opencode-swarm-plugin 0.43.0 → 0.44.1

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 (208) hide show
  1. package/bin/cass.characterization.test.ts +422 -0
  2. package/bin/swarm.serve.test.ts +6 -4
  3. package/bin/swarm.test.ts +68 -0
  4. package/bin/swarm.ts +81 -8
  5. package/dist/compaction-prompt-scoring.js +139 -0
  6. package/dist/contributor-tools.d.ts +42 -0
  7. package/dist/contributor-tools.d.ts.map +1 -0
  8. package/dist/eval-capture.js +12811 -0
  9. package/dist/hive.d.ts.map +1 -1
  10. package/dist/index.d.ts +12 -0
  11. package/dist/index.d.ts.map +1 -1
  12. package/dist/index.js +7728 -62590
  13. package/dist/plugin.js +23833 -78695
  14. package/dist/sessions/agent-discovery.d.ts +59 -0
  15. package/dist/sessions/agent-discovery.d.ts.map +1 -0
  16. package/dist/sessions/index.d.ts +10 -0
  17. package/dist/sessions/index.d.ts.map +1 -0
  18. package/dist/swarm-orchestrate.d.ts.map +1 -1
  19. package/dist/swarm-prompts.d.ts.map +1 -1
  20. package/dist/swarm-review.d.ts.map +1 -1
  21. package/package.json +17 -5
  22. package/.changeset/swarm-insights-data-layer.md +0 -63
  23. package/.hive/analysis/eval-failure-analysis-2025-12-25.md +0 -331
  24. package/.hive/analysis/session-data-quality-audit.md +0 -320
  25. package/.hive/eval-results.json +0 -483
  26. package/.hive/issues.jsonl +0 -138
  27. package/.hive/memories.jsonl +0 -729
  28. package/.opencode/eval-history.jsonl +0 -327
  29. package/.turbo/turbo-build.log +0 -9
  30. package/CHANGELOG.md +0 -2255
  31. package/SCORER-ANALYSIS.md +0 -598
  32. package/docs/analysis/subagent-coordination-patterns.md +0 -902
  33. package/docs/analysis-socratic-planner-pattern.md +0 -504
  34. package/docs/planning/ADR-001-monorepo-structure.md +0 -171
  35. package/docs/planning/ADR-002-package-extraction.md +0 -393
  36. package/docs/planning/ADR-003-performance-improvements.md +0 -451
  37. package/docs/planning/ADR-004-message-queue-features.md +0 -187
  38. package/docs/planning/ADR-005-devtools-observability.md +0 -202
  39. package/docs/planning/ADR-007-swarm-enhancements-worktree-review.md +0 -168
  40. package/docs/planning/ADR-008-worker-handoff-protocol.md +0 -293
  41. package/docs/planning/ADR-009-oh-my-opencode-patterns.md +0 -353
  42. package/docs/planning/ROADMAP.md +0 -368
  43. package/docs/semantic-memory-cli-syntax.md +0 -123
  44. package/docs/swarm-mail-architecture.md +0 -1147
  45. package/docs/testing/context-recovery-test.md +0 -470
  46. package/evals/ARCHITECTURE.md +0 -1189
  47. package/evals/README.md +0 -768
  48. package/evals/compaction-prompt.eval.ts +0 -149
  49. package/evals/compaction-resumption.eval.ts +0 -289
  50. package/evals/coordinator-behavior.eval.ts +0 -307
  51. package/evals/coordinator-session.eval.ts +0 -154
  52. package/evals/evalite.config.ts.bak +0 -15
  53. package/evals/example.eval.ts +0 -31
  54. package/evals/fixtures/compaction-cases.ts +0 -350
  55. package/evals/fixtures/compaction-prompt-cases.ts +0 -311
  56. package/evals/fixtures/coordinator-sessions.ts +0 -328
  57. package/evals/fixtures/decomposition-cases.ts +0 -105
  58. package/evals/lib/compaction-loader.test.ts +0 -248
  59. package/evals/lib/compaction-loader.ts +0 -320
  60. package/evals/lib/data-loader.evalite-test.ts +0 -289
  61. package/evals/lib/data-loader.test.ts +0 -345
  62. package/evals/lib/data-loader.ts +0 -281
  63. package/evals/lib/llm.ts +0 -115
  64. package/evals/scorers/compaction-prompt-scorers.ts +0 -145
  65. package/evals/scorers/compaction-scorers.ts +0 -305
  66. package/evals/scorers/coordinator-discipline.evalite-test.ts +0 -539
