amd-gaia 0.14.3__py3-none-any.whl → 0.15.1__py3-none-any.whl

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (181) hide show
  1. {amd_gaia-0.14.3.dist-info → amd_gaia-0.15.1.dist-info}/METADATA +223 -223
  2. amd_gaia-0.15.1.dist-info/RECORD +178 -0
  3. {amd_gaia-0.14.3.dist-info → amd_gaia-0.15.1.dist-info}/entry_points.txt +1 -0
  4. {amd_gaia-0.14.3.dist-info → amd_gaia-0.15.1.dist-info}/licenses/LICENSE.md +20 -20
  5. gaia/__init__.py +29 -29
  6. gaia/agents/__init__.py +19 -19
  7. gaia/agents/base/__init__.py +9 -9
  8. gaia/agents/base/agent.py +2177 -2177
  9. gaia/agents/base/api_agent.py +120 -120
  10. gaia/agents/base/console.py +1841 -1841
  11. gaia/agents/base/errors.py +237 -237
  12. gaia/agents/base/mcp_agent.py +86 -86
  13. gaia/agents/base/tools.py +83 -83
  14. gaia/agents/blender/agent.py +556 -556
  15. gaia/agents/blender/agent_simple.py +133 -135
  16. gaia/agents/blender/app.py +211 -211
  17. gaia/agents/blender/app_simple.py +41 -41
  18. gaia/agents/blender/core/__init__.py +16 -16
  19. gaia/agents/blender/core/materials.py +506 -506
  20. gaia/agents/blender/core/objects.py +316 -316
  21. gaia/agents/blender/core/rendering.py +225 -225
  22. gaia/agents/blender/core/scene.py +220 -220
  23. gaia/agents/blender/core/view.py +146 -146
  24. gaia/agents/chat/__init__.py +9 -9
  25. gaia/agents/chat/agent.py +835 -835
  26. gaia/agents/chat/app.py +1058 -1058
  27. gaia/agents/chat/session.py +508 -508
  28. gaia/agents/chat/tools/__init__.py +15 -15
  29. gaia/agents/chat/tools/file_tools.py +96 -96
  30. gaia/agents/chat/tools/rag_tools.py +1729 -1729
  31. gaia/agents/chat/tools/shell_tools.py +436 -436
  32. gaia/agents/code/__init__.py +7 -7
  33. gaia/agents/code/agent.py +549 -549
  34. gaia/agents/code/cli.py +377 -0
  35. gaia/agents/code/models.py +135 -135
  36. gaia/agents/code/orchestration/__init__.py +24 -24
  37. gaia/agents/code/orchestration/checklist_executor.py +1763 -1763
  38. gaia/agents/code/orchestration/checklist_generator.py +713 -713
  39. gaia/agents/code/orchestration/factories/__init__.py +9 -9
  40. gaia/agents/code/orchestration/factories/base.py +63 -63
  41. gaia/agents/code/orchestration/factories/nextjs_factory.py +118 -118
  42. gaia/agents/code/orchestration/factories/python_factory.py +106 -106
  43. gaia/agents/code/orchestration/orchestrator.py +841 -841
  44. gaia/agents/code/orchestration/project_analyzer.py +391 -391
  45. gaia/agents/code/orchestration/steps/__init__.py +67 -67
  46. gaia/agents/code/orchestration/steps/base.py +188 -188
  47. gaia/agents/code/orchestration/steps/error_handler.py +314 -314
  48. gaia/agents/code/orchestration/steps/nextjs.py +828 -828
  49. gaia/agents/code/orchestration/steps/python.py +307 -307
  50. gaia/agents/code/orchestration/template_catalog.py +469 -469
  51. gaia/agents/code/orchestration/workflows/__init__.py +14 -14
  52. gaia/agents/code/orchestration/workflows/base.py +80 -80
  53. gaia/agents/code/orchestration/workflows/nextjs.py +186 -186
  54. gaia/agents/code/orchestration/workflows/python.py +94 -94
  55. gaia/agents/code/prompts/__init__.py +11 -11
  56. gaia/agents/code/prompts/base_prompt.py +77 -77
  57. gaia/agents/code/prompts/code_patterns.py +2036 -2036
