File: test_ai_agents_instrumentor.py

package info (click to toggle)
python-azure 20251118%2Bgit-1
  • links: PTS, VCS
  • area: main
  • in suites: sid
  • size: 783,356 kB
  • sloc: python: 6,474,533; ansic: 804; javascript: 287; sh: 205; makefile: 198; xml: 109
file content (468 lines) | stat: -rw-r--r-- 19,847 bytes parent folder | download
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
# pylint: disable=too-many-lines,line-too-long,useless-suppression
# ------------------------------------
# Copyright (c) Microsoft Corporation.
# Licensed under the MIT License.
# ------------------------------------
# cSpell:disable# cSpell:disable
import pytest
import os
from typing import Optional
from azure.ai.projects.telemetry import AIProjectInstrumentor, _utils
from azure.core.settings import settings
from gen_ai_trace_verifier import GenAiTraceVerifier
from azure.ai.projects.models import PromptAgentDefinition, PromptAgentDefinitionText

from azure.ai.projects.models import (
    Reasoning,
    FunctionTool,
    ResponseTextFormatConfigurationText,
)
from devtools_testutils import (
    recorded_by_proxy,
)

from test_base import servicePreparer
from test_ai_instrumentor_base import TestAiAgentsInstrumentorBase, MessageCreationMode, CONTENT_TRACING_ENV_VARIABLE

settings.tracing_implementation = "OpenTelemetry"
_utils._span_impl_type = settings.tracing_implementation()


class TestAiAgentsInstrumentor(TestAiAgentsInstrumentorBase):
    """Tests for AI agents instrumentor."""

    @pytest.fixture(scope="function")
    def instrument_with_content(self):
        os.environ.update({CONTENT_TRACING_ENV_VARIABLE: "True"})
        self.setup_telemetry()
        yield
        self.cleanup()

    @pytest.fixture(scope="function")
    def instrument_without_content(self):
        os.environ.update({CONTENT_TRACING_ENV_VARIABLE: "False"})
        self.setup_telemetry()
        yield
        self.cleanup()

    def test_instrumentation(self, **kwargs):
        # Make sure code is not instrumented due to a previous test exception
        AIProjectInstrumentor().uninstrument()
        exception_caught = False
        try:
            assert AIProjectInstrumentor().is_instrumented() == False
            AIProjectInstrumentor().instrument()
            assert AIProjectInstrumentor().is_instrumented() == True
            AIProjectInstrumentor().uninstrument()
            assert AIProjectInstrumentor().is_instrumented() == False
        except RuntimeError as e:
            exception_caught = True
            print(e)
        assert exception_caught == False

    def test_instrumenting_twice_does_not_cause_exception(self, **kwargs):
        # Make sure code is not instrumented due to a previous test exception
        AIProjectInstrumentor().uninstrument()
        exception_caught = False
        try:
            AIProjectInstrumentor().instrument()
            AIProjectInstrumentor().instrument()
        except RuntimeError as e:
            exception_caught = True
            print(e)
        AIProjectInstrumentor().uninstrument()
        assert exception_caught == False

    def test_uninstrumenting_uninstrumented_does_not_cause_exception(self, **kwargs):
        # Make sure code is not instrumented due to a previous test exception
        AIProjectInstrumentor().uninstrument()
        exception_caught = False
        try:
            AIProjectInstrumentor().uninstrument()
        except RuntimeError as e:
            exception_caught = True
            print(e)
        assert exception_caught == False

    def test_uninstrumenting_twice_does_not_cause_exception(self, **kwargs):
        # Make sure code is not instrumented due to a previous test exception
        AIProjectInstrumentor().uninstrument()
        exception_caught = False
        try:
            AIProjectInstrumentor().instrument()
            AIProjectInstrumentor().uninstrument()
            AIProjectInstrumentor().uninstrument()
        except RuntimeError as e:
            exception_caught = True
            print(e)
        assert exception_caught == False