  67. package/evals/scorers/coordinator-discipline.ts +0 -325
  68. package/evals/scorers/index.test.ts +0 -146
  69. package/evals/scorers/index.ts +0 -328
  70. package/evals/scorers/outcome-scorers.evalite-test.ts +0 -27
  71. package/evals/scorers/outcome-scorers.ts +0 -349
  72. package/evals/swarm-decomposition.eval.ts +0 -121
  73. package/examples/commands/swarm.md +0 -745
  74. package/examples/plugin-wrapper-template.ts +0 -2426
  75. package/examples/skills/hive-workflow/SKILL.md +0 -212
  76. package/examples/skills/skill-creator/SKILL.md +0 -223
  77. package/examples/skills/swarm-coordination/SKILL.md +0 -292
  78. package/global-skills/cli-builder/SKILL.md +0 -344
  79. package/global-skills/cli-builder/references/advanced-patterns.md +0 -244
  80. package/global-skills/learning-systems/SKILL.md +0 -644
  81. package/global-skills/skill-creator/LICENSE.txt +0 -202
  82. package/global-skills/skill-creator/SKILL.md +0 -352
  83. package/global-skills/skill-creator/references/output-patterns.md +0 -82
  84. package/global-skills/skill-creator/references/workflows.md +0 -28
  85. package/global-skills/swarm-coordination/SKILL.md +0 -995
  86. package/global-skills/swarm-coordination/references/coordinator-patterns.md +0 -235
  87. package/global-skills/swarm-coordination/references/strategies.md +0 -138
  88. package/global-skills/system-design/SKILL.md +0 -213
  89. package/global-skills/testing-patterns/SKILL.md +0 -430
  90. package/global-skills/testing-patterns/references/dependency-breaking-catalog.md +0 -586
  91. package/opencode-swarm-plugin-0.30.7.tgz +0 -0
  92. package/opencode-swarm-plugin-0.31.0.tgz +0 -0
  93. package/scripts/cleanup-test-memories.ts +0 -346
  94. package/scripts/init-skill.ts +0 -222
  95. package/scripts/migrate-unknown-sessions.ts +0 -349
  96. package/scripts/validate-skill.ts +0 -204
  97. package/src/agent-mail.ts +0 -1724
  98. package/src/anti-patterns.test.ts +0 -1167
  99. package/src/anti-patterns.ts +0 -448
  100. package/src/compaction-capture.integration.test.ts +0 -257
  101. package/src/compaction-hook.test.ts +0 -838
  102. package/src/compaction-hook.ts +0 -1204
  103. package/src/compaction-observability.integration.test.ts +0 -139
  104. package/src/compaction-observability.test.ts +0 -187
  105. package/src/compaction-observability.ts +0 -324
  106. package/src/compaction-prompt-scorers.test.ts +0 -475
  107. package/src/compaction-prompt-scoring.ts +0 -300
  108. package/src/dashboard.test.ts +0 -611
  109. package/src/dashboard.ts +0 -462
  110. package/src/error-enrichment.test.ts +0 -403
  111. package/src/error-enrichment.ts +0 -219
  112. package/src/eval-capture.test.ts +0 -1015
  113. package/src/eval-capture.ts +0 -929
  114. package/src/eval-gates.test.ts +0 -306
  115. package/src/eval-gates.ts +0 -218
  116. package/src/eval-history.test.ts +0 -508
  117. package/src/eval-history.ts +0 -214
  118. package/src/eval-learning.test.ts +0 -378
  119. package/src/eval-learning.ts +0 -360
  120. package/src/eval-runner.test.ts +0 -223
  121. package/src/eval-runner.ts +0 -402
  122. package/src/export-tools.test.ts +0 -476
  123. package/src/export-tools.ts +0 -257
  124. package/src/hive.integration.test.ts +0 -2241
  125. package/src/hive.ts +0 -1628
  126. package/src/index.ts +0 -935
  127. package/src/learning.integration.test.ts +0 -1815
  128. package/src/learning.ts +0 -1079
  129. package/src/logger.test.ts +0 -189
  130. package/src/logger.ts +0 -135
  131. package/src/mandate-promotion.test.ts +0 -473
  132. package/src/mandate-promotion.ts +0 -239
  133. package/src/mandate-storage.integration.test.ts +0 -601
  134. package/src/mandate-storage.test.ts +0 -578