  58. gaia/agents/code/prompts/nextjs_prompt.py +40 -40
  59. gaia/agents/code/prompts/python_prompt.py +109 -109
  60. gaia/agents/code/schema_inference.py +365 -365
  61. gaia/agents/code/system_prompt.py +41 -41
  62. gaia/agents/code/tools/__init__.py +42 -42
  63. gaia/agents/code/tools/cli_tools.py +1138 -1138
  64. gaia/agents/code/tools/code_formatting.py +319 -319
  65. gaia/agents/code/tools/code_tools.py +769 -769
  66. gaia/agents/code/tools/error_fixing.py +1347 -1347
  67. gaia/agents/code/tools/external_tools.py +180 -180
  68. gaia/agents/code/tools/file_io.py +845 -845
  69. gaia/agents/code/tools/prisma_tools.py +190 -190
  70. gaia/agents/code/tools/project_management.py +1016 -1016
  71. gaia/agents/code/tools/testing.py +321 -321
  72. gaia/agents/code/tools/typescript_tools.py +122 -122
  73. gaia/agents/code/tools/validation_parsing.py +461 -461
  74. gaia/agents/code/tools/validation_tools.py +806 -806
  75. gaia/agents/code/tools/web_dev_tools.py +1758 -1758
  76. gaia/agents/code/validators/__init__.py +16 -16
  77. gaia/agents/code/validators/antipattern_checker.py +241 -241
  78. gaia/agents/code/validators/ast_analyzer.py +197 -197
  79. gaia/agents/code/validators/requirements_validator.py +145 -145
  80. gaia/agents/code/validators/syntax_validator.py +171 -171
  81. gaia/agents/docker/__init__.py +7 -7
  82. gaia/agents/docker/agent.py +642 -642
  83. gaia/agents/emr/__init__.py +8 -8
  84. gaia/agents/emr/agent.py +1506 -1506
  85. gaia/agents/emr/cli.py +1322 -1322
  86. gaia/agents/emr/constants.py +475 -475
  87. gaia/agents/emr/dashboard/__init__.py +4 -4
  88. gaia/agents/emr/dashboard/server.py +1974 -1974
  89. gaia/agents/jira/__init__.py +11 -11
  90. gaia/agents/jira/agent.py +894 -894
  91. gaia/agents/jira/jql_templates.py +299 -299
  92. gaia/agents/routing/__init__.py +7 -7
  93. gaia/agents/routing/agent.py +567 -570
  94. gaia/agents/routing/system_prompt.py +75 -75
  95. gaia/agents/summarize/__init__.py +11 -0
  96. gaia/agents/summarize/agent.py +885 -0
  97. gaia/agents/summarize/prompts.py +129 -0
  98. gaia/api/__init__.py +23 -23
  99. gaia/api/agent_registry.py +238 -238
  100. gaia/api/app.py +305 -305
  101. gaia/api/openai_server.py +575 -575
  102. gaia/api/schemas.py +186 -186
  103. gaia/api/sse_handler.py +373 -373
  104. gaia/apps/__init__.py +4 -4
  105. gaia/apps/llm/__init__.py +6 -6
  106. gaia/apps/llm/app.py +173 -169
  107. gaia/apps/summarize/app.py +116 -633
  108. gaia/apps/summarize/html_viewer.py +133 -133
  109. gaia/apps/summarize/pdf_formatter.py +284 -284
  110. gaia/audio/__init__.py +2 -2
  111. gaia/audio/audio_client.py +439 -439
  112. gaia/audio/audio_recorder.py +269 -269
  113. gaia/audio/kokoro_tts.py +599 -599
  114. gaia/audio/whisper_asr.py +432 -432
  115. gaia/chat/__init__.py +16 -16
  116. gaia/chat/app.py +430 -430
  117. gaia/chat/prompts.py +522 -522
  118. gaia/chat/sdk.py +1228 -1225
  119. gaia/cli.py +5481 -5621
  120. gaia/database/__init__.py +10 -10
  121. gaia/database/agent.py +176 -176
  122. gaia/database/mixin.py +290 -290
  123. gaia/database/testing.py +64 -64
  124. gaia/eval/batch_experiment.py +2332 -2332
  125. gaia/eval/claude.py +542 -542
