@illuma-ai/agents 1.1.20 → 1.1.22

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 (246) hide show
  1. package/dist/cjs/graphs/Graph.cjs +12 -1
  2. package/dist/cjs/graphs/Graph.cjs.map +1 -1
  3. package/dist/cjs/graphs/MultiAgentGraph.cjs +85 -1
  4. package/dist/cjs/graphs/MultiAgentGraph.cjs.map +1 -1
  5. package/dist/cjs/llm/bedrock/index.cjs +14 -0
  6. package/dist/cjs/llm/bedrock/index.cjs.map +1 -1
  7. package/dist/cjs/run.cjs +20 -9
  8. package/dist/cjs/run.cjs.map +1 -1
  9. package/dist/esm/graphs/Graph.mjs +12 -1
  10. package/dist/esm/graphs/Graph.mjs.map +1 -1
  11. package/dist/esm/graphs/MultiAgentGraph.mjs +85 -1
  12. package/dist/esm/graphs/MultiAgentGraph.mjs.map +1 -1
  13. package/dist/esm/llm/bedrock/index.mjs +14 -0
  14. package/dist/esm/llm/bedrock/index.mjs.map +1 -1
  15. package/dist/esm/run.mjs +20 -9
  16. package/dist/esm/run.mjs.map +1 -1
  17. package/dist/types/graphs/MultiAgentGraph.d.ts +17 -0
  18. package/package.json +1 -1
  19. package/src/graphs/Graph.ts +12 -1
  20. package/src/graphs/MultiAgentGraph.ts +105 -1
  21. package/src/graphs/__tests__/multi-agent-delegate.test.ts +191 -0
  22. package/src/llm/bedrock/index.ts +17 -0
  23. package/src/run.ts +20 -11
  24. package/src/scripts/test-bedrock-handoff-autonomous.ts +231 -0
  25. package/src/agents/AgentContext.js +0 -782
  26. package/src/agents/AgentContext.test.js +0 -421
  27. package/src/agents/__tests__/AgentContext.test.js +0 -678
  28. package/src/agents/__tests__/resolveStructuredOutputMode.test.js +0 -117
  29. package/src/common/enum.js +0 -192
  30. package/src/common/index.js +0 -3
  31. package/src/events.js +0 -166
  32. package/src/graphs/Graph.js +0 -1857
  33. package/src/graphs/MultiAgentGraph.js +0 -1092
  34. package/src/graphs/__tests__/structured-output.integration.test.js +0 -624
  35. package/src/graphs/__tests__/structured-output.test.js +0 -144
  36. package/src/graphs/contextManagement.e2e.test.js +0 -718
  37. package/src/graphs/contextManagement.test.js +0 -485
  38. package/src/graphs/handoffValidation.test.js +0 -276
  39. package/src/graphs/index.js +0 -3
  40. package/src/index.js +0 -28
  41. package/src/instrumentation.js +0 -21
  42. package/src/llm/anthropic/index.js +0 -319
  43. package/src/llm/anthropic/types.js +0 -46
  44. package/src/llm/anthropic/utils/message_inputs.js +0 -627
  45. package/src/llm/anthropic/utils/message_outputs.js +0 -290
  46. package/src/llm/anthropic/utils/output_parsers.js +0 -89
  47. package/src/llm/anthropic/utils/tools.js +0 -25
  48. package/src/llm/bedrock/__tests__/bedrock-caching.test.js +0 -392
  49. package/src/llm/bedrock/index.js +0 -303
  50. package/src/llm/bedrock/types.js +0 -2
  51. package/src/llm/bedrock/utils/index.js +0 -6
  52. package/src/llm/bedrock/utils/message_inputs.js +0 -463
  53. package/src/llm/bedrock/utils/message_outputs.js +0 -269
  54. package/src/llm/fake.js +0 -92
  55. package/src/llm/google/index.js +0 -215
  56. package/src/llm/google/types.js +0 -12
  57. package/src/llm/google/utils/common.js +0 -670
  58. package/src/llm/google/utils/tools.js +0 -111
  59. package/src/llm/google/utils/zod_to_genai_parameters.js +0 -47
  60. package/src/llm/openai/index.js +0 -1033
  61. package/src/llm/openai/types.js +0 -2
  62. package/src/llm/openai/utils/index.js +0 -756
  63. package/src/llm/openai/utils/isReasoningModel.test.js +0 -79
  64. package/src/llm/openrouter/index.js +0 -261
  65. package/src/llm/openrouter/reasoning.test.js +0 -181
  66. package/src/llm/providers.js +0 -36
  67. package/src/llm/text.js +0 -65
  68. package/src/llm/vertexai/index.js +0 -402
  69. package/src/messages/__tests__/tools.test.js +0 -392
  70. package/src/messages/cache.js +0 -404
  71. package/src/messages/cache.test.js +0 -1167
  72. package/src/messages/content.js +0 -48
  73. package/src/messages/content.test.js +0 -314
  74. package/src/messages/core.js +0 -359
  75. package/src/messages/ensureThinkingBlock.test.js +0 -997
  76. package/src/messages/format.js +0 -973
  77. package/src/messages/formatAgentMessages.test.js +0 -2278
  78. package/src/messages/formatAgentMessages.tools.test.js +0 -362
  79. package/src/messages/formatMessage.test.js +0 -608
  80. package/src/messages/ids.js +0 -18
  81. package/src/messages/index.js +0 -9
  82. package/src/messages/labelContentByAgent.test.js +0 -725
  83. package/src/messages/prune.js +0 -438