  135. package/src/mandate-storage.ts +0 -794
  136. package/src/mandates.ts +0 -540
  137. package/src/memory-tools.test.ts +0 -195
  138. package/src/memory-tools.ts +0 -344
  139. package/src/memory.integration.test.ts +0 -334
  140. package/src/memory.test.ts +0 -158
  141. package/src/memory.ts +0 -527
  142. package/src/model-selection.test.ts +0 -188
  143. package/src/model-selection.ts +0 -68
  144. package/src/observability-tools.test.ts +0 -359
  145. package/src/observability-tools.ts +0 -871
  146. package/src/output-guardrails.test.ts +0 -438
  147. package/src/output-guardrails.ts +0 -381
  148. package/src/pattern-maturity.test.ts +0 -1160
  149. package/src/pattern-maturity.ts +0 -525
  150. package/src/planning-guardrails.test.ts +0 -491
  151. package/src/planning-guardrails.ts +0 -438
  152. package/src/plugin.ts +0 -23
  153. package/src/post-compaction-tracker.test.ts +0 -251
  154. package/src/post-compaction-tracker.ts +0 -237
  155. package/src/query-tools.test.ts +0 -636
  156. package/src/query-tools.ts +0 -324
  157. package/src/rate-limiter.integration.test.ts +0 -466
  158. package/src/rate-limiter.ts +0 -774
  159. package/src/replay-tools.test.ts +0 -496
  160. package/src/replay-tools.ts +0 -240
  161. package/src/repo-crawl.integration.test.ts +0 -441
  162. package/src/repo-crawl.ts +0 -610
  163. package/src/schemas/cell-events.test.ts +0 -347
  164. package/src/schemas/cell-events.ts +0 -807
  165. package/src/schemas/cell.ts +0 -257
  166. package/src/schemas/evaluation.ts +0 -166
  167. package/src/schemas/index.test.ts +0 -199
  168. package/src/schemas/index.ts +0 -286
  169. package/src/schemas/mandate.ts +0 -232
  170. package/src/schemas/swarm-context.ts +0 -115
  171. package/src/schemas/task.ts +0 -161
  172. package/src/schemas/worker-handoff.test.ts +0 -302
  173. package/src/schemas/worker-handoff.ts +0 -131
  174. package/src/skills.integration.test.ts +0 -1192
  175. package/src/skills.test.ts +0 -643
  176. package/src/skills.ts +0 -1549
  177. package/src/storage.integration.test.ts +0 -341
  178. package/src/storage.ts +0 -884
  179. package/src/structured.integration.test.ts +0 -817
  180. package/src/structured.test.ts +0 -1046
  181. package/src/structured.ts +0 -762
  182. package/src/swarm-decompose.test.ts +0 -188
  183. package/src/swarm-decompose.ts +0 -1302
  184. package/src/swarm-deferred.integration.test.ts +0 -157
  185. package/src/swarm-deferred.test.ts +0 -38
  186. package/src/swarm-insights.test.ts +0 -214
  187. package/src/swarm-insights.ts +0 -459
  188. package/src/swarm-mail.integration.test.ts +0 -970
  189. package/src/swarm-mail.ts +0 -739
  190. package/src/swarm-orchestrate.integration.test.ts +0 -282
  191. package/src/swarm-orchestrate.test.ts +0 -548
  192. package/src/swarm-orchestrate.ts +0 -3084
  193. package/src/swarm-prompts.test.ts +0 -1270
  194. package/src/swarm-prompts.ts +0 -2077
  195. package/src/swarm-research.integration.test.ts +0 -701
  196. package/src/swarm-research.test.ts +0 -698
  197. package/src/swarm-research.ts +0 -472
  198. package/src/swarm-review.integration.test.ts +0 -285
  199. package/src/swarm-review.test.ts +0 -879
  200. package/src/swarm-review.ts +0 -709
  201. package/src/swarm-strategies.ts +0 -407
  202. package/src/swarm-worktree.test.ts +0 -501
  203. package/src/swarm-worktree.ts +0 -575
  204. package/src/swarm.integration.test.ts +0 -2377
  205. package/src/swarm.ts +0 -38
  206. package/src/tool-adapter.integration.test.ts +0 -1221
  207. package/src/tool-availability.ts +0 -461
  208. package/tsconfig.json +0 -28
@@ -1,345 +0,0 @@
1
- /**
2