    @pytest.mark.parametrize(
        "env1, env2, expected",
        [
            (None, None, False),
            (None, False, False),
            (None, True, False),
            (False, None, False),
            (False, False, False),
            (False, True, False),
            (True, None, True),
            (True, False, True),
            (True, True, True),
        ],
    )
    def test_content_recording_verify_old_env_variable_ignored(
        self, env1: Optional[bool], env2: Optional[bool], expected: bool
    ):
        """
        Test content recording enablement with both old and new environment variables.
        This test verifies the behavior of content recording when both the current
        and legacy environment variables are set to different combinations of values.
        The method tests all possible combinations of None, True, and False for both
        environment variables to ensure the old one is no longer having impact, since
        support for it has been dropped.
        Args:
            env1: Value for the current content tracing environment variable.
                  Can be None (unset), True, or False.
            env2: Value for the old/legacy content tracing environment variable.
                  Can be None (unset), True, or False.
            expected: The expected result of is_content_recording_enabled() given
                      the environment variable combination.
        The test ensures that only if one or both of the environment variables are
        defined and set to "true" content recording is enabled.
        """

        OLD_CONTENT_TRACING_ENV_VARIABLE = "AZURE_TRACING_GEN_AI_CONTENT_RECORDING_ENABLED"  # Deprecated, undocumented.

        def set_env_var(var_name, value):
            if value is None:
                os.environ.pop(var_name, None)
            else:
                os.environ[var_name] = "true" if value else "false"

        set_env_var(CONTENT_TRACING_ENV_VARIABLE, env1)
        set_env_var(OLD_CONTENT_TRACING_ENV_VARIABLE, env2)

        self.setup_telemetry()
        try:
            assert AIProjectInstrumentor().is_content_recording_enabled() == expected
        finally:
            self.cleanup()  # This also undefines CONTENT_TRACING_ENV_VARIABLE
            os.environ.pop(OLD_CONTENT_TRACING_ENV_VARIABLE, None)

    @pytest.mark.usefixtures("instrument_with_content")
    @servicePreparer()
    @recorded_by_proxy
    def test_agent_creation_with_tracing_content_recording_enabled(self, **kwargs):
        self.cleanup()
        os.environ.update({CONTENT_TRACING_ENV_VARIABLE: "True"})
        self.setup_telemetry()
        assert True == AIProjectInstrumentor().is_content_recording_enabled()
        assert True == AIProjectInstrumentor().is_instrumented()

        with self.create_client(operation_group="tracing", **kwargs) as project_client:

            model = self.test_agents_params["model_deployment_name"]
            print(f"Using model deployment: {model}")

            agent_definition = PromptAgentDefinition(
                # Required parameter
                model=model,
                # Optional parameters
                instructions="You are a helpful AI assistant. Be polite and provide accurate information.",
                # temperature=0.7,
                # top_p=0.9,
                # # Reasoning configuration
                # reasoning=Reasoning(
                #     effort="medium",
                #     summary="auto",
                # ),
                # # Tools that the model can use
                # tools=[
                #     # Function tool for custom functions
                #     FunctionTool(
                #         name="get_weather",
                #         description="Get the current weather for a location",
                #         parameters={
                #             "type": "object",
                #             "properties": {
                #                 "location": {"type": "string", "description": "The city and state, e.g. San Francisco, CA"},
                #                 "unit": {
                #                     "type": "string",
                #                     "enum": ["celsius", "fahrenheit"],
                #                     "description": "The temperature unit",
                #                 },
                #             },
                #             "required": ["location"],
                #         },
                #         strict=True,  # Enforce strict parameter validation
                #     ),
                # ],
                # # Text response configuration
                # text=PromptAgentDefinitionText(format=ResponseTextFormatConfigurationText()),
            )

            agent = project_client.agents.create_version(agent_name="myagent", definition=agent_definition)
            version = agent.version

            # delete agent and close client
            project_client.agents.delete_version(agent_name=agent.name, agent_version=agent.version)
            print("Deleted agent")

        # ------------------------- Validate "create_agent" span ---------------------------------
        self.exporter.force_flush()
        spans = self.exporter.get_spans_by_name("create_agent myagent")
        assert len(spans) == 1
        span = spans[0]
        expected_attributes = [
            ("gen_ai.system", "az.ai.agents"),
            ("gen_ai.provider.name", "azure.ai.agents"),
            ("gen_ai.operation.name", "create_agent"),
            ("server.address", ""),
            ("gen_ai.request.model", model),
            # ("gen_ai.request.temperature", "0.7"),
            # ("gen_ai.request.top_p", "0.9"),
            # ("gen_ai.request.response_format", "text"),
            # ("gen_ai.request.reasoning.effort", "medium"),
            # ("gen_ai.request.reasoning.summary", "auto"),
            ("gen_ai.agent.name", "myagent"),
            ("gen_ai.agent.id", "myagent:" + str(version)),
            ("gen_ai.agent.version", str(version)),
            ("gen_ai.agent.type", "prompt"),
        ]
        attributes_match = GenAiTraceVerifier().check_span_attributes(span, expected_attributes)
        assert attributes_match == True

        expected_events = [
            {
                "name": "gen_ai.system.instruction",
                "attributes": {
                    "gen_ai.system": "az.ai.agents",
                    "gen_ai.event.content": '{"content": [{"type": "text", "text": "You are a helpful AI assistant. Be polite and provide accurate information."}]}',
                },
            }
        ]
        events_match = GenAiTraceVerifier().check_span_events(span, expected_events)
        assert events_match == True