  84. package/src/messages/reducer.js +0 -60
  85. package/src/messages/shiftIndexTokenCountMap.test.js +0 -63
  86. package/src/messages/summarize.js +0 -146
  87. package/src/messages/summarize.test.js +0 -332
  88. package/src/messages/tools.js +0 -90
  89. package/src/mockStream.js +0 -81
  90. package/src/prompts/collab.js +0 -7
  91. package/src/prompts/index.js +0 -3
  92. package/src/prompts/taskmanager.js +0 -58
  93. package/src/run.js +0 -427
  94. package/src/schemas/index.js +0 -3
  95. package/src/schemas/schema-preparation.test.js +0 -370
  96. package/src/schemas/validate.js +0 -314
  97. package/src/schemas/validate.test.js +0 -264
  98. package/src/scripts/abort.js +0 -127
  99. package/src/scripts/ant_web_search.js +0 -130
  100. package/src/scripts/ant_web_search_edge_case.js +0 -133
  101. package/src/scripts/ant_web_search_error_edge_case.js +0 -119
  102. package/src/scripts/args.js +0 -41
  103. package/src/scripts/bedrock-cache-debug.js +0 -186
  104. package/src/scripts/bedrock-content-aggregation-test.js +0 -195
  105. package/src/scripts/bedrock-merge-test.js +0 -80
  106. package/src/scripts/bedrock-parallel-tools-test.js +0 -150
  107. package/src/scripts/caching.js +0 -106
  108. package/src/scripts/cli.js +0 -152
  109. package/src/scripts/cli2.js +0 -119
  110. package/src/scripts/cli3.js +0 -163
  111. package/src/scripts/cli4.js +0 -165
  112. package/src/scripts/cli5.js +0 -165
  113. package/src/scripts/code_exec.js +0 -171
  114. package/src/scripts/code_exec_files.js +0 -180
  115. package/src/scripts/code_exec_multi_session.js +0 -185
  116. package/src/scripts/code_exec_ptc.js +0 -265
  117. package/src/scripts/code_exec_session.js +0 -217
  118. package/src/scripts/code_exec_simple.js +0 -120
  119. package/src/scripts/content.js +0 -111
  120. package/src/scripts/empty_input.js +0 -125
  121. package/src/scripts/handoff-test.js +0 -96
  122. package/src/scripts/image.js +0 -138
  123. package/src/scripts/memory.js +0 -83
  124. package/src/scripts/multi-agent-chain.js +0 -271
  125. package/src/scripts/multi-agent-conditional.js +0 -185
  126. package/src/scripts/multi-agent-document-review-chain.js +0 -171
  127. package/src/scripts/multi-agent-hybrid-flow.js +0 -264
  128. package/src/scripts/multi-agent-parallel-start.js +0 -214
  129. package/src/scripts/multi-agent-parallel.js +0 -346
  130. package/src/scripts/multi-agent-sequence.js +0 -184
  131. package/src/scripts/multi-agent-supervisor.js +0 -324
  132. package/src/scripts/multi-agent-test.js +0 -147
  133. package/src/scripts/parallel-asymmetric-tools-test.js +0 -202
  134. package/src/scripts/parallel-full-metadata-test.js +0 -176
  135. package/src/scripts/parallel-tools-test.js +0 -256
  136. package/src/scripts/programmatic_exec.js +0 -277
  137. package/src/scripts/programmatic_exec_agent.js +0 -168
  138. package/src/scripts/search.js +0 -118
  139. package/src/scripts/sequential-full-metadata-test.js +0 -143
  140. package/src/scripts/simple.js +0 -174
  141. package/src/scripts/single-agent-metadata-test.js +0 -152
  142. package/src/scripts/stream.js +0 -113
  143. package/src/scripts/test-custom-prompt-key.js +0 -132
  144. package/src/scripts/test-handoff-input.js +0 -143
  145. package/src/scripts/test-handoff-preamble.js +0 -227
  146. package/src/scripts/test-handoff-steering.js +0 -353
  147. package/src/scripts/test-multi-agent-list-handoff.js +0 -318
  148. package/src/scripts/test-parallel-agent-labeling.js +0 -253
  149. package/src/scripts/test-parallel-handoffs.js +0 -229
  150. package/src/scripts/test-thinking-handoff-bedrock.js +0 -132
  151. package/src/scripts/test-thinking-handoff.js +0 -132
  152. package/src/scripts/test-thinking-to-thinking-handoff-bedrock.js +0 -140
  153. package/src/scripts/test-tool-before-handoff-role-order.js +0 -223
  154. package/src/scripts/test-tools-before-handoff.js +0 -187
  155. package/src/scripts/test_code_api.js +0 -263
  156. package/src/scripts/thinking-bedrock.js +0 -128
  157. package/src/scripts/thinking-vertexai.js +0 -130
  158. package/src/scripts/thinking.js +0 -134
  159. package/src/scripts/tool_search.js +0 -114
  160. package/src/scripts/tools.js +0 -125
  161. package/src/specs/agent-handoffs-bedrock.integration.test.js +0 -280
  162. package/src/specs/agent-handoffs.test.js +0 -924
  163. package/src/specs/anthropic.simple.test.js +0 -287
  164. package/src/specs/azure.simple.test.js +0 -381