- * Tests for data-loader quality filters
3
- *
4
- * TDD approach: RED → GREEN → REFACTOR
5
- */
6
- import { afterEach, beforeEach, describe, expect, test } from "bun:test";
7
- import * as fs from "node:fs";
8
- import * as os from "node:os";
9
- import * as path from "node:path";
10
- import type { CoordinatorEvent } from "../../src/eval-capture.js";
11
- import { loadCapturedSessions } from "./data-loader.js";
12
-
13
- // Test helper: create a temp session directory
14
- let tempSessionDir: string;
15
-
16
- beforeEach(() => {
17
- tempSessionDir = fs.mkdtempSync(path.join(os.tmpdir(), "test-sessions-"));
18
- });
19
-
20
- afterEach(() => {
21
- if (fs.existsSync(tempSessionDir)) {
22
- fs.rmSync(tempSessionDir, { recursive: true });
23
- }
24
- });
25
-
26
- /**
27
- * Helper: create a session JSONL file with events
28
- */
29
- function createSessionFile(
30
- sessionId: string,
31
- events: CoordinatorEvent[],
32
- ): void {
33
- const filePath = path.join(tempSessionDir, `${sessionId}.jsonl`);
34
- const lines = events.map((e) => JSON.stringify(e)).join("\n") + "\n";
35
- fs.writeFileSync(filePath, lines, "utf-8");
36
- }
37
-
38
- /**
39
- * Helper: create minimal events
40
- */
41
- function createEvent(
42
- sessionId: string,
43
- epicId: string,
44
- type: "DECISION" | "VIOLATION" | "OUTCOME",
45
- subtype: string,
46
- ): CoordinatorEvent {
47
- const base = {
48
- session_id: sessionId,
49
- epic_id: epicId,
50
- timestamp: new Date().toISOString(),
51
- payload: {},
52
- };
53
-
54
- if (type === "DECISION") {
55
- return {
56
- ...base,
57
- event_type: "DECISION" as const,
58
- decision_type: subtype as any,
59
- };
60
- } else if (type === "VIOLATION") {
61
- return {
62
- ...base,
63
- event_type: "VIOLATION" as const,
64
- violation_type: subtype as any,
65
- };
66
- } else {
67
- return {
68
- ...base,
69
- event_type: "OUTCOME" as const,
70
- outcome_type: subtype as any,
71
- };
72
- }
73
- }
74
-
75
- describe("loadCapturedSessions - quality filters", () => {
76
- test("filters out sessions with fewer than minEvents (default: 3)", async () => {
77
- // Create sessions with different event counts
78
- createSessionFile("session-2-events", [
79
- createEvent("session-2-events", "epic-1", "DECISION", "worker_spawned"),
80
- createEvent("session-2-events", "epic-1", "OUTCOME", "subtask_success"),
81
- ]);
82
-
83
- createSessionFile("session-3-events", [
84
- createEvent("session-3-events", "epic-2", "DECISION", "worker_spawned"),
85
- createEvent("session-3-events", "epic-2", "DECISION", "review_completed"),
86
- createEvent("session-3-events", "epic-2", "OUTCOME", "subtask_success"),
87
- ]);
88
-
89
- createSessionFile("session-5-events", [
90
- createEvent("session-5-events", "epic-3", "DECISION", "worker_spawned"),
91
- createEvent("session-5-events", "epic-3", "DECISION", "review_completed"),
92
- createEvent("session-5-events", "epic-3", "OUTCOME", "subtask_success"),
93
- createEvent("session-5-events", "epic-3", "OUTCOME", "subtask_success"),
94
- createEvent("session-5-events", "epic-3", "OUTCOME", "epic_complete"),
95
- ]);
96
-
97
- const sessions = await loadCapturedSessions({
98
- minEvents: 3,
99
- sessionDir: tempSessionDir,
100
- });
101
-
102
- // Should only get sessions with >= 3 events
103
- expect(sessions.length).toBe(2);
104
- expect(
105
- sessions.some((s) => s.session.session_id === "session-3-events"),
106
- ).toBe(true);
107
- expect(
108
- sessions.some((s) => s.session.session_id === "session-5-events"),
109
- ).toBe(true);
110
- expect(
111
- sessions.some((s) => s.session.session_id === "session-2-events"),
112
- ).toBe(false);
113
- });
