1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472
|
import http
import math
import os
import pathlib
import json, html, re
from typing import Any, Iterator, MutableMapping, Optional
from unittest.mock import MagicMock, patch
import pytest
from azure.ai.evaluation._common.constants import EvaluationMetrics, HarmSeverityLevel, RAIService
from azure.ai.evaluation._common.rai_service import (
_get_service_discovery_url,
ensure_service_availability,
evaluate_with_rai_service,
fetch_or_reuse_token,
fetch_result,
get_rai_svc_url,
parse_response,
submit_request,
Tasks,
USER_TEXT_TEMPLATE_DICT,
get_formatted_template,
)
from azure.core.exceptions import HttpResponseError
from azure.core.rest import AsyncHttpResponse, HttpRequest
from azure.identity import DefaultAzureCredential
@pytest.fixture
def data_file():
data_path = os.path.join(pathlib.Path(__file__).parent.resolve(), "data")
return os.path.join(data_path, "evaluate_test_data.jsonl")
class MockAsyncHttpResponse(AsyncHttpResponse):
"""A mocked implementation of azure.core.rest.HttpResponse."""
def __init__(
self,
status_code: int,
*,
text: Optional[str] = None,
json: Optional[Any] = None,
headers: Optional[MutableMapping[str, str]] = None,
request: Optional[HttpRequest] = None,
content_type: Optional[str] = None,
) -> None:
self._status_code = status_code
self._text = text or ""
self._json = json
self._request = request
self._headers = headers or {}
self._content_type = content_type
def json(self) -> Any:
return self._json
def text(self, encoding: Optional[str] = None) -> str:
return self._text
@property
def status_code(self) -> int:
return self._status_code
@property
def request(self) -> HttpRequest:
return self._request
@property
def reason(self) -> str:
return f"{self.status_code} {http.client.responses[self.status_code]}"
@property
def headers(self) -> MutableMapping[str, str]:
return self._headers
@property
def content_type(self) -> Optional[str]:
return self._content_type
@property
def is_closed(self) -> bool:
return True
@property
def is_stream_consumed(self) -> bool:
return True
@property
def encoding(self) -> Optional[str]:
return None
def raise_for_status(self) -> None:
if self.status_code >= 400:
raise HttpResponseError(response=self)
async def close(self) -> None:
pass
async def __aenter__(self) -> object:
raise NotImplementedError()
async def __aexit__(self, *args) -> None:
raise NotImplementedError()
@property
def url(self) -> str:
raise NotImplementedError()
@property
def content(self) -> bytes:
raise NotImplementedError()
async def read(self) -> bytes:
raise NotImplementedError()
async def iter_bytes(self, **kwargs) -> Iterator[bytes]:
raise NotImplementedError()
async def iter_raw(self, **kwargs) -> Iterator[bytes]:
raise NotImplementedError()
@pytest.mark.usefixtures("mock_project_scope")
@pytest.mark.unittest
class TestContentSafetyEvaluator:
def test_rai_subscript_functions(self):
# ensure_service_availability()
"""
evaluate_with_rai_service()
fetch_or_reuse_token()
get_rai_svc_url()
_get_service_discovery_url()
parse_response()
fetch_result()
submit_request()
ensure_service_availability()"""
@pytest.mark.asyncio
@patch("azure.ai.evaluation._http_utils.AsyncHttpPipeline.get", return_value=MockAsyncHttpResponse(200, json={}))
async def test_ensure_service_availability(self, client_mock):
_ = await ensure_service_availability("dummy_url", "dummy_token")
assert client_mock._mock_await_count == 1
@pytest.mark.asyncio
@patch("azure.ai.evaluation._http_utils.AsyncHttpPipeline.get", return_value=MockAsyncHttpResponse(9001, json={}))
async def test_ensure_service_availability_service_unavailable(self, client_mock):
with pytest.raises(Exception) as exc_info:
_ = await ensure_service_availability("dummy_url", "dummy_token")