  165. package/src/specs/cache.simple.test.js +0 -282
  166. package/src/specs/custom-event-await.test.js +0 -148
  167. package/src/specs/deepseek.simple.test.js +0 -189
  168. package/src/specs/emergency-prune.test.js +0 -308
  169. package/src/specs/moonshot.simple.test.js +0 -237
  170. package/src/specs/observability.integration.test.js +0 -1337
  171. package/src/specs/openai.simple.test.js +0 -233
  172. package/src/specs/openrouter.simple.test.js +0 -202
  173. package/src/specs/prune.test.js +0 -733
  174. package/src/specs/reasoning.test.js +0 -144
  175. package/src/specs/spec.utils.js +0 -4
  176. package/src/specs/thinking-handoff.test.js +0 -486
  177. package/src/specs/thinking-prune.test.js +0 -600
  178. package/src/specs/token-distribution-edge-case.test.js +0 -246
  179. package/src/specs/token-memoization.test.js +0 -32
  180. package/src/specs/tokens.test.js +0 -49
  181. package/src/specs/tool-error.test.js +0 -139
  182. package/src/splitStream.js +0 -204
  183. package/src/splitStream.test.js +0 -504
  184. package/src/stream.js +0 -650
  185. package/src/stream.test.js +0 -225
  186. package/src/test/mockTools.js +0 -340
  187. package/src/tools/BrowserTools.js +0 -245
  188. package/src/tools/Calculator.js +0 -38
  189. package/src/tools/Calculator.test.js +0 -225
  190. package/src/tools/CodeExecutor.js +0 -233
  191. package/src/tools/ProgrammaticToolCalling.js +0 -602
  192. package/src/tools/StreamingToolCallBuffer.js +0 -179
  193. package/src/tools/ToolNode.js +0 -930
  194. package/src/tools/ToolSearch.js +0 -904
  195. package/src/tools/__tests__/BrowserTools.test.js +0 -306
  196. package/src/tools/__tests__/ProgrammaticToolCalling.integration.test.js +0 -276
  197. package/src/tools/__tests__/ProgrammaticToolCalling.test.js +0 -807
  198. package/src/tools/__tests__/StreamingToolCallBuffer.test.js +0 -175
  199. package/src/tools/__tests__/ToolApproval.test.js +0 -675
  200. package/src/tools/__tests__/ToolNode.recovery.test.js +0 -200
  201. package/src/tools/__tests__/ToolNode.session.test.js +0 -319
  202. package/src/tools/__tests__/ToolSearch.integration.test.js +0 -125
  203. package/src/tools/__tests__/ToolSearch.test.js +0 -812
  204. package/src/tools/__tests__/handlers.test.js +0 -799
  205. package/src/tools/__tests__/truncation-recovery.integration.test.js +0 -362
  206. package/src/tools/handlers.js +0 -306
  207. package/src/tools/schema.js +0 -25
  208. package/src/tools/search/anthropic.js +0 -34
  209. package/src/tools/search/content.js +0 -116
  210. package/src/tools/search/content.test.js +0 -133
  211. package/src/tools/search/firecrawl.js +0 -173
  212. package/src/tools/search/format.js +0 -198
  213. package/src/tools/search/highlights.js +0 -241
  214. package/src/tools/search/index.js +0 -3
  215. package/src/tools/search/jina-reranker.test.js +0 -106
  216. package/src/tools/search/rerankers.js +0 -165
  217. package/src/tools/search/schema.js +0 -102
  218. package/src/tools/search/search.js +0 -561
  219. package/src/tools/search/serper-scraper.js +0 -126
  220. package/src/tools/search/test.js +0 -129
  221. package/src/tools/search/tool.js +0 -453
  222. package/src/tools/search/types.js +0 -2
  223. package/src/tools/search/utils.js +0 -59
  224. package/src/types/graph.js +0 -24
  225. package/src/types/graph.test.js +0 -192
  226. package/src/types/index.js +0 -7
  227. package/src/types/llm.js +0 -2
  228. package/src/types/messages.js +0 -2
  229. package/src/types/run.js +0 -2
  230. package/src/types/stream.js +0 -2
  231. package/src/types/tools.js +0 -2
  232. package/src/utils/contextAnalytics.js +0 -79
  233. package/src/utils/contextAnalytics.test.js +0 -166
  234. package/src/utils/events.js +0 -26
  235. package/src/utils/graph.js +0 -11
  236. package/src/utils/handlers.js +0 -65
  237. package/src/utils/index.js +0 -10
  238. package/src/utils/llm.js +0 -21
  239. package/src/utils/llmConfig.js +0 -205
  240. package/src/utils/logging.js +0 -37
  241. package/src/utils/misc.js +0 -51
  242. package/src/utils/run.js +0 -69
  243. package/src/utils/schema.js +0 -21
  244. package/src/utils/title.js +0 -119
  245. package/src/utils/tokens.js +0 -92
  246. package/src/utils/toonFormat.js +0 -379
@@ -1,263 +0,0 @@
1
- // src/scripts/test_code_api.ts
2
- /**
3
- * Direct test of the Code API to verify session file persistence.
4
- * This bypasses the LLM and tests the API directly.