114
-
115
- test("filters out sessions without worker_spawned event when requireWorkerSpawn=true", async () => {
116
- // Session WITH worker_spawned
117
- createSessionFile("session-with-spawn", [
118
- createEvent("session-with-spawn", "epic-1", "DECISION", "worker_spawned"),
119
- createEvent(
120
- "session-with-spawn",
121
- "epic-1",
122
- "DECISION",
123
- "review_completed",
124
- ),
125
- createEvent("session-with-spawn", "epic-1", "OUTCOME", "subtask_success"),
126
- ]);
127
-
128
- // Session WITHOUT worker_spawned
129
- createSessionFile("session-no-spawn", [
130
- createEvent(
131
- "session-no-spawn",
132
- "epic-2",
133
- "DECISION",
134
- "strategy_selected",
135
- ),
136
- createEvent(
137
- "session-no-spawn",
138
- "epic-2",
139
- "DECISION",
140
- "decomposition_complete",
141
- ),
142
- createEvent("session-no-spawn", "epic-2", "OUTCOME", "epic_complete"),
143
- ]);
144
-
145
- const sessions = await loadCapturedSessions({
146
- requireWorkerSpawn: true,
147
- sessionDir: tempSessionDir,
148
- });
149
-
150
- expect(sessions.length).toBe(1);
151
- expect(sessions[0]?.session.session_id).toBe("session-with-spawn");
152
- });
153
-
154
- test("filters out sessions without review_completed event when requireReview=true", async () => {
155
- // Session WITH review
156
- createSessionFile("session-with-review", [
157
- createEvent(
158
- "session-with-review",
159
- "epic-1",
160
- "DECISION",
161
- "worker_spawned",
162
- ),
163
- createEvent(
164
- "session-with-review",
165
- "epic-1",
166
- "DECISION",
167
- "review_completed",
168
- ),
169
- createEvent("session-with-review", "epic-1", "OUTCOME", "subtask_success"),
170
- ]);
171
-
172
- // Session WITHOUT review
173
- createSessionFile("session-no-review", [
174
- createEvent("session-no-review", "epic-2", "DECISION", "worker_spawned"),
175
- createEvent("session-no-review", "epic-2", "OUTCOME", "subtask_success"),
176
- createEvent("session-no-review", "epic-2", "OUTCOME", "epic_complete"),
177
- ]);
178
-
179
- const sessions = await loadCapturedSessions({
180
- requireReview: true,
181
- sessionDir: tempSessionDir,
182
- });
183
-
184
- expect(sessions.length).toBe(1);
185
- expect(sessions[0]?.session.session_id).toBe("session-with-review");
186
- });
187
-
188
- test("allows disabling filters individually", async () => {
189
- // Session with only 2 events, no worker_spawned, no review
190
- createSessionFile("session-low-quality", [
191
- createEvent(
192
- "session-low-quality",
193
- "epic-1",
194
- "DECISION",
195
- "strategy_selected",
196
- ),
197
- createEvent("session-low-quality", "epic-1", "OUTCOME", "epic_complete"),
198
- ]);
199
-
200
- // Disable all filters
201
- const sessions = await loadCapturedSessions({
202
- minEvents: 0,
203
- requireWorkerSpawn: false,
204
- requireReview: false,
205
- sessionDir: tempSessionDir,
206
- });
207
-
208
- expect(sessions.length).toBe(1);
209
- expect(sessions[0]?.session.session_id).toBe("session-low-quality");
210
- });
211
-
212
- test("applies limit AFTER filtering", async () => {
213
- // Create 5 high-quality sessions
214
- for (let i = 1; i <= 5; i++) {
215
- createSessionFile(`session-${i}`, [
216
- createEvent(`session-${i}`, `epic-${i}`, "DECISION", "worker_spawned"),
217
- createEvent(
218
- `session-${i}`,
219
- `epic-${i}`,
220
- "DECISION",
221
- "review_completed",
222
- ),
223
- createEvent(`session-${i}`, `epic-${i}`, "OUTCOME", "subtask_success"),
224
- ]);
225
- }
226
-
227
- // Create 3 low-quality sessions (will be filtered out)
228
- for (let i = 6; i <= 8; i++) {
229
- createSessionFile(`session-${i}`, [
230