  126. gaia/eval/config.py +37 -37
  127. gaia/eval/email_generator.py +512 -512
  128. gaia/eval/eval.py +3179 -3179
  129. gaia/eval/groundtruth.py +1130 -1130
  130. gaia/eval/transcript_generator.py +582 -582
  131. gaia/eval/webapp/README.md +167 -167
  132. gaia/eval/webapp/package-lock.json +875 -875
  133. gaia/eval/webapp/package.json +20 -20
  134. gaia/eval/webapp/public/app.js +3402 -3402
  135. gaia/eval/webapp/public/index.html +87 -87
  136. gaia/eval/webapp/public/styles.css +3661 -3661
  137. gaia/eval/webapp/server.js +415 -415
  138. gaia/eval/webapp/test-setup.js +72 -72
  139. gaia/llm/__init__.py +9 -2
  140. gaia/llm/base_client.py +60 -0
  141. gaia/llm/exceptions.py +12 -0
  142. gaia/llm/factory.py +70 -0
  143. gaia/llm/lemonade_client.py +3236 -3221
  144. gaia/llm/lemonade_manager.py +294 -294
  145. gaia/llm/providers/__init__.py +9 -0
  146. gaia/llm/providers/claude.py +108 -0
  147. gaia/llm/providers/lemonade.py +120 -0
  148. gaia/llm/providers/openai_provider.py +79 -0
  149. gaia/llm/vlm_client.py +382 -382
  150. gaia/logger.py +189 -189
  151. gaia/mcp/agent_mcp_server.py +245 -245
  152. gaia/mcp/blender_mcp_client.py +138 -138
  153. gaia/mcp/blender_mcp_server.py +648 -648
  154. gaia/mcp/context7_cache.py +332 -332
  155. gaia/mcp/external_services.py +518 -518
  156. gaia/mcp/mcp_bridge.py +811 -550
  157. gaia/mcp/servers/__init__.py +6 -6
  158. gaia/mcp/servers/docker_mcp.py +83 -83
  159. gaia/perf_analysis.py +361 -0
  160. gaia/rag/__init__.py +10 -10
  161. gaia/rag/app.py +293 -293
  162. gaia/rag/demo.py +304 -304
  163. gaia/rag/pdf_utils.py +235 -235
  164. gaia/rag/sdk.py +2194 -2194
  165. gaia/security.py +163 -163
  166. gaia/talk/app.py +289 -289
  167. gaia/talk/sdk.py +538 -538
  168. gaia/testing/__init__.py +87 -87
  169. gaia/testing/assertions.py +330 -330
  170. gaia/testing/fixtures.py +333 -333
  171. gaia/testing/mocks.py +493 -493
  172. gaia/util.py +46 -46
  173. gaia/utils/__init__.py +33 -33
  174. gaia/utils/file_watcher.py +675 -675
  175. gaia/utils/parsing.py +223 -223
  176. gaia/version.py +100 -100
  177. amd_gaia-0.14.3.dist-info/RECORD +0 -168
  178. gaia/agents/code/app.py +0 -266
  179. gaia/llm/llm_client.py +0 -729
  180. {amd_gaia-0.14.3.dist-info → amd_gaia-0.15.1.dist-info}/WHEEL +0 -0
  181. {amd_gaia-0.14.3.dist-info → amd_gaia-0.15.1.dist-info}/top_level.txt +0 -0
gaia/testing/fixtures.py CHANGED
@@ -1,333 +1,333 @@
1
- # Copyright(C) 2025-2026 Advanced Micro Devices, Inc. All rights reserved.
2
- # SPDX-License-Identifier: MIT
3
-
4
- """Test fixtures and factories for GAIA agent testing."""
5
-
6
- import tempfile
7
- from contextlib import contextmanager
8
- from pathlib import Path
9
- from typing import List, Optional, Type, TypeVar
10
-
11
- from gaia.testing.mocks import MockLLMProvider, MockVLMClient
12
-
13
- # Type variable for agent classes
14
- AgentT = TypeVar("AgentT")
15
-
16
-
17
- @contextmanager
18
- def temp_directory():
19
- """
20
- Create a temporary directory for testing with automatic cleanup.
21
-
22
- Yields a Path object to the temporary directory. The directory
23
- and all its contents are automatically deleted when the context exits.