5
- *
6
- * Run with: npx ts-node -r dotenv/config src/scripts/test_code_api.ts
7
- */
8
- import { config } from 'dotenv';
9
- config();
10
- import fetch from 'node-fetch';
11
- import { HttpsProxyAgent } from 'https-proxy-agent';
12
- const API_KEY = process.env.CODE_EXECUTOR_API_KEY ?? '';
13
- const BASE_URL = process.env.ILLUMA_CODE_BASEURL ?? 'https://api.illuma.ai/v1';
14
- const PROXY = process.env.PROXY;
15
- if (!API_KEY) {
16
- console.error('CODE_EXECUTOR_API_KEY not set');
17
- process.exit(1);
18
- }
19
- async function makeRequest(endpoint, body) {
20
- const fetchOptions = {
21
- method: body ? 'POST' : 'GET',
22
- headers: {
23
- 'Content-Type': 'application/json',
24
- 'User-Agent': 'Illuma/1.0',
25
- 'X-API-Key': API_KEY,
26
- },
27
- };
28
- if (body) {
29
- fetchOptions.body = JSON.stringify(body);
30
- }
31
- if (PROXY) {
32
- fetchOptions.agent = new HttpsProxyAgent(PROXY);
33
- }
34
- console.log(`\n>>> ${body ? 'POST' : 'GET'} ${endpoint}`);
35
- if (body) {
36
- console.log('Body:', JSON.stringify(body, null, 2));
37
- }
38
- const response = await fetch(endpoint, fetchOptions);
39
- const result = await response.json();
40
- console.log(`<<< Response (${response.status}):`);
41
- console.log(JSON.stringify(result, null, 2));
42
- if (!response.ok) {
43
- throw new Error(`HTTP ${response.status}: ${JSON.stringify(result)}`);
44
- }
45
- return result;
46
- }
47
- async function testCodeAPI() {
48
- console.log('='.repeat(60));
49
- console.log('TEST 1: Create a file');
50
- console.log('='.repeat(60));
51
- const createCode = `
52
- import json
53
-
54
- config = {
55
- "app_name": "TestApp",
56
- "version": "1.0.0",
57
- "debug": True
58
- }
59
-
60
- with open("/mnt/data/test_config.json", "w") as f:
61
- json.dump(config, f, indent=2)
62
-
63
- with open("/mnt/data/test_config.json", "r") as f:
64
- print(f.read())
65
- `;
66
- const result1 = (await makeRequest(`${BASE_URL}/exec`, {
67
- lang: 'py',
68
- code: createCode,
69
- }));
70
- const sessionId = result1.session_id;
71
- const files = result1.files ?? [];
72
- console.log('\n--- Result Summary ---');
73
- console.log('session_id:', sessionId);
74
- console.log('files:', files);
75
- console.log('stdout:', result1.stdout);
76
- console.log('stderr:', result1.stderr);
77
- if (!sessionId || files.length === 0) {
78
- console.error('\n❌ No session_id or files returned!');
79
- return;
80
- }
81
- // Check if files now include session_id (new API feature)
82
- const hasSessionIdInFiles = files.some((f) => f.session_id != null);
83
- console.log('\n✅ Files include session_id:', hasSessionIdInFiles);
84
- console.log('\n' + '='.repeat(60));
85
- console.log('TEST 2: Fetch files IMMEDIATELY (no delay - testing race condition fix)');
86
- console.log('='.repeat(60));
87
- const filesResult = (await makeRequest(`${BASE_URL}/files/${sessionId}?detail=full`));
88
- console.log('\n--- Files in session (detail=full) ---');
89
- for (const file of filesResult) {
90
- console.log('File:', file.name);
91
- console.log(' metadata:', file.metadata);
92
- }
93
- if (filesResult.length === 0) {
94
- console.log('\n⚠️ Files endpoint returned empty - race condition may still exist');
95
- }
96
- else {
97
- console.log('\n✅ Files available immediately!');
98
- }
99
- // Test new normalized detail level
100
- console.log('\n' + '='.repeat(60));
101
- console.log('TEST 2b: Fetch files with detail=normalized');
102
- console.log('='.repeat(60));
103
- const normalizedResult = (await makeRequest(`${BASE_URL}/files/${sessionId}?detail=normalized`));
104
- console.log('\n--- Files in session (detail=normalized) ---');
105
- console.log(JSON.stringify(normalizedResult, null, 2));
106
- console.log('\n' + '='.repeat(60));
107
- console.log('TEST 3: Read file IMMEDIATELY using files from original response');
108
- console.log('='.repeat(60));
109
- // Use files directly - if API returns session_id, use that; otherwise add it
110
- const fileReferences = files.map((file) => ({
111
- session_id: file.session_id ?? sessionId,
112
- id: file.id,
113
- name: file.name,
114
- }));
115
- console.log('\nFile references we will send:', JSON.stringify(fileReferences, null, 2));
116
- const readCode = `
117
- import json
118
-
119
- with open("/mnt/data/test_config.json", "r") as f:
120
- config = json.load(f)
121
- print("Read config:")
122
- print(json.dumps(config, indent=2))
123
- print("Version:", config.get("version"))
124
- `;
125
- const result2 = (await makeRequest(`${BASE_URL}/exec`, {
126
- lang: 'py',
127
- code: readCode,
128
- files: fileReferences,
129
- }));
130
- console.log('\n--- Result Summary ---');
131
- console.log('stdout:', result2.stdout);
132
- console.log('stderr:', result2.stderr);
133
- if (result2.stderr && result2.stderr.includes('FileNotFoundError')) {
134
- console.log('\n❌ File not found! The file reference format might be wrong.');
135
- // Try alternative format - just session_id
136
- console.log('\n' + '='.repeat(60));
137
- console.log('TEST 4: Try with just session_id in request');
138
- console.log('='.repeat(60));
139
- const result3 = (await makeRequest(`${BASE_URL}/exec`, {
140
- lang: 'py',
141
- code: readCode,
142
- session_id: sessionId,
143
- }));
144
- console.log('\n--- Result Summary ---');
145
- console.log('stdout:', result3.stdout);
146
- console.log('stderr:', result3.stderr);
147
- }
148
- else {
149
- console.log('\n✅ File read successfully!');
150
- }
151
- // ============================================================
152
- // TEST 4: MODIFY the file (same filename) - tests editable files
153
- // ============================================================
154
- console.log('\n' + '='.repeat(60));
155
- console.log('TEST 4: MODIFY file in-place (testing editable files feature)');
156
- console.log('='.repeat(60));
157
- const modifyCode = `
158
- import json
159
-
160
- # Read the existing file
161
- with open("/mnt/data/test_config.json", "r") as f:
162
- config = json.load(f)
163
-
164
- print("Original config:")
165
- print(json.dumps(config, indent=2))
166
-
167
- # Modify the config
168
- config["version"] = "2.0.0"
169
- config["modified"] = True
170
-
171
- # Write BACK to the SAME filename (should work now!)