- createEvent(`session-${i}`, `epic-${i}`, "DECISION", "strategy_selected"),
231
- ]);
232
- }
233
-
234
- // Filter first (remove 3 low-quality), then limit to 2
235
- const sessions = await loadCapturedSessions({
236
- minEvents: 3,
237
- requireWorkerSpawn: true,
238
- requireReview: true,
239
- limit: 2,
240
- sessionDir: tempSessionDir,
241
- });
242
-
243
- // Should get 2 sessions from the 5 high-quality ones
244
- expect(sessions.length).toBe(2);
245
- expect(sessions.every((s) => s.session.events.length >= 3)).toBe(true);
246
- });
247
-
248
- test("combines all filters correctly", async () => {
249
- // High-quality session (passes all filters)
250
- createSessionFile("session-high-quality", [
251
- createEvent(
252
- "session-high-quality",
253
- "epic-1",
254
- "DECISION",
255
- "worker_spawned",
256
- ),
257
- createEvent(
258
- "session-high-quality",
259
- "epic-1",
260
- "DECISION",
261
- "review_completed",
262
- ),
263
- createEvent("session-high-quality", "epic-1", "OUTCOME", "subtask_success"),
264
- createEvent("session-high-quality", "epic-1", "OUTCOME", "epic_complete"),
265
- ]);
266
-
267
- // Missing worker_spawned
268
- createSessionFile("session-no-spawn", [
269
- createEvent(
270
- "session-no-spawn",
271
- "epic-2",
272
- "DECISION",
273
- "review_completed",
274
- ),
275
- createEvent("session-no-spawn", "epic-2", "OUTCOME", "subtask_success"),
276
- createEvent("session-no-spawn", "epic-2", "OUTCOME", "epic_complete"),
277
- ]);
278
-
279
- // Missing review_completed
280
- createSessionFile("session-no-review", [
281
- createEvent("session-no-review", "epic-3", "DECISION", "worker_spawned"),
282
- createEvent("session-no-review", "epic-3", "OUTCOME", "subtask_success"),
283
- createEvent("session-no-review", "epic-3", "OUTCOME", "epic_complete"),
284
- ]);
285
-
286
- // Too few events
287
- createSessionFile("session-too-few", [
288
- createEvent("session-too-few", "epic-4", "DECISION", "worker_spawned"),
289
- createEvent("session-too-few", "epic-4", "DECISION", "review_completed"),
290
- ]);
291
-
292
- const sessions = await loadCapturedSessions({
293
- minEvents: 3,
294
- requireWorkerSpawn: true,
295
- requireReview: true,
296
- sessionDir: tempSessionDir,
297
- });
298
-
299
- // Only high-quality session should pass
300
- expect(sessions.length).toBe(1);
301
- expect(sessions[0]?.session.session_id).toBe("session-high-quality");
302
- });
303
-
304
- test("defaults are: minEvents=3, requireWorkerSpawn=true, requireReview=true", async () => {
305
- // Create one session that meets defaults
306
- createSessionFile("session-meets-defaults", [
307
- createEvent(
308
- "session-meets-defaults",
309
- "epic-1",
310
- "DECISION",
311
- "worker_spawned",
312
- ),
313
- createEvent(
314
- "session-meets-defaults",
315
- "epic-1",
316
- "DECISION",
317
- "review_completed",
318
- ),
319
- createEvent(
320
- "session-meets-defaults",
321
- "epic-1",
322
- "OUTCOME",
323
- "subtask_success",
324
- ),
325
- ]);
326
-
327
- // Create one that doesn't
328
- createSessionFile("session-fails-defaults", [
329
- createEvent(
330
- "session-fails-defaults",
331
- "epic-2",
332
- "DECISION",
333
- "strategy_selected",
334
- ),
335
- ]);
336
-
337
- // Call with NO options except sessionDir - should use defaults
338
- const sessions = await loadCapturedSessions({
339
- sessionDir: tempSessionDir,
340
- });
341
-
342
- expect(sessions.length).toBe(1);
343
- expect(sessions[0]?.session.session_id).toBe("session-meets-defaults");
344
- });
345
- });
@@ -1,281 +0,0 @@
1
- /**
2
- * PGlite-backed eval data loader
3
- *
4
- * Loads real decomposition outcomes from the eval_records table
5
- * for use in Evalite evals.