24
-
25
- Example:
26
- from gaia.testing import temp_directory
27
-
28
- def test_file_processing():
29
- with temp_directory() as tmp_dir:
30
- # Create test files
31
- test_file = tmp_dir / "test.txt"
32
- test_file.write_text("Hello World")
33
-
34
- # Test your agent
35
- agent = MyAgent(data_dir=str(tmp_dir))
36
- result = agent.process_file(str(test_file))
37
-
38
- assert result["status"] == "success"
39
-
40
- # Directory automatically deleted after context exits
41
-
42
- Yields:
43
- Path: Path to the temporary directory
44
- """
45
- with tempfile.TemporaryDirectory() as tmp_dir:
46
- yield Path(tmp_dir)
47
-
48
-
49
- @contextmanager
50
- def temp_file(
51
- content: str = "",
52
- suffix: str = ".txt",
53
- prefix: str = "test_",
54
- ):
55
- """
56
- Create a temporary file for testing with automatic cleanup.
57
-
58
- Args:
59
- content: Initial content to write to the file
60
- suffix: File extension (default: .txt)
61
- prefix: Filename prefix (default: test_)
62
-
63
- Yields:
64
- Path: Path to the temporary file
65
-
66
- Example:
67
- from gaia.testing import temp_file
68
-
69
- def test_file_reading():
70
- with temp_file(content="test data", suffix=".json") as tmp_path:
71
- result = my_agent.read_file(str(tmp_path))
72
- assert "test data" in result
73
- """
74
- with tempfile.NamedTemporaryFile(
75
- mode="w",
76
- suffix=suffix,
77
- prefix=prefix,
78
- delete=False,
79
- ) as f:
80
- f.write(content)
81
- tmp_path = Path(f.name)
82
-
83
- try:
84
- yield tmp_path
85
- finally:
86
- tmp_path.unlink(missing_ok=True)
87
-
88
-
89
- def create_test_agent(
90
- agent_class: Type[AgentT],
91
- mock_responses: Optional[List[str]] = None,
92
- mock_vlm_text: Optional[str] = None,
93
- inject_mocks: bool = True,
94
- **agent_kwargs,
95
- ) -> AgentT:
96
- """
97
- Create an agent instance configured for testing.
98
-
99
- Creates an agent with silent mode enabled and optionally injects
100
- mock LLM and VLM providers for testing without real API calls.
101
-
102
- Args:
103
- agent_class: The Agent class to instantiate
104
- mock_responses: List of LLM responses to return in sequence
105
- mock_vlm_text: Text for mock VLM to return
106
- inject_mocks: Whether to inject mock providers (default: True)
107
- **agent_kwargs: Additional arguments passed to agent constructor
108
-
109
- Returns:
110
- Configured agent instance
111
-
112
- Example:
113
- from gaia.testing import create_test_agent
114
- from my_agent import CustomerAgent
115
-
116
- def test_customer_search():
117
- agent = create_test_agent(
118
- CustomerAgent,
119
- mock_responses=[
120
- '{"tool": "search", "args": {"name": "John"}}',
121
- "Found 1 customer named John."
122
- ],
123
- db_url="sqlite:///:memory:",
124
- )
125
-
126
- result = agent.process_query("Find customer John")
127
- assert "John" in result["answer"]
128
-
129
- Note:
130
- The agent is created with `skip_lemonade=True` and `silent_mode=True`
131
- to avoid requiring a running Lemonade server during tests.
132
- """
133
- # Ensure testing-friendly defaults
134
- agent_kwargs.setdefault("skip_lemonade", True)
135
- agent_kwargs.setdefault("silent_mode", True)
136
-
137
- # Create the agent
138
- agent = agent_class(**agent_kwargs)
139
-
140
- # Inject mocks if requested
141
- if inject_mocks:
142
- # Create and attach mock LLM
143
- mock_llm = MockLLMProvider(responses=mock_responses)
144
- agent._mock_llm = mock_llm # pylint: disable=protected-access
145
-
146
- # Inject into chat attribute if it exists
147
- if hasattr(agent, "chat"):
148
- agent.chat = mock_llm
149
-
150
- # Create and attach mock VLM if text provided
151
- if mock_vlm_text is not None:
152
- mock_vlm = MockVLMClient(extracted_text=mock_vlm_text)
153
- agent._mock_vlm = mock_vlm # pylint: disable=protected-access
154
-
155
- if hasattr(agent, "vlm"):
156
- agent.vlm = mock_vlm
157
-
158
- return agent
159
-
160
-
161
- class AgentTestContext:
162
- """
163
- Context manager for comprehensive agent testing.
164
-
165
- Provides a clean testing environment with temporary directories,
166
- mock providers, and automatic cleanup.