172
- with open("/mnt/data/test_config.json", "w") as f:
173
- json.dump(config, f, indent=2)
174
-
175
- # Verify the write
176
- with open("/mnt/data/test_config.json", "r") as f:
177
- updated = json.load(f)
178
-
179
- print("\\nUpdated config:")
180
- print(json.dumps(updated, indent=2))
181
- `;
182
- const result3 = (await makeRequest(`${BASE_URL}/exec`, {
183
- lang: 'py',
184
- code: modifyCode,
185
- files: fileReferences,
186
- }));
187
- console.log('\n--- Result Summary ---');
188
- console.log('stdout:', result3.stdout);
189
- console.log('stderr:', result3.stderr);
190
- console.log('files:', JSON.stringify(result3.files, null, 2));
191
- if (result3.stderr && result3.stderr.includes('Permission denied')) {
192
- console.log('\n❌ Permission denied - files are still read-only!');
193
- }
194
- else if (result3.stderr && result3.stderr.includes('Error')) {
195
- console.log('\n❌ Error modifying file:', result3.stderr);
196
- }
197
- else {
198
- console.log('\n✅ File modified successfully!');
199
- // Check for modified_from lineage
200
- const modifiedFile = result3.files?.find((f) => f.name === 'test_config.json');
201
- if (modifiedFile) {
202
- console.log('\n--- Modified File Details ---');
203
- console.log(' id:', modifiedFile.id);
204
- console.log(' name:', modifiedFile.name);
205
- console.log(' session_id:', modifiedFile.session_id);
206
- if (modifiedFile.modified_from) {
207
- console.log(' modified_from:', JSON.stringify(modifiedFile.modified_from));
208
- console.log('\n✅ Lineage tracking working! File shows it was modified from previous session.');
209
- }
210
- else {
211
- console.log('\n⚠️ No modified_from field - lineage tracking not present');
212
- }
213
- }
214
- else {
215
- console.log('\n⚠️ Modified file not found in response files array');
216
- }
217
- }
218
- // ============================================================
219
- // TEST 5: Verify modification persists in next execution
220
- // ============================================================
221
- console.log('\n' + '='.repeat(60));
222
- console.log('TEST 5: Verify modified file can be read in subsequent execution');
223
- console.log('='.repeat(60));
224
- // Use the new file references from the modify response
225
- const newFileRefs = (result3.files ?? []).map((file) => ({
226
- session_id: file.session_id ?? result3.session_id,
227
- id: file.id,
228
- name: file.name,
229
- }));
230
- if (newFileRefs.length === 0) {
231
- console.log('\n⚠️ No files returned from modification, skipping verification');
232
- }
233
- else {
234
- console.log('\nUsing new file references:', JSON.stringify(newFileRefs, null, 2));
235
- const verifyCode = `
236
- import json
237
-
238
- with open("/mnt/data/test_config.json", "r") as f:
239
- config = json.load(f)
240
-
241
- print("Verified config:")
242
- print(json.dumps(config, indent=2))
243
-
244
- if config.get("version") == "2.0.0" and config.get("modified") == True:
245
- print("\\n✅ Modification persisted correctly!")
246
- else:
247
- print("\\n❌ Modification did NOT persist!")