assert "RAI service is unavailable in this region" in str(exc_info._excinfo[1])
assert "Status Code: 9001" in str(exc_info._excinfo[1])
assert client_mock._mock_await_count == 1
@pytest.mark.asyncio
@patch("azure.ai.evaluation._http_utils.AsyncHttpPipeline.get", return_value=MockAsyncHttpResponse(200, json={}))
async def test_ensure_service_availability_exception_capability_unavailable(self, client_mock):
with pytest.raises(Exception) as exc_info:
_ = await ensure_service_availability("dummy_url", "dummy_token", capability="does not exist")
assert "The needed capability 'does not exist' is not supported by the RAI service in this region" in str(
exc_info._excinfo[1]
)
assert client_mock._mock_await_count == 1
@pytest.mark.asyncio
@patch(
"azure.ai.evaluation._http_utils.AsyncHttpPipeline.post",
return_value=MockAsyncHttpResponse(
202,
json={"location": "this/is/the/dummy-operation-id"},
),
)
async def test_submit_request(self, client_mock):
result = await submit_request(
data={"query": "What is the meaning of life", "response": "42"},
metric="points",
rai_svc_url="www.notarealurl.com",
token="dummy",
annotation_task=Tasks.CONTENT_HARM,
evaluator_name="dummy-evaluator",
)
assert result == "dummy-operation-id"
@pytest.mark.asyncio
@patch(
"azure.ai.evaluation._http_utils.AsyncHttpPipeline.post",
return_value=MockAsyncHttpResponse(
404,
json={"location": "this/is/the/dummy-operation-id"},
content_type="application/json",
),
)
async def test_submit_request_not_found(self, client_mock):
with pytest.raises(HttpResponseError) as exc_info:
_ = await submit_request(
data={"query": "What is the meaning of life", "response": "42"},
metric="points",
rai_svc_url="www.notarealurl.com",
token="dummy",
annotation_task=Tasks.CONTENT_HARM,
evaluator_name="dummy-evaluator",
)
assert "Operation returned an invalid status '404 Not Found'" in str(exc_info._excinfo[1])
@pytest.mark.usefixtures("mock_token")
@pytest.mark.usefixtures("mock_expired_token")
@pytest.mark.asyncio
async def test_fetch_or_reuse_token(self, mock_token, mock_expired_token):
mock_cred = MagicMock(Spec=DefaultAzureCredential)
mock_cred.get_token.return_value = type("", (object,), {"token": 100})()
res = await fetch_or_reuse_token(credential=mock_cred, token=mock_token)
assert res == mock_token
res = await fetch_or_reuse_token(credential=mock_cred, token=mock_expired_token)
assert res == 100
res = await fetch_or_reuse_token(credential=mock_cred, token="not-a-token")
assert res == 100
@patch(
"azure.ai.evaluation._http_utils.AsyncHttpPipeline.get",
return_value=MockAsyncHttpResponse(200, json={"result": "stuff"}),
)
@patch("azure.ai.evaluation._common.constants.RAIService.TIMEOUT", 1)
@patch("azure.ai.evaluation._common.constants.RAIService.SLEEP_TIME", 1.2)
@pytest.mark.usefixtures("mock_token")
@pytest.mark.asyncio
async def test_fetch_result(self, client_mock, mock_token):
# These asserts aren't necessary, but given the scarcity of constant patches,
# it's nice to include an obvious example.
assert RAIService.TIMEOUT == 1
assert RAIService.SLEEP_TIME == 1.2
res = await fetch_result(
operation_id="op-id", rai_svc_url="www.notarealurl.com", credential=None, token=mock_token
)
assert client_mock._mock_await_count == 1
assert res["result"] == "stuff"
@patch(
"azure.ai.evaluation._http_utils.AsyncHttpPipeline.get",
return_value=MockAsyncHttpResponse(404, json={"result": "stuff"}),
)
@patch("azure.ai.evaluation._common.constants.RAIService.TIMEOUT", 1)
@patch("azure.ai.evaluation._common.constants.RAIService.SLEEP_TIME", 1.2)
@pytest.mark.usefixtures("mock_token")
@pytest.mark.asyncio
async def test_fetch_result_timeout(self, client_mock, mock_token):
with pytest.raises(TimeoutError) as exc_info:
_ = await fetch_result(
operation_id="op-id", rai_svc_url="www.notarealurl.com", credential=None, token=mock_token
)