167
-
168
- Example:
169
- from gaia.testing import AgentTestContext
170
- from my_agent import IntakeAgent
171
-
172
- def test_intake_processing():
173
- with AgentTestContext(IntakeAgent) as ctx:
174
- # Create test file
175
- test_file = ctx.create_file("form.txt", "Patient: John Doe")
176
-
177
- # Process with agent
178
- result = ctx.agent.process_query(f"Process {test_file}")
179
-
180
- # Verify
181
- assert ctx.mock_llm.was_called
182
- assert "John" in result["answer"]
183
- """
184
-
185
- def __init__(
186
- self,
187
- agent_class: Type[AgentT],
188
- mock_responses: Optional[List[str]] = None,
189
- mock_vlm_text: Optional[str] = None,
190
- **agent_kwargs,
191
- ):
192
- """
193
- Initialize test context.
194
-
195
- Args:
196
- agent_class: Agent class to instantiate
197
- mock_responses: LLM responses to use
198
- mock_vlm_text: VLM extraction text to use
199
- **agent_kwargs: Additional agent arguments
200
- """
201
- self.agent_class = agent_class
202
- self.mock_responses = mock_responses
203
- self.mock_vlm_text = mock_vlm_text
204
- self.agent_kwargs = agent_kwargs
205
-
206
- self.agent: Optional[AgentT] = None
207
- self.mock_llm: Optional[MockLLMProvider] = None
208
- self.mock_vlm: Optional[MockVLMClient] = None
209
- self._temp_dir: Optional[Path] = None
210
- self._temp_dir_context = None
211
-
212
- def __enter__(self) -> "AgentTestContext":
213
- """Enter context and set up test environment."""
214
- # Create temporary directory
215
- self._temp_dir_context = tempfile.TemporaryDirectory()
216
- self._temp_dir = Path(self._temp_dir_context.name)
217
-
218
- # Create mock providers
219
- self.mock_llm = MockLLMProvider(responses=self.mock_responses)
220
- if self.mock_vlm_text is not None:
221
- self.mock_vlm = MockVLMClient(extracted_text=self.mock_vlm_text)
222
-
223
- # Create agent with mocks
224
- self.agent_kwargs.setdefault("skip_lemonade", True)
225
- self.agent_kwargs.setdefault("silent_mode", True)
226
-
227
- self.agent = self.agent_class(**self.agent_kwargs)
228
-
229
- # Inject mocks
230
- self.agent._mock_llm = self.mock_llm # pylint: disable=protected-access
231
- if hasattr(self.agent, "chat"):
232
- self.agent.chat = self.mock_llm
233
-
234
- if self.mock_vlm is not None:
235
- self.agent._mock_vlm = self.mock_vlm # pylint: disable=protected-access
236
- if hasattr(self.agent, "vlm"):
237
- self.agent.vlm = self.mock_vlm
238
-
239
- return self
240
-
241
- def __exit__(self, exc_type, exc_val, exc_tb):
242
- """Exit context and clean up."""
243
- # Clean up agent if it has cleanup method
244
- if self.agent is not None:
245
- if hasattr(self.agent, "close"):
246
- try:
247
- self.agent.close()
248
- except Exception:
249
- pass
250
- if hasattr(self.agent, "close_db"):
251
- try:
252
- self.agent.close_db()
253
- except Exception:
254
- pass
255
- if hasattr(self.agent, "stop_all_watchers"):
256
- try:
257
- self.agent.stop_all_watchers()
258
- except Exception:
259
- pass
260
-
261
- # Clean up temp directory
262
- if self._temp_dir_context is not None:
263
- self._temp_dir_context.cleanup()
264
-
265
- return False # Don't suppress exceptions
266
-
267
- @property
268
- def temp_dir(self) -> Path:
269
- """Get the temporary directory path."""
270
- if self._temp_dir is None:
271
- raise RuntimeError("Context not entered. Use 'with' statement.")
272
- return self._temp_dir
273
-
274
- def create_file(
275
- self,
276
- name: str,
277
- content: str = "",
278
- subdir: Optional[str] = None,
279
- ) -> Path:
280
- """
281
- Create a file in the temporary directory.