248
- `;
249
- const result4 = (await makeRequest(`${BASE_URL}/exec`, {
250
- lang: 'py',
251
- code: verifyCode,
252
- files: newFileRefs,
253
- }));
254
- console.log('\n--- Result Summary ---');
255
- console.log('stdout:', result4.stdout);
256
- console.log('stderr:', result4.stderr);
257
- }
258
- }
259
- testCodeAPI().catch((err) => {
260
- console.error('Error:', err);
261
- process.exit(1);
262
- });
263
- //# sourceMappingURL=test_code_api.js.map
@@ -1,128 +0,0 @@
1
- // src/scripts/thinking-bedrock.ts
2
- import { config } from 'dotenv';
3
- config();
4
- import { HumanMessage } from '@langchain/core/messages';
5
- import { ChatModelStreamHandler, createContentAggregator } from '@/stream';
6
- import { createCodeExecutionTool } from '@/tools/CodeExecutor';
7
- import { ToolEndHandler, ModelEndHandler } from '@/events';
8
- import { GraphEvents, Providers } from '@/common';
9
- import { getLLMConfig } from '@/utils/llmConfig';
10
- import { getArgs } from '@/scripts/args';
11
- import { Run } from '@/run';
12
- const conversationHistory = [];
13
- let _contentParts = [];
14
- const collectedUsage = [];
15
- async function testBedrockThinking() {
16
- const { userName } = await getArgs();
17
- const instructions = `You are a helpful AI assistant for ${userName}. When answering questions, be thorough in your reasoning.`;
18
- const { contentParts, aggregateContent } = createContentAggregator();
19
- _contentParts = contentParts;
20
- // Set up event handlers
21
- const customHandlers = {
22
- [GraphEvents.TOOL_END]: new ToolEndHandler(),
23
- [GraphEvents.CHAT_MODEL_END]: new ModelEndHandler(collectedUsage),
24
- [GraphEvents.CHAT_MODEL_STREAM]: new ChatModelStreamHandler(),
25
- [GraphEvents.ON_RUN_STEP_COMPLETED]: {
26
- handle: (event, data) => {
27
- console.log('====== ON_RUN_STEP_COMPLETED ======');
28
- aggregateContent({
29
- event,
30
- data: data,
31
- });
32
- },
33
- },
34
- [GraphEvents.ON_RUN_STEP]: {
35
- handle: (event, data) => {
36
- aggregateContent({ event, data });
37
- },
38
- },
39
- [GraphEvents.ON_RUN_STEP_DELTA]: {
40
- handle: (event, data) => {
41
- aggregateContent({ event, data });
42
- },
43
- },
44
- [GraphEvents.ON_MESSAGE_DELTA]: {
45
- handle: (event, data) => {
46
- aggregateContent({ event, data });
47
- },
48
- },
49
- [GraphEvents.ON_REASONING_DELTA]: {
50
- handle: (event, data) => {
51
- aggregateContent({ event, data });
52
- },
53
- },
54
- };
55
- const baseLlmConfig = getLLMConfig(Providers.BEDROCK);
56
- // Enable thinking with token budget for Bedrock
57
- const llmConfig = {
58
- ...baseLlmConfig,
59
- model: 'us.anthropic.claude-3-7-sonnet-20250219-v1:0',
60
- maxTokens: 5000,
61
- additionalModelRequestFields: {
62
- thinking: { type: 'enabled', budget_tokens: 2000 },
63
- },
64
- };
65
- const run = await Run.create({
66
- runId: 'test-bedrock-thinking-id',
67
- graphConfig: {
68
- instructions,
69
- type: 'standard',
70
- tools: [createCodeExecutionTool()],
71
- llmConfig,
72
- },
73
- returnContent: true,
74
- skipCleanup: true,
75
- customHandlers: customHandlers,
76
- });
77
- const config = {
78
- configurable: {
79
- thread_id: 'bedrock-thinking-test-thread',
80
- },
81
- streamMode: 'values',
82
- version: 'v2',
83
- };
84
- // Test 1: Regular thinking mode
85
- console.log('\n\nTest 1: Bedrock Regular thinking mode');
86
- const userMessage1 = `Please print 'hello world' in python`;
87
- conversationHistory.push(new HumanMessage(userMessage1));
88
- console.log('Running first query with Bedrock thinking enabled...');
89
- const firstInputs = { messages: [...conversationHistory] };
90
- await run.processStream(firstInputs, config);
91
- // Extract and display thinking blocks
92
- const finalMessages = run.getRunMessages();
93
- console.log('\n\nFinal messages after Test 1:');
94
- console.dir(finalMessages, { depth: null });
95
- // Test 2: Try multi-turn conversation
96
- console.log('\n\nTest 2: Multi-turn conversation with Bedrock thinking enabled');
97
- const userMessage2 = `Given your previous analysis, what would be the most significant technical challenges in making this transition?`;
98
- conversationHistory.push(new HumanMessage(userMessage2));
99
- console.log('Running second query with Bedrock thinking enabled...');
100
- const secondInputs = { messages: [...conversationHistory] };
101
- await run.processStream(secondInputs, config);
102
- // Display thinking blocks for second response
103
- const finalMessages2 = run.getRunMessages();
104
- console.log('\n\nBedrock thinking feature test completed!');
105
- console.dir(finalMessages2, { depth: null });
106
- console.log('\n\nContent parts:');
107
- console.dir(_contentParts, { depth: null });
108
- }
109
- process.on('unhandledRejection', (reason, promise) => {
110
- console.error('Unhandled Rejection at:', promise, 'reason:', reason);
111
- console.log('Conversation history:');
112
- console.dir(conversationHistory, { depth: null });
113
- console.log('Content parts:');
114
- console.dir(_contentParts, { depth: null });
115
- process.exit(1);
116
- });
117
- process.on('uncaughtException', (err) => {
118
- console.error('Uncaught Exception:', err);
119
- });
120
- testBedrockThinking().catch((err) => {
121
- console.error(err);
122
- console.log('Conversation history:');
123
- console.dir(conversationHistory, { depth: null });
124
- console.log('Content parts:');
125