    @pytest.mark.usefixtures("instrument_without_content")
    @servicePreparer()
    @recorded_by_proxy
    def test_agent_creation_with_tracing_content_recording_disabled(self, **kwargs):
        self.cleanup()
        os.environ.update({CONTENT_TRACING_ENV_VARIABLE: "False"})
        self.setup_telemetry()
        assert False == AIProjectInstrumentor().is_content_recording_enabled()
        assert True == AIProjectInstrumentor().is_instrumented()

        with self.create_client(operation_group="agents", **kwargs) as project_client:

            model = self.test_agents_params["model_deployment_name"]
            agent_definition = PromptAgentDefinition(
                # Required parameter
                model=model,
                # Optional parameters
                instructions="You are a helpful AI assistant. Always be polite and provide accurate information.",
                # temperature=0.7,
                # top_p=0.9,
                # # Reasoning configuration
                # reasoning=Reasoning(
                #     effort="medium",
                #     summary="auto",
                # ),
                # # Tools that the model can use
                # tools=[
                #     # Function tool for custom functions
                #     FunctionTool(
                #         name="get_weather",
                #         description="Get the current weather for a location",
                #         parameters={
                #             "type": "object",
                #             "properties": {
                #                 "location": {"type": "string", "description": "The city and state, e.g. San Francisco, CA"},
                #                 "unit": {
                #                     "type": "string",
                #                     "enum": ["celsius", "fahrenheit"],
                #                     "description": "The temperature unit",
                #                 },
                #             },
                #             "required": ["location"],
                #         },
                #         strict=True,  # Enforce strict parameter validation
                #     ),
                # ],
                # Text response configuration
                # text=PromptAgentDefinitionText(format=ResponseTextFormatConfigurationText()),
            )

            agent = project_client.agents.create_version(agent_name="myagent", definition=agent_definition)
            version = agent.version

            # delete agent and close client
            project_client.agents.delete_version(agent_name=agent.name, agent_version=agent.version)
            print("Deleted agent")

        # ------------------------- Validate "create_agent" span ---------------------------------
        self.exporter.force_flush()
        spans = self.exporter.get_spans_by_name("create_agent myagent")
        assert len(spans) == 1
        span = spans[0]
        expected_attributes = [
            ("gen_ai.system", "az.ai.agents"),
            ("gen_ai.provider.name", "azure.ai.agents"),
            ("gen_ai.operation.name", "create_agent"),
            ("server.address", ""),
            ("gen_ai.request.model", model),
            # ("gen_ai.request.temperature", "0.7"),
            # ("gen_ai.request.top_p", "0.9"),
            # ("gen_ai.request.response_format", "text"),
            # ("gen_ai.request.reasoning.effort", "medium"),
            # ("gen_ai.request.reasoning.summary", "auto"),
            ("gen_ai.agent.name", "myagent"),
            ("gen_ai.agent.id", "myagent:" + str(version)),
            ("gen_ai.agent.version", str(version)),
            ("gen_ai.agent.type", "prompt"),
        ]
        attributes_match = GenAiTraceVerifier().check_span_attributes(span, expected_attributes)
        assert attributes_match == True

        expected_events = [
            {
                "name": "gen_ai.system.instruction",
                "attributes": {
                    "gen_ai.system": "az.ai.agents",
                    "gen_ai.event.content": "{}",
                },
            }
        ]
        events_match = GenAiTraceVerifier().check_span_events(span, expected_events)
        assert events_match == True