282
-
283
- Args:
284
- name: Filename
285
- content: File content
286
- subdir: Optional subdirectory to create file in
287
-
288
- Returns:
289
- Path to created file
290
- """
291
- if subdir:
292
- file_dir = self.temp_dir / subdir
293
- file_dir.mkdir(parents=True, exist_ok=True)
294
- else:
295
- file_dir = self.temp_dir
296
-
297
- file_path = file_dir / name
298
- file_path.write_text(content)
299
- return file_path
300
-
301
- def create_directory(self, name: str) -> Path:
302
- """
303
- Create a subdirectory in the temporary directory.
304
-
305
- Args:
306
- name: Directory name
307
-
308
- Returns:
309
- Path to created directory
310
- """
311
- dir_path = self.temp_dir / name
312
- dir_path.mkdir(parents=True, exist_ok=True)
313
- return dir_path
314
-
315
- def set_llm_responses(self, responses: List[str]) -> None:
316
- """
317
- Set new LLM responses.
318
-
319
- Args:
320
- responses: New list of responses
321
- """
322
- if self.mock_llm:
323
- self.mock_llm.set_responses(responses)
324
-
325
- def set_vlm_text(self, text: str) -> None:
326
- """
327
- Set new VLM extraction text.
328
-
329
- Args:
330
- text: New extraction text
331
- """
332
- if self.mock_vlm:
333
- self.mock_vlm.set_extracted_text(text)
1
+ # Copyright(C) 2025-2026 Advanced Micro Devices, Inc. All rights reserved.
2
+ # SPDX-License-Identifier: MIT
3
+
4
+ """Test fixtures and factories for GAIA agent testing."""
5
+
6
+ import tempfile
7
+ from contextlib import contextmanager
8
+ from pathlib import Path
9
+ from typing import List, Optional, Type, TypeVar
10
+
11
+ from gaia.testing.mocks import MockLLMProvider, MockVLMClient
12
+
13
+ # Type variable for agent classes
14
+ AgentT = TypeVar("AgentT")
15
+
16
+
17
+ @contextmanager
18
+ def temp_directory():
19
+ """
20
+ Create a temporary directory for testing with automatic cleanup.
21
+
22
+ Yields a Path object to the temporary directory. The directory
23
+ and all its contents are automatically deleted when the context exits.
24
+
25
+ Example:
26
+ from gaia.testing import temp_directory
27
+
28
+ def test_file_processing():
29
+ with temp_directory() as tmp_dir:
30
+ # Create test files
31
+ test_file = tmp_dir / "test.txt"
32
+ test_file.write_text("Hello World")
33
+
34
+ # Test your agent
35
+ agent = MyAgent(data_dir=str(tmp_dir))
36
+ result = agent.process_file(str(test_file))
37
+
38
+ assert result["status"] == "success"
39
+
40
+ # Directory automatically deleted after context exits
41
+
42
+ Yields:
43
+ Path: Path to the temporary directory
44
+ """
45
+ with tempfile.TemporaryDirectory() as tmp_dir:
46
+ yield Path(tmp_dir)
47
+
48
+
49
+ @contextmanager
50
+ def temp_file(
51
+ content: str = "",
52
+ suffix: str = ".txt",
53
+ prefix: str = "test_",
54
+ ):
55
+ """
56
+ Create a temporary file for testing with automatic cleanup.
57
+
58
+ Args:
59
+ content: Initial content to write to the file
60
+ suffix: File extension (default: .txt)
61
+ prefix: Filename prefix (default: test_)
62
+
63
+ Yields:
64
+ Path: Path to the temporary file
65
+
66
+ Example:
67
+ from gaia.testing import temp_file
68
+
69
+ def test_file_reading():
70
+ with temp_file(content="test data", suffix=".json") as tmp_path:
71
+ result = my_agent.read_file(str(tmp_path))
72
+ assert "test data" in result
73
+ """
74
+ with tempfile.NamedTemporaryFile(
75
+ mode="w",
76
+ suffix=suffix,
77
+ prefix=prefix,
78
+ delete=False,
79
+ ) as f:
80
+ f.write(content)
81
+ tmp_path = Path(f.name)
82
+
83
+ try:
84
+ yield tmp_path
85
+ finally:
86
+ tmp_path.unlink(missing_ok=True)
87
+
88
+
89
+ def create_test_agent(
90
+ agent_class: Type[AgentT],
91
+ mock_responses: Optional[List[str]] = None,
92
+ mock_vlm_text: Optional[str] = None,
93
+ inject_mocks: bool = True,
94
+ **agent_kwargs,
95
+ ) -> AgentT:
96
+ """
97
+ Create an agent instance configured for testing.