# We expect 2 calls; the initial call, then one more ~2 seconds later.
assert client_mock._mock_await_count == 2
# Don't bother checking exact time beyond seconds, that's never going to be consistent across machines.
assert "Fetching annotation result 2 times out after 1" in str(exc_info._excinfo[1])
def test_parse_response(self):
batch_response = [{"not-a-metric": "not-a-value"}]
metric_name = EvaluationMetrics.HATE_FAIRNESS
result = parse_response(
batch_response=batch_response,
metric_name=metric_name,
metric_display_name=EvaluationMetrics.HATE_UNFAIRNESS,
)
assert math.isnan(result[EvaluationMetrics.HATE_UNFAIRNESS])
assert math.isnan(result[EvaluationMetrics.HATE_UNFAIRNESS + "_score"])
assert math.isnan(result[EvaluationMetrics.HATE_UNFAIRNESS + "_reason"])
metric_name = EvaluationMetrics.VIOLENCE
response_value = {
"label": 0,
"reasoning": "This is a sample reasoning.",
"version": "0.3",
}
# The parse_response function has a TON of conditional logic that depends
# on the exact structure of batch_response[0][metric_name].
# This tests ALL of it.
batch_response[0] = {metric_name: str(response_value)}
result = parse_response(batch_response=batch_response, metric_name=metric_name, metric_display_name=metric_name)
assert result[metric_name] == HarmSeverityLevel.VeryLow.value
assert result[metric_name + "_score"] == 0
assert result[metric_name + "_reason"] == response_value["reasoning"]
response_value["output"] = {
"valid": True,
"reason": "This is a sample reason.",
}
batch_response[0] = {metric_name: str(response_value)}
result = parse_response(batch_response=batch_response, metric_name=metric_name, metric_display_name=metric_name)
assert result[metric_name] == HarmSeverityLevel.VeryLow.value
assert result[metric_name + "_score"] == 0
assert result[metric_name + "_reason"] == response_value["output"]["reason"]
response_value.pop("output")
response_value.pop("reasoning")
response_value.pop("label")
batch_response[0] = {metric_name: str(response_value)}
result = parse_response(batch_response=batch_response, metric_name=metric_name)
assert math.isnan(result[metric_name])
assert math.isnan(result[metric_name + "_score"])
assert result[metric_name + "_reason"] == ""
batch_response[0] = {metric_name: 5}
result = parse_response(batch_response=batch_response, metric_name=metric_name)
assert result[metric_name] == HarmSeverityLevel.Medium.value
assert result[metric_name + "_score"] == 5
assert result[metric_name + "_reason"] == ""
batch_response[0] = {metric_name: 8}
result = parse_response(batch_response=batch_response, metric_name=metric_name)
assert math.isnan(result[metric_name])
assert math.isnan(result[metric_name + "_score"])
batch_response[0] = {metric_name: "value is 7"}
result = parse_response(batch_response=batch_response, metric_name=metric_name)
assert result[metric_name] == HarmSeverityLevel.High.value
assert result[metric_name + "_score"] == 7
assert result[metric_name + "_reason"] == "value is 7"
batch_response[0] = {metric_name: "not a number"}
result = parse_response(batch_response=batch_response, metric_name=metric_name)
assert math.isnan(result[metric_name])
assert math.isnan(result[metric_name + "_score"])
batch_response[0] = {metric_name: ["still not a number"]}
result = parse_response(batch_response=batch_response, metric_name=metric_name, metric_display_name=metric_name)
assert math.isnan(result[metric_name])
assert math.isnan(result[metric_name + "_score"])
@pytest.mark.asyncio
@patch(
"azure.ai.evaluation._http_utils.AsyncHttpPipeline.get",
return_value=MockAsyncHttpResponse(
200, json={"properties": {"discoveryUrl": "https://www.url.com:123/thePath"}}
),
)
async def test_get_service_discovery_url(self, client_mock):
token = "fake-token"
azure_ai_project = {
"subscription_id": "fake-id",
"project_name": "fake-name",
"resource_group_name": "fake-group",
}
url = await _get_service_discovery_url(azure_ai_project=azure_ai_project, token=token)
assert url == "https://www.url.com:123"
@pytest.mark.asyncio
@patch(
"azure.ai.evaluation._http_utils.AsyncHttpPipeline.get",