    @pytest.mark.usefixtures("instrument_with_content")
    @servicePreparer()
    @recorded_by_proxy
    def test_workflow_agent_creation_with_tracing_content_recording_enabled(self, **kwargs):
        """Test workflow agent creation with content recording enabled."""
        self.cleanup()
        os.environ.update({CONTENT_TRACING_ENV_VARIABLE: "True"})
        self.setup_telemetry()
        assert True == AIProjectInstrumentor().is_content_recording_enabled()
        assert True == AIProjectInstrumentor().is_instrumented()

        from azure.ai.projects.models import WorkflowAgentDefinition

        with self.create_client(operation_group="tracing", **kwargs) as project_client:

            workflow_yaml = """
kind: workflow
trigger:
  kind: OnConversationStart
  id: test_workflow
  actions:
    - kind: SetVariable
      id: set_variable
      variable: Local.TestVar
      value: "test"
"""

            agent = project_client.agents.create_version(
                agent_name="test-workflow-agent",
                definition=WorkflowAgentDefinition(workflow=workflow_yaml),
            )
            version = agent.version

            # delete agent
            project_client.agents.delete_version(agent_name=agent.name, agent_version=agent.version)
            print("Deleted workflow agent")

        # ------------------------- Validate "create_agent" span ---------------------------------
        self.exporter.force_flush()
        spans = self.exporter.get_spans_by_name("create_agent test-workflow-agent")
        assert len(spans) == 1
        span = spans[0]
        expected_attributes = [
            ("gen_ai.system", "az.ai.agents"),
            ("gen_ai.provider.name", "azure.ai.agents"),
            ("gen_ai.operation.name", "create_agent"),
            ("server.address", ""),
            ("gen_ai.agent.name", "test-workflow-agent"),
            ("gen_ai.agent.id", "test-workflow-agent:" + str(version)),
            ("gen_ai.agent.version", str(version)),
            ("gen_ai.agent.type", "workflow"),
        ]
        attributes_match = GenAiTraceVerifier().check_span_attributes(span, expected_attributes)
        assert attributes_match == True

        # Verify workflow event with standard content format
        events = span.events
        assert len(events) == 1
        workflow_event = events[0]
        assert workflow_event.name == "gen_ai.agent.workflow"

        import json

        event_content = json.loads(workflow_event.attributes["gen_ai.event.content"])
        assert "content" in event_content
        assert len(event_content["content"]) == 1
        assert event_content["content"][0]["type"] == "workflow"
        assert "workflow" in event_content["content"][0]
        assert "kind: workflow" in event_content["content"][0]["workflow"]

    @pytest.mark.usefixtures("instrument_without_content")
    @servicePreparer()
    @recorded_by_proxy
    def test_workflow_agent_creation_with_tracing_content_recording_disabled(self, **kwargs):
        """Test workflow agent creation with content recording disabled."""
        self.cleanup()
        os.environ.update({CONTENT_TRACING_ENV_VARIABLE: "False"})
        self.setup_telemetry()
        assert False == AIProjectInstrumentor().is_content_recording_enabled()
        assert True == AIProjectInstrumentor().is_instrumented()

        from azure.ai.projects.models import WorkflowAgentDefinition

        with self.create_client(operation_group="agents", **kwargs) as project_client:

            workflow_yaml = """
kind: workflow
trigger:
  kind: OnConversationStart
  id: test_workflow
  actions:
    - kind: SetVariable
      id: set_variable
      variable: Local.TestVar
      value: "test"
"""

            agent = project_client.agents.create_version(
                agent_name="test-workflow-agent",
                definition=WorkflowAgentDefinition(workflow=workflow_yaml),
            )
            version = agent.version

            # delete agent
            project_client.agents.delete_version(agent_name=agent.name, agent_version=agent.version)
            print("Deleted workflow agent")

        # ------------------------- Validate "create_agent" span ---------------------------------
        self.exporter.force_flush()
        spans = self.exporter.get_spans_by_name("create_agent test-workflow-agent")
        assert len(spans) == 1
        span = spans[0]
        expected_attributes = [
            ("gen_ai.system", "az.ai.agents"),
            ("gen_ai.provider.name", "azure.ai.agents"),
            ("gen_ai.operation.name", "create_agent"),
            ("server.address", ""),
            ("gen_ai.agent.name", "test-workflow-agent"),
            ("gen_ai.agent.id", "test-workflow-agent:" + str(version)),
            ("gen_ai.agent.version", str(version)),
            ("gen_ai.agent.type", "workflow"),
        ]
        attributes_match = GenAiTraceVerifier().check_span_attributes(span, expected_attributes)
        assert attributes_match == True

        # Verify no workflow event when content recording is disabled
        events = span.events
        assert len(events) == 0