graphrag/tests/unit/config/test_config.py
KennyZhang1 83cc2daf91
Multi-index query CLI support (#1675)
* Add vector store id reference to embeddings config.

* changed structure of output config section

* added cli integration for multi index global

* added cli integration for multi index local

* added cli integration for multi index drift and basic

* finished local testing of multi-index cli

* ruff fixes

* partially refactored test code to align with new output section

* more test changes for new output structure

* semversioner

* refactored to align with new multi index config proposal

* locally tested new multi-index output proposal

* cleaned up tests to align with new structure

---------

Co-authored-by: Derek Worthen <worthend.derek@gmail.com>
2025-02-07 12:56:48 -05:00

184 lines
6.1 KiB
Python

# Copyright (c) 2024 Microsoft Corporation.
# Licensed under the MIT License
import os
from pathlib import Path
from unittest import mock
import pytest
from pydantic import ValidationError
import graphrag.config.defaults as defs
from graphrag.config.create_graphrag_config import create_graphrag_config
from graphrag.config.enums import AuthType, LLMType
from graphrag.config.load_config import load_config
from tests.unit.config.utils import (
DEFAULT_EMBEDDING_MODEL_CONFIG,
DEFAULT_MODEL_CONFIG,
FAKE_API_KEY,
assert_graphrag_configs,
get_default_graphrag_config,
)
def test_missing_openai_required_api_key() -> None:
model_config_missing_api_key = {
defs.DEFAULT_CHAT_MODEL_ID: {
"type": LLMType.OpenAIChat,
"model": defs.LLM_MODEL,
},
defs.DEFAULT_EMBEDDING_MODEL_ID: DEFAULT_EMBEDDING_MODEL_CONFIG,
}
# API Key required for OpenAIChat
with pytest.raises(ValidationError):
create_graphrag_config({"models": model_config_missing_api_key})
# API Key required for OpenAIEmbedding
model_config_missing_api_key[defs.DEFAULT_CHAT_MODEL_ID]["type"] = (
LLMType.OpenAIEmbedding
)
with pytest.raises(ValidationError):
create_graphrag_config({"models": model_config_missing_api_key})
def test_missing_azure_api_key() -> None:
model_config_missing_api_key = {
defs.DEFAULT_CHAT_MODEL_ID: {
"type": LLMType.AzureOpenAIChat,
"auth_type": AuthType.APIKey,
"model": defs.LLM_MODEL,
"api_base": "some_api_base",
"api_version": "some_api_version",
"deployment_name": "some_deployment_name",
},
defs.DEFAULT_EMBEDDING_MODEL_ID: DEFAULT_EMBEDDING_MODEL_CONFIG,
}
with pytest.raises(ValidationError):
create_graphrag_config({"models": model_config_missing_api_key})
# API Key not required for managed identity
model_config_missing_api_key[defs.DEFAULT_CHAT_MODEL_ID]["auth_type"] = (
AuthType.AzureManagedIdentity
)
create_graphrag_config({"models": model_config_missing_api_key})
def test_conflicting_auth_type() -> None:
model_config_invalid_auth_type = {
defs.DEFAULT_CHAT_MODEL_ID: {
"auth_type": AuthType.AzureManagedIdentity,
"type": LLMType.OpenAIChat,
"model": defs.LLM_MODEL,
},
defs.DEFAULT_EMBEDDING_MODEL_ID: DEFAULT_EMBEDDING_MODEL_CONFIG,
}
with pytest.raises(ValidationError):
create_graphrag_config({"models": model_config_invalid_auth_type})
def test_conflicting_azure_api_key() -> None:
model_config_conflicting_api_key = {
defs.DEFAULT_CHAT_MODEL_ID: {
"type": LLMType.AzureOpenAIChat,
"auth_type": AuthType.AzureManagedIdentity,
"model": defs.LLM_MODEL,
"api_base": "some_api_base",
"api_version": "some_api_version",
"deployment_name": "some_deployment_name",
"api_key": "THIS_SHOULD_NOT_BE_SET_WHEN_USING_MANAGED_IDENTITY",
},
defs.DEFAULT_EMBEDDING_MODEL_ID: DEFAULT_EMBEDDING_MODEL_CONFIG,
}
with pytest.raises(ValidationError):
create_graphrag_config({"models": model_config_conflicting_api_key})
base_azure_model_config = {
"type": LLMType.AzureOpenAIChat,
"auth_type": AuthType.AzureManagedIdentity,
"model": defs.LLM_MODEL,
"api_base": "some_api_base",
"api_version": "some_api_version",
"deployment_name": "some_deployment_name",
}
def test_missing_azure_api_base() -> None:
missing_api_base_config = base_azure_model_config.copy()
del missing_api_base_config["api_base"]
with pytest.raises(ValidationError):
create_graphrag_config({
"models": {
defs.DEFAULT_CHAT_MODEL_ID: missing_api_base_config,
defs.DEFAULT_EMBEDDING_MODEL_ID: DEFAULT_EMBEDDING_MODEL_CONFIG,
}
})
def test_missing_azure_api_version() -> None:
missing_api_version_config = base_azure_model_config.copy()
del missing_api_version_config["api_version"]
with pytest.raises(ValidationError):
create_graphrag_config({
"models": {
defs.DEFAULT_CHAT_MODEL_ID: missing_api_version_config,
defs.DEFAULT_EMBEDDING_MODEL_ID: DEFAULT_EMBEDDING_MODEL_CONFIG,
}
})
def test_missing_azure_deployment_name() -> None:
missing_deployment_name_config = base_azure_model_config.copy()
del missing_deployment_name_config["deployment_name"]
with pytest.raises(ValidationError):
create_graphrag_config({
"models": {
defs.DEFAULT_CHAT_MODEL_ID: missing_deployment_name_config,
defs.DEFAULT_EMBEDDING_MODEL_ID: DEFAULT_EMBEDDING_MODEL_CONFIG,
}
})
def test_default_config() -> None:
expected = get_default_graphrag_config()
actual = create_graphrag_config({"models": DEFAULT_MODEL_CONFIG})
assert_graphrag_configs(actual, expected)
@mock.patch.dict(os.environ, {"CUSTOM_API_KEY": FAKE_API_KEY}, clear=True)
def test_load_minimal_config() -> None:
cwd = Path(__file__).parent
root_dir = (cwd / "fixtures" / "minimal_config").resolve()
expected = get_default_graphrag_config(str(root_dir))
actual = load_config(root_dir=root_dir)
assert_graphrag_configs(actual, expected)
@mock.patch.dict(os.environ, {"CUSTOM_API_KEY": FAKE_API_KEY}, clear=True)
def test_load_config_with_cli_overrides() -> None:
cwd = Path(__file__).parent
root_dir = (cwd / "fixtures" / "minimal_config").resolve()
output_dir = "some_output_dir"
expected_output_base_dir = root_dir / output_dir
expected = get_default_graphrag_config(str(root_dir))
expected.output.base_dir = str(expected_output_base_dir)
actual = load_config(
root_dir=root_dir,
cli_overrides={"output.base_dir": output_dir},
)
assert_graphrag_configs(actual, expected)
def test_load_config_missing_env_vars() -> None:
cwd = Path(__file__).parent
root_dir = (cwd / "fixtures" / "minimal_config_missing_env_var").resolve()
with pytest.raises(KeyError):
load_config(root_dir=root_dir)