6
- */
7
- import * as fs from "node:fs";
8
- import {
9
- getEvalRecords,
10
- getEvalStats,
11
- type EvalRecord,
12
- } from "swarm-mail";
13
-
14
- export interface EvalCase {
15
- input: { task: string; context?: string };
16
- expected: {
17
- minSubtasks: number;
18
- maxSubtasks: number;
19
- requiredFiles?: string[];
20
- overallSuccess?: boolean;
21
- };
22
- actual?: EvalRecord;
23
- }
24
-
25
- /**
26
- * Load eval cases from PGlite
27
- *
28
- * @param projectKey - Project key for filtering records
29
- * @param options - Filter options
30
- * @returns Array of eval cases ready for Evalite
31
- */
32
- export async function loadEvalCases(
33
- projectKey: string,
34
- options?: {
35
- limit?: number;
36
- strategy?: "file-based" | "feature-based" | "risk-based";
37
- successOnly?: boolean;
38
- projectPath?: string;
39
- },
40
- ): Promise<EvalCase[]> {
41
- const { limit, strategy, successOnly, projectPath } = options ?? {};
42
-
43
- // Query eval records from PGlite
44
- const records = await getEvalRecords(
45
- projectKey,
46
- { limit, strategy },
47
- projectPath,
48
- );
49
-
50
- // Filter by success if requested
51
- const filtered = successOnly
52
- ? records.filter((r) => r.overall_success === true)
53
- : records;
54
-
55
- // Transform to EvalCase format
56
- return filtered.map((record) => ({
57
- input: {
58
- task: record.task,
59
- context: record.context ?? undefined,
60
- },
61
- expected: {
62
- minSubtasks: 2,
63
- maxSubtasks: record.subtasks.length,
64
- requiredFiles: record.subtasks.flatMap((s) => s.files),
65
- overallSuccess: record.overall_success ?? undefined,
66
- },
67
- actual: record,
68
- }));
69
- }
70
-
71
- /**
72
- * Check if we have enough real data to run evals
73
- *
74
- * @param projectKey - Project key to check
75
- * @param minRecords - Minimum number of records required (default: 5)
76
- * @param projectPath - Optional project path for database lookup
77
- * @returns True if enough data exists
78
- */
79
- export async function hasRealEvalData(
80
- projectKey: string,
81
- minRecords: number = 5,
82
- projectPath?: string,
83
- ): Promise<boolean> {
84
- const stats = await getEvalStats(projectKey, projectPath);
85
- return stats.totalRecords >= minRecords;
86
- }
87
-
88
- /**
89
- * Get eval data stats for reporting
90
- *
91
- * @param projectKey - Project key to query
92
- * @param projectPath - Optional project path for database lookup
93
- * @returns Summary of available eval data
94
- */
95
- export async function getEvalDataSummary(
96
- projectKey: string,
97
- projectPath?: string,
98
- ): Promise<{
99
- totalRecords: number;
100
- successRate: number;
101
- byStrategy: Record<string, number>;
102
- hasEnoughData: boolean;
103
- }> {
104
- const stats = await getEvalStats(projectKey, projectPath);
105
-
106
- return {
107
- totalRecords: stats.totalRecords,
108
- successRate: stats.successRate,
109
- byStrategy: stats.byStrategy,
110
- hasEnoughData: stats.totalRecords >= 5,
111
- };
112
- }
113
-
114
- /**
115
- * Check if a session meets quality criteria
116
- */
117
- function meetsQualityCriteria(
118
- session: import("../../src/eval-capture.js").CoordinatorSession,
119
- criteria: {
120
- minEvents: number;
121
- requireWorkerSpawn: boolean;
122
- requireReview: boolean;
123
- },
124
- ): boolean {
125
- // Filter 1: minEvents
126
- if (session.events.length < criteria.minEvents) {
127
- return false;
128
- }
129
-
130
- // Filter 2: requireWorkerSpawn
131
- if (
132
- criteria.requireWorkerSpawn &&
133
- !session.events.some(
134
- (e) => e.event_type === "DECISION" && e.decision_type === "worker_spawned",
135
- )
136
- ) {
137
- return false;
138
- }
139
-
140
- // Filter 3: requireReview
141
- if (
142
- criteria.requireReview &&
143
- !session.events.some(
144
- (e) =>
145
- e.event_type === "DECISION" && e.decision_type === "review_completed",
146
- )
147
- ) {
148
- return false;
149
- }
150
-
151
- return true;
152
- }
153
-
154
- /**
155
- * Load captured coordinator sessions from ~/.config/swarm-tools/sessions/
156
- *
157
- * Reads all JSONL session files and returns CoordinatorSession objects.
158
- *
159
- * Quality filters are applied to focus on high-signal coordinator sessions:
160
- * - minEvents: Filter out incomplete/aborted sessions (default: 3)
161
- * - requireWorkerSpawn: Ensure session delegated to workers (default: true)
162
- * - requireReview: Ensure coordinator reviewed work (default: true)
163
- *
164
- * Filters are applied BEFORE the limit for accurate sampling.