- console.dir(_contentParts, { depth: null });
126
- process.exit(1);
127
- });
128
- //# sourceMappingURL=thinking-bedrock.js.map
@@ -1,130 +0,0 @@
1
- // src/scripts/thinking-vertexai.ts
2
- import { config } from 'dotenv';
3
- config();
4
- import { HumanMessage } from '@langchain/core/messages';
5
- import { ChatModelStreamHandler, createContentAggregator } from '@/stream';
6
- import { ToolEndHandler, ModelEndHandler } from '@/events';
7
- import { GraphEvents, Providers } from '@/common';
8
- import { getLLMConfig } from '@/utils/llmConfig';
9
- import { getArgs } from '@/scripts/args';
10
- import { Run } from '@/run';
11
- const conversationHistory = [];
12
- let _contentParts = [];
13
- const collectedUsage = [];
14
- async function testVertexAIThinking() {
15
- const { userName } = await getArgs();
16
- const instructions = `You are a helpful AI assistant for ${userName}. When answering questions, be thorough in your reasoning.`;
17
- const { contentParts, aggregateContent } = createContentAggregator();
18
- _contentParts = contentParts;
19
- // Set up event handlers
20
- const customHandlers = {
21
- [GraphEvents.TOOL_END]: new ToolEndHandler(),
22
- [GraphEvents.CHAT_MODEL_END]: new ModelEndHandler(collectedUsage),
23
- [GraphEvents.CHAT_MODEL_STREAM]: new ChatModelStreamHandler(),
24
- [GraphEvents.ON_RUN_STEP_COMPLETED]: {
25
- handle: (event, data) => {
26
- console.log('====== ON_RUN_STEP_COMPLETED ======');
27
- aggregateContent({
28
- event,
29
- data: data,
30
- });
31
- },
32
- },
33
- [GraphEvents.ON_RUN_STEP]: {
34
- handle: (event, data) => {
35
- aggregateContent({ event, data });
36
- },
37
- },
38
- [GraphEvents.ON_RUN_STEP_DELTA]: {
39
- handle: (event, data) => {
40
- aggregateContent({ event, data });
41
- },
42
- },
43
- [GraphEvents.ON_MESSAGE_DELTA]: {
44
- handle: (event, data) => {
45
- aggregateContent({ event, data });
46
- },
47
- },
48
- [GraphEvents.ON_REASONING_DELTA]: {
49
- handle: (event, data) => {
50
- console.log('[ON_REASONING_DELTA]', JSON.stringify(data.delta.content?.[0]).slice(0, 100));
51
- aggregateContent({ event, data });
52
- },
53
- },
54
- };
55
- const baseLlmConfig = getLLMConfig(Providers.VERTEXAI);
56
- const llmConfig = {
57
- ...baseLlmConfig,
58
- model: 'gemini-3-flash-preview',
59
- location: 'global',
60
- streaming: true,
61
- streamUsage: true,
62
- thinkingConfig: {
63
- thinkingLevel: 'HIGH',
64
- includeThoughts: true,
65
- },
66
- };
67
- const run = await Run.create({
68
- runId: 'test-vertexai-thinking-id',
69
- graphConfig: {
70
- instructions,
71
- type: 'standard',
72
- llmConfig,
73
- },
74
- returnContent: true,
75
- skipCleanup: true,
76
- customHandlers: customHandlers,
77
- });
78
- const streamConfig = {
79
- configurable: {
80
- thread_id: 'vertexai-thinking-test-thread',
81
- },
82
- streamMode: 'values',
83
- version: 'v2',
84
- };
85
- // Test 1: Regular thinking mode
86
- console.log('\n\nTest 1: Vertex AI thinking mode with thinkingLevel=HIGH');
87
- const userMessage1 = 'How many r\'s are in the word "strawberry"? Think carefully.';
88
- conversationHistory.push(new HumanMessage(userMessage1));
89
- console.log('Running first query with Vertex AI thinking enabled...');
90
- const firstInputs = { messages: [...conversationHistory] };
91
- await run.processStream(firstInputs, streamConfig);
92
- // Extract and display results
93
- const finalMessages = run.getRunMessages();
94
- console.log('\n\nFinal messages after Test 1:');
95
- console.dir(finalMessages, { depth: null });
96
- // Test 2: Multi-turn conversation
97
- console.log('\n\nTest 2: Multi-turn conversation with Vertex AI thinking enabled');
98
- const userMessage2 = 'Now count the number of letters in "Mississippi". Explain step by step.';
99
- conversationHistory.push(new HumanMessage(userMessage2));
100
- console.log('Running second query with Vertex AI thinking enabled...');
101
- const secondInputs = { messages: [...conversationHistory] };
102
- await run.processStream(secondInputs, streamConfig);
103
- const finalMessages2 = run.getRunMessages();
104
- console.log('\n\nVertex AI thinking feature test completed!');
105
- console.dir(finalMessages2, { depth: null });
106
- console.log('\n\nContent parts:');
107
- console.dir(_contentParts, { depth: null });
108
- console.log('\n\nCollected usage:');
109
- console.dir(collectedUsage, { depth: null });
110
- }
111
- process.on('unhandledRejection', (reason, promise) => {
112
- console.error('Unhandled Rejection at:', promise, 'reason:', reason);
113
- console.log('Conversation history:');
114
- console.dir(conversationHistory, { depth: null });
115
- console.log('Content parts:');
116
- console.dir(_contentParts, { depth: null });
117
- process.exit(1);
118
- });
119
- process.on('uncaughtException', (err) => {
120
- console.error('Uncaught Exception:', err);
121
- });
122
- testVertexAIThinking().catch((err) => {
123
- console.error(err);
124
- console.log('Conversation history:');
125
- console.dir(conversationHistory, { depth: null });
126
- console.log('Content parts:');
127
- console.dir(_contentParts, { depth: null });
128
- process.exit(1);
129
- });
130
- //# sourceMappingURL=thinking-vertexai.js.map
@@ -1,134 +0,0 @@
1
- // src/scripts/test-thinking.ts
2
- import { config } from 'dotenv';
3
- config();
4
- import { HumanMessage, } from '@langchain/core/messages';
5
