e.g., ZHIPU-AI) necessary for model-dependent tests.

Usage Summary

1. Configure test run options using command-line arguments (--level, --client-type).

2. On pytest startup, user registration and login occur automatically.

3. Session-scoped fixtures provide authentication tokens and tenant setup.

4. Tests can access these fixtures to perform authenticated and tenant-aware API calls.

5. Models are registered and tenant information is set automatically before tests run.

Mermaid Flowchart Diagram

flowchart TD
    A[pytest_addoption] --> B[pytest_configure]
    B --> C[register()]
    C --> D[login()]
    D --> E[auth fixture]
    E --> F[token fixture]
    E --> G[add_models()]
    G --> H[get_my_llms()]
    E --> I[get_tenant_info()]
    G & I --> J[set_tenant_info fixture]

Diagram Explanation:

• pytest_addoption and pytest_configure set up test run parameters.

• register and login handle user lifecycle.

• auth fixture provides authorization token.

• token fixture uses auth to get a system token.

• add_models and get_my_llms manage model registration.

• get_tenant_info fetches tenant details.

• set_tenant_info depends on add_models and get_tenant_info to configure tenant settings automatically.

End of Documentation for conftest.py