98
+
99
+ Creates an agent with silent mode enabled and optionally injects
100
+ mock LLM and VLM providers for testing without real API calls.
101
+
102
+ Args:
103
+ agent_class: The Agent class to instantiate
104
+ mock_responses: List of LLM responses to return in sequence
105
+ mock_vlm_text: Text for mock VLM to return
106
+ inject_mocks: Whether to inject mock providers (default: True)
107
+ **agent_kwargs: Additional arguments passed to agent constructor
108
+
109
+ Returns:
110
+ Configured agent instance
111
+
112
+ Example:
113
+ from gaia.testing import create_test_agent
114
+ from my_agent import CustomerAgent
115
+
116
+ def test_customer_search():
117
+ agent = create_test_agent(
118
+ CustomerAgent,
119
+ mock_responses=[
120
+ '{"tool": "search", "args": {"name": "John"}}',
121
+ "Found 1 customer named John."
122
+ ],
123
+ db_url="sqlite:///:memory:",
124
+ )
125
+
126
+ result = agent.process_query("Find customer John")
127
+ assert "John" in result["answer"]
128
+
129
+ Note:
130
+ The agent is created with `skip_lemonade=True` and `silent_mode=True`
131
+ to avoid requiring a running Lemonade server during tests.
132
+ """
133
+ # Ensure testing-friendly defaults
134
+ agent_kwargs.setdefault("skip_lemonade", True)
135
+ agent_kwargs.setdefault("silent_mode", True)
136
+
137
+ # Create the agent
138
+ agent = agent_class(**agent_kwargs)
139
+
140
+ # Inject mocks if requested
141
+ if inject_mocks:
142
+ # Create and attach mock LLM
143
+ mock_llm = MockLLMProvider(responses=mock_responses)
144
+ agent._mock_llm = mock_llm # pylint: disable=protected-access
145
+
146
+ # Inject into chat attribute if it exists
147
+ if hasattr(agent, "chat"):
148
+ agent.chat = mock_llm
149
+
150
+ # Create and attach mock VLM if text provided
151
+ if mock_vlm_text is not None:
152
+ mock_vlm = MockVLMClient(extracted_text=mock_vlm_text)
153
+ agent._mock_vlm = mock_vlm # pylint: disable=protected-access
154
+
155
+ if hasattr(agent, "vlm"):
156
+ agent.vlm = mock_vlm
157
+
158
+ return agent
159
+
160
+
161
+ class AgentTestContext:
162
+ """
163
+ Context manager for comprehensive agent testing.
164
+
165
+ Provides a clean testing environment with temporary directories,
166
+ mock providers, and automatic cleanup.
167
+
168
+ Example:
169
+ from gaia.testing import AgentTestContext
170
+ from my_agent import IntakeAgent
171
+
172
+ def test_intake_processing():
173
+ with AgentTestContext(IntakeAgent) as ctx:
174
+ # Create test file
175
+ test_file = ctx.create_file("form.txt", "Patient: John Doe")
176
+
177
+ # Process with agent
178
+ result = ctx.agent.process_query(f"Process {test_file}")
179
+
180
+ # Verify
181
+ assert ctx.mock_llm.was_called
182
+ assert "John" in result["answer"]
183
+ """
184
+
185
+ def __init__(
186
+ self,
187
+ agent_class: Type[AgentT],
188
+ mock_responses: Optional[List[str]] = None,
189
+ mock_vlm_text: Optional[str] = None,
190
+ **agent_kwargs,
191
+ ):
192
+ """
193
+ Initialize test context.
194
+
195
+ Args:
196
+ agent_class: Agent class to instantiate
197
+ mock_responses: LLM responses to use
198
+ mock_vlm_text: VLM extraction text to use
199
+ **agent_kwargs: Additional agent arguments
200
+ """
201
+ self.agent_class = agent_class
202
+ self.mock_responses = mock_responses
203
+ self.mock_vlm_text = mock_vlm_text
204
+ self.agent_kwargs = agent_kwargs
205
+
206
+ self.agent: Optional[AgentT] = None
207
+ self.mock_llm: Optional[MockLLMProvider] = None
208
+ self.mock_vlm: Optional[MockVLMClient] = None
209
+ self._temp_dir: Optional[Path] = None
210
+ self._temp_dir_context = None
211
+
212
+ def __enter__(self) -> "AgentTestContext":
213
+ """Enter context and set up test environment."""