- import { ChatModelStreamHandler, createContentAggregator } from '@/stream';
6
- import { createCodeExecutionTool } from '@/tools/CodeExecutor';
7
- import { ToolEndHandler, ModelEndHandler } from '@/events';
8
- import { GraphEvents, Providers } from '@/common';
9
- import { getLLMConfig } from '@/utils/llmConfig';
10
- import { getArgs } from '@/scripts/args';
11
- import { Run } from '@/run';
12
- const conversationHistory = [];
13
- let _contentParts = [];
14
- const collectedUsage = [];
15
- async function testThinking() {
16
- const { userName } = await getArgs();
17
- const instructions = `You are a helpful AI assistant for ${userName}. When answering questions, be thorough in your reasoning.`;
18
- const { contentParts, aggregateContent } = createContentAggregator();
19
- _contentParts = contentParts;
20
- // Set up event handlers
21
- const customHandlers = {
22
- [GraphEvents.TOOL_END]: new ToolEndHandler(),
23
- [GraphEvents.CHAT_MODEL_END]: new ModelEndHandler(collectedUsage),
24
- [GraphEvents.CHAT_MODEL_STREAM]: new ChatModelStreamHandler(),
25
- [GraphEvents.ON_RUN_STEP_COMPLETED]: {
26
- handle: (event, data) => {
27
- console.log('====== ON_RUN_STEP_COMPLETED ======');
28
- aggregateContent({
29
- event,
30
- data: data,
31
- });
32
- },
33
- },
34
- [GraphEvents.ON_RUN_STEP]: {
35
- handle: (event, data) => {
36
- aggregateContent({ event, data });
37
- },
38
- },
39
- [GraphEvents.ON_RUN_STEP_DELTA]: {
40
- handle: (event, data) => {
41
- aggregateContent({ event, data });
42
- },
43
- },
44
- [GraphEvents.ON_MESSAGE_DELTA]: {
45
- handle: (event, data) => {
46
- aggregateContent({ event, data });
47
- },
48
- },
49
- [GraphEvents.ON_REASONING_DELTA]: {
50
- handle: (event, data) => {
51
- aggregateContent({ event, data });
52
- },
53
- },
54
- };
55
- const baseLlmConfig = getLLMConfig(Providers.ANTHROPIC);
56
- // Enable thinking with token budget
57
- const llmConfig = {
58
- ...baseLlmConfig,
59
- model: 'claude-3-7-sonnet-latest',
60
- thinking: { type: 'enabled', budget_tokens: 2000 },
61
- };
62
- const run = await Run.create({
63
- runId: 'test-thinking-id',
64
- graphConfig: {
65
- instructions,
66
- type: 'standard',
67
- tools: [createCodeExecutionTool()],
68
- llmConfig,
69
- },
70
- returnContent: true,
71
- skipCleanup: true,
72
- customHandlers: customHandlers,
73
- });
74
- const config = {
75
- configurable: {
76
- thread_id: 'thinking-test-thread',
77
- },
78
- streamMode: 'values',
79
- version: 'v2',
80
- };
81
- // Test 1: Regular thinking mode
82
- console.log('\n\nTest 1: Regular thinking mode');
83
- // const userMessage1 = `What would be the environmental and economic impacts if all cars globally were replaced by electric vehicles overnight?`;
84
- const userMessage1 = `Please print 'hello world' in python`;
85
- conversationHistory.push(new HumanMessage(userMessage1));
86
- console.log('Running first query with thinking enabled...');
87
- const firstInputs = { messages: [...conversationHistory] };
88
- await run.processStream(firstInputs, config);
89
- // Extract and display thinking blocks
90
- const finalMessages = run.getRunMessages();
91
- // Test 2: Try multi-turn conversation
92
- console.log('\n\nTest 2: Multi-turn conversation with thinking enabled');
93
- const userMessage2 = `Given your previous analysis, what would be the most significant technical challenges in making this transition?`;
94
- conversationHistory.push(new HumanMessage(userMessage2));
95
- console.log('Running second query with thinking enabled...');
96
- const secondInputs = { messages: [...conversationHistory] };
97
- await run.processStream(secondInputs, config);
98
- // Display thinking blocks for second response
99
- const finalMessages2 = run.getRunMessages();
100
- // Test 3: Redacted thinking mode
101
- console.log('\n\nTest 3: Redacted thinking mode');
102
- const magicString = 'ANTHROPIC_MAGIC_STRING_TRIGGER_REDACTED_THINKING_46C9A13E193C177646C7398A98432ECCCE4C1253D5E2D82641AC0E52CC2876CB';
103
- const userMessage3 = `${magicString}\n\nExplain how quantum computing works in simple terms.`;
104
- // Reset conversation for clean test
105
- conversationHistory.length = 0;
106
- conversationHistory.push(new HumanMessage(userMessage3));
107
- console.log('Running query with redacted thinking...');
108
- const thirdInputs = { messages: [...conversationHistory] };
109
- await run.processStream(thirdInputs, config);
110
- // Display redacted thinking blocks
111
- const finalMessages3 = run.getRunMessages();
112
- console.log('\n\nThinking feature test completed!');
113
- console.dir(finalMessages3, { depth: null });
114
- }
115
- process.on('unhandledRejection', (reason, promise) => {
116
- console.error('Unhandled Rejection at:', promise, 'reason:', reason);
117
- console.log('Conversation history:');
118
- console.dir(conversationHistory, { depth: null });
119
- console.log('Content parts:');
120
- console.dir(_contentParts, { depth: null });
121
- process.exit(1);
122
- });
123
- process.on('uncaughtException', (err) => {
124
- console.error('Uncaught Exception:', err);
125
- });
126
- testThinking().catch((err) => {
127
- console.error(err);
128
- console.log('Conversation history:');
129
- console.dir(conversationHistory, { depth: null });
130
- console.log('Content parts:');
131
- console.dir(_contentParts, { depth: null });
132
- process.exit(1);
133
- });
134
- //# sourceMappingURL=thinking.js.map