165
- *
166
- * @param options - Filter options
167
- * @returns Array of coordinator sessions that meet quality criteria
168
- */
169
- export async function loadCapturedSessions(options?: {
170
- sessionIds?: string[];
171
- limit?: number;
172
- /** Minimum number of events required (default: 3) */
173
- minEvents?: number;
174
- /** Require at least one worker_spawned event (default: true) */
175
- requireWorkerSpawn?: boolean;
176
- /** Require at least one review_completed event (default: true) */
177
- requireReview?: boolean;
178
- /** Override session directory for testing */
179
- sessionDir?: string;
180
- }): Promise<
181
- Array<{ session: import("../../src/eval-capture.js").CoordinatorSession }>
182
- > {
183
- const { getSessionDir, readSessionEvents, saveSession } = await import(
184
- "../../src/eval-capture.js"
185
- );
186
- const sessionDir = options?.sessionDir ?? getSessionDir();
187
-
188
- // Default quality filters
189
- const qualityCriteria = {
190
- minEvents: options?.minEvents ?? 3,
191
- requireWorkerSpawn: options?.requireWorkerSpawn ?? true,
192
- requireReview: options?.requireReview ?? true,
193
- };
194
-
195
- // If session dir doesn't exist, return empty
196
- if (!fs.existsSync(sessionDir)) {
197
- return [];
198
- }
199
-
200
- // Read all .jsonl files in session directory
201
- const files = fs
202
- .readdirSync(sessionDir)
203
- .filter((f) => f.endsWith(".jsonl"));
204
-
205
- // Filter by sessionIds if provided
206
- const targetFiles = options?.sessionIds
207
- ? files.filter((f) => options.sessionIds?.includes(f.replace(".jsonl", "")))
208
- : files;
209
-
210
- // Load each session
211
- const sessions: Array<{
212
- session: import("../../src/eval-capture.js").CoordinatorSession;
213
- }> = [];
214
- let filteredOutCount = 0;
215
-
216
- for (const file of targetFiles) {
217
- const sessionId = file.replace(".jsonl", "");
218
-
219
- try {
220
- let events: import("../../src/eval-capture.js").CoordinatorEvent[];
221
-
222
- // If custom sessionDir, read directly; otherwise use eval-capture functions
223
- if (options?.sessionDir) {
224
- const sessionPath = `${sessionDir}/${sessionId}.jsonl`;
225
- if (!fs.existsSync(sessionPath)) continue;
226
-
227
- const content = fs.readFileSync(sessionPath, "utf-8");
228
- const lines = content.trim().split("\n").filter(Boolean);
229
- const { CoordinatorEventSchema } = await import(
230
- "../../src/eval-capture.js"
231
- );
232
- events = lines.map((line) => {
233
- const parsed = JSON.parse(line);
234
- return CoordinatorEventSchema.parse(parsed);
235
- });
236
- } else {
237
- events = readSessionEvents(sessionId);
238
- }
239
-
240
- if (events.length === 0) continue;
241
-
242
- // Find epic_id from first event
243
- const epicId = events[0]?.epic_id;
244
- if (!epicId) continue;
245
-
246
- // Build session object
247
- const session: import("../../src/eval-capture.js").CoordinatorSession = {
248
- session_id: sessionId,
249
- epic_id: epicId,
250
- start_time: events[0]?.timestamp ?? new Date().toISOString(),
251
- end_time: events[events.length - 1]?.timestamp,
252
- events,
253
- };
254
- if (!session) continue;
255
-
256
- // Apply quality filters BEFORE limit
257
- if (meetsQualityCriteria(session, qualityCriteria)) {
258
- sessions.push({ session });
259
- } else {
260
- filteredOutCount++;
261
- }
262
- } catch (error) {
263
- // Skip invalid sessions
264
- console.warn(`Failed to load session ${sessionId}:`, error);
265
- }
266
-
267
- // Apply limit AFTER filtering
268
- if (options?.limit && sessions.length >= options.limit) {
269
- break;
270
- }
271
- }
272
-
273
- // Log filtering stats for visibility
274
- if (filteredOutCount > 0) {
275
- console.log(
276
- `Filtered out ${filteredOutCount} sessions (minEvents=${qualityCriteria.minEvents}, requireWorkerSpawn=${qualityCriteria.requireWorkerSpawn}, requireReview=${qualityCriteria.requireReview})`,
277
- );
278
- }
279
-
280
- return sessions;
281
- }