214
+ # Create temporary directory
215
+ self._temp_dir_context = tempfile.TemporaryDirectory()
216
+ self._temp_dir = Path(self._temp_dir_context.name)
217
+
218
+ # Create mock providers
219
+ self.mock_llm = MockLLMProvider(responses=self.mock_responses)
220
+ if self.mock_vlm_text is not None:
221
+ self.mock_vlm = MockVLMClient(extracted_text=self.mock_vlm_text)
222
+
223
+ # Create agent with mocks
224
+ self.agent_kwargs.setdefault("skip_lemonade", True)
225
+ self.agent_kwargs.setdefault("silent_mode", True)
226
+
227
+ self.agent = self.agent_class(**self.agent_kwargs)
228
+
229
+ # Inject mocks
230
+ self.agent._mock_llm = self.mock_llm # pylint: disable=protected-access
231
+ if hasattr(self.agent, "chat"):
232
+ self.agent.chat = self.mock_llm
233
+
234
+ if self.mock_vlm is not None:
235
+ self.agent._mock_vlm = self.mock_vlm # pylint: disable=protected-access
236
+ if hasattr(self.agent, "vlm"):
237
+ self.agent.vlm = self.mock_vlm
238
+
239
+ return self
240
+
241
+ def __exit__(self, exc_type, exc_val, exc_tb):
242
+ """Exit context and clean up."""
243
+ # Clean up agent if it has cleanup method
244
+ if self.agent is not None:
245
+ if hasattr(self.agent, "close"):
246
+ try:
247
+ self.agent.close()
248
+ except Exception:
249
+ pass
250
+ if hasattr(self.agent, "close_db"):
251
+ try:
252
+ self.agent.close_db()
253
+ except Exception:
254
+ pass
255
+ if hasattr(self.agent, "stop_all_watchers"):
256
+ try:
257
+ self.agent.stop_all_watchers()
258
+ except Exception:
259
+ pass
260
+
261
+ # Clean up temp directory
262
+ if self._temp_dir_context is not None:
263
+ self._temp_dir_context.cleanup()
264
+
265
+ return False # Don't suppress exceptions
266
+
267
+ @property
268
+ def temp_dir(self) -> Path:
269
+ """Get the temporary directory path."""
270
+ if self._temp_dir is None:
271
+ raise RuntimeError("Context not entered. Use 'with' statement.")
272
+ return self._temp_dir
273
+
274
+ def create_file(
275
+ self,
276
+ name: str,
277
+ content: str = "",
278
+ subdir: Optional[str] = None,
279
+ ) -> Path:
280
+ """
281
+ Create a file in the temporary directory.
282
+
283
+ Args:
284
+ name: Filename
285
+ content: File content
286
+ subdir: Optional subdirectory to create file in
287
+
288
+ Returns:
289
+ Path to created file
290
+ """
291
+ if subdir:
292
+ file_dir = self.temp_dir / subdir
293
+ file_dir.mkdir(parents=True, exist_ok=True)
294
+ else:
295
+ file_dir = self.temp_dir
296
+
297
+ file_path = file_dir / name
298
+ file_path.write_text(content)
299
+ return file_path
300
+
301
+ def create_directory(self, name: str) -> Path:
302
+ """
303
+ Create a subdirectory in the temporary directory.
304
+
305
+ Args:
306
+ name: Directory name
307
+
308
+ Returns:
309
+ Path to created directory
310
+ """
311
+ dir_path = self.temp_dir / name
312
+ dir_path.mkdir(parents=True, exist_ok=True)
313
+ return dir_path
314
+
315
+ def set_llm_responses(self, responses: List[str]) -> None:
316
+ """
317
+ Set new LLM responses.
318
+
319
+ Args:
320
+ responses: New list of responses
321
+ """
322
+ if self.mock_llm:
323
+ self.mock_llm.set_responses(responses)
324
+
325
+ def set_vlm_text(self, text: str) -> None:
326
+ """
327
+ Set new VLM extraction text.
328
+
329
+ Args:
330
+ text: New extraction text
331
+ """
332
+ if self.mock_vlm:
333
+ self.mock_vlm.set_extracted_text(text)