return_value=MockAsyncHttpResponse(
201, json={"properties": {"discoveryUrl": "https://www.url.com:123/thePath"}}
),
)
async def test_get_service_discovery_url_exception(self, client_mock):
token = "fake-token"
azure_ai_project = {
"subscription_id": "fake-id",
"project_name": "fake-name",
"resource_group_name": "fake-group",
}
with pytest.raises(Exception) as exc_info:
_ = await _get_service_discovery_url(azure_ai_project=azure_ai_project, token=token)
assert "Failed to connect to your Azure AI project." in str(exc_info._excinfo[1])
@pytest.mark.asyncio
@patch(
"azure.ai.evaluation._http_utils.AsyncHttpPipeline.get",
return_value=MockAsyncHttpResponse(
200, json={"properties": {"discoveryUrl": "https://www.url.com:123/thePath"}}
),
)
@patch(
"azure.ai.evaluation._common.rai_service._get_service_discovery_url",
return_value="https://www.url.com:123",
)
async def test_get_rai_svc_url(self, client_mock, discovery_mock):
token = "fake-token"
project_scope = {
"subscription_id": "fake-id",
"project_name": "fake-name",
"resource_group_name": "fake-group",
}
rai_url = await get_rai_svc_url(project_scope=project_scope, token=token)
assert rai_url == (
"https://www.url.com:123/raisvc/v1.0/subscriptions/fake-id/"
+ "resourceGroups/fake-group/providers/Microsoft.MachineLearningServices/workspaces/fake-name"
)
@pytest.mark.asyncio
@patch(
"azure.ai.evaluation._common.rai_service.fetch_or_reuse_token",
return_value="dummy-token",
)
@patch(
"azure.ai.evaluation._common.rai_service.get_rai_svc_url",
return_value="www.rai_url.com",
)
@patch(
"azure.ai.evaluation._common.rai_service.ensure_service_availability",
return_value=None,
)
@patch(
"azure.ai.evaluation._common.rai_service.submit_request",
return_value="op_id",
)
@patch(
"azure.ai.evaluation._common.rai_service.fetch_result",
return_value="response_object",
)
@patch(
"azure.ai.evaluation._common.rai_service.parse_response",
return_value="wow-that's-a-lot-of-patches",
)
@patch("azure.identity.DefaultAzureCredential")
async def test_evaluate_with_rai_service(
self, cred_mock, fetch_token_mock, scv_mock, avail_mock, submit_mock, fetch_result_mock, parse_mock
):
result = await evaluate_with_rai_service(
"what is the weather outside?",
EvaluationMetrics.HATE_FAIRNESS,
{"subscription_id": "fake-id", "project_name": "fake-name", "resource_group_name": "fake-group"},
DefaultAzureCredential(),
)
assert result == "wow-that's-a-lot-of-patches"
assert fetch_token_mock._mock_call_count == 1
assert scv_mock._mock_call_count == 1
assert avail_mock._mock_call_count == 1
assert submit_mock._mock_call_count == 1
assert fetch_result_mock._mock_call_count == 1
assert parse_mock._mock_call_count == 1
# RAI service templates are so different that it's not worth trying to test them all in one test.
# Groundedness is JSON
def test_get_formatted_template_groundedness(self):
tagged_text = "This text </> has <> tags."
bracketed_text = "{This text has {brackets}, and I didn't even both to even them out {."
quoted_text = (
'This text has \'quotes\', also it has "quotes", and it even has `backticks` and """ triple quotes""".'
)
all_texts = [tagged_text, quoted_text, bracketed_text]
for text in all_texts:
input_kwargs = {
"query": text,
"response": text,
"context": text,
}
formatted_payload = get_formatted_template(input_kwargs, Tasks.GROUNDEDNESS)
assert json.loads(formatted_payload)["question"] == text
# Default is basic markup.
def test_get_formatted_template_default(self):
tagged_text = "This text </> has <> tags."
bracketed_text = "{This text has {brackets}, and I didn't even both to even them out {."
quoted_text = (
'This text has \'quotes\', also it has "quotes", and it even has `backticks` and """ triple quotes""".'
)
all_texts = [tagged_text, quoted_text, bracketed_text]
for text in all_texts:
input_kwargs = {
"query": text,
"response": text,
"context": text,
}
formatted_payload = get_formatted_template(input_kwargs, "DEFAULT")
assert html.unescape(re.match("\<Human\>{(.*?)}\<", formatted_payload)[1]) == text
|