@openpond/evals 0.5.0 → 0.7.0

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 (118) hide show
  1. package/CONTRACT.md +13 -1
  2. package/LEARNING.md +106 -0
  3. package/README.md +28 -2
  4. package/conformance/telemetry/v1/invalid-batch.json +17 -0
  5. package/conformance/telemetry/v1/valid-batch.json +58 -0
  6. package/dist/graders.js +26 -5
  7. package/dist/index.js +8 -0
  8. package/dist/javascript-verifier-contract.js +7 -0
  9. package/dist/javascript-verifier-node.js +56 -0
  10. package/dist/javascript-verifier-worker-source.js +1 -0
  11. package/dist/javascript-verifier-worker.js +6 -0
  12. package/dist/javascript-verifier.js +64 -0
  13. package/dist/learned-preference.js +334 -0
  14. package/dist/learning/admission.js +209 -0
  15. package/dist/learning/assets.js +37 -0
  16. package/dist/learning/contracts.js +267 -0
  17. package/dist/learning/errors.js +12 -0
  18. package/dist/learning/grade-worker.js +99 -0
  19. package/dist/learning/index.js +9 -0
  20. package/dist/learning/operations.js +33 -0
  21. package/dist/learning/repository.js +48 -0
  22. package/dist/learning/service.js +367 -0
  23. package/dist/learning/transport.js +25 -0
  24. package/dist/preferences.js +77 -11
  25. package/dist/rewards.js +230 -0
  26. package/dist/task-schema.js +142 -0
  27. package/dist/tasksets.js +3 -0
  28. package/dist/telemetry/index.js +4 -0
  29. package/dist/telemetry-analysis.js +183 -0
  30. package/dist/telemetry-bundle.js +60 -0
  31. package/dist/telemetry-catalog.js +50 -0
  32. package/dist/telemetry.js +112 -0
  33. package/dist/types/builtin-benchmarks/harness-refiner.d.ts +7 -0
  34. package/dist/types/builtin-benchmarks/harness-refiner.d.ts.map +1 -1
  35. package/dist/types/conformance.d.ts +14 -0
  36. package/dist/types/conformance.d.ts.map +1 -1
  37. package/dist/types/execution-contracts.d.ts +14 -0
  38. package/dist/types/execution-contracts.d.ts.map +1 -1
  39. package/dist/types/graders.d.ts.map +1 -1
  40. package/dist/types/index.d.ts +8 -0
  41. package/dist/types/index.d.ts.map +1 -1
  42. package/dist/types/javascript-verifier-contract.d.ts +9 -0
  43. package/dist/types/javascript-verifier-contract.d.ts.map +1 -0
  44. package/dist/types/javascript-verifier-node.d.ts +5 -0
  45. package/dist/types/javascript-verifier-node.d.ts.map +1 -0
  46. package/dist/types/javascript-verifier-worker-source.d.ts +2 -0
  47. package/dist/types/javascript-verifier-worker.d.ts +2 -0
  48. package/dist/types/javascript-verifier-worker.d.ts.map +1 -0
  49. package/dist/types/javascript-verifier.d.ts +15 -0
  50. package/dist/types/javascript-verifier.d.ts.map +1 -0
  51. package/dist/types/learned-preference.d.ts +282 -0
  52. package/dist/types/learned-preference.d.ts.map +1 -0
  53. package/dist/types/learning/admission.d.ts +335 -0
  54. package/dist/types/learning/admission.d.ts.map +1 -0
  55. package/dist/types/learning/assets.d.ts +49 -0
  56. package/dist/types/learning/assets.d.ts.map +1 -0
  57. package/dist/types/learning/contracts.d.ts +2149 -0
  58. package/dist/types/learning/contracts.d.ts.map +1 -0
  59. package/dist/types/learning/errors.d.ts +8 -0
  60. package/dist/types/learning/errors.d.ts.map +1 -0
  61. package/dist/types/learning/grade-worker.d.ts +31 -0
  62. package/dist/types/learning/grade-worker.d.ts.map +1 -0
  63. package/dist/types/learning/index.d.ts +10 -0
  64. package/dist/types/learning/index.d.ts.map +1 -0
  65. package/dist/types/learning/operations.d.ts +2061 -0
  66. package/dist/types/learning/operations.d.ts.map +1 -0
  67. package/dist/types/learning/repository.d.ts +1143 -0
  68. package/dist/types/learning/repository.d.ts.map +1 -0
  69. package/dist/types/learning/service.d.ts +22 -0
  70. package/dist/types/learning/service.d.ts.map +1 -0
  71. package/dist/types/learning/transport.d.ts +2139 -0
  72. package/dist/types/learning/transport.d.ts.map +1 -0
  73. package/dist/types/preferences.d.ts +55 -7
  74. package/dist/types/preferences.d.ts.map +1 -1
  75. package/dist/types/review-conformance.d.ts +5 -5
  76. package/dist/types/rewards.d.ts +611 -0
  77. package/dist/types/rewards.d.ts.map +1 -0
  78. package/dist/types/task-schema.d.ts +18 -0
  79. package/dist/types/task-schema.d.ts.map +1 -0
  80. package/dist/types/tasksets.d.ts +28 -0
  81. package/dist/types/tasksets.d.ts.map +1 -1
  82. package/dist/types/telemetry/index.d.ts +5 -0
  83. package/dist/types/telemetry/index.d.ts.map +1 -0
  84. package/dist/types/telemetry-analysis.d.ts +116 -0
  85. package/dist/types/telemetry-analysis.d.ts.map +1 -0
  86. package/dist/types/telemetry-bundle.d.ts +309 -0
  87. package/dist/types/telemetry-bundle.d.ts.map +1 -0
  88. package/dist/types/telemetry-catalog.d.ts +269 -0
  89. package/dist/types/telemetry-catalog.d.ts.map +1 -0
  90. package/dist/types/telemetry.d.ts +266 -0
  91. package/dist/types/telemetry.d.ts.map +1 -0
  92. package/package.json +40 -3
  93. package/schemas/learning/v1/asset.schema.json +89 -0
  94. package/schemas/learning/v1/batch.schema.json +132 -0
  95. package/schemas/learning/v1/binding.schema.json +207 -0
  96. package/schemas/learning/v1/command-request.schema.json +2414 -0
  97. package/schemas/learning/v1/decision.schema.json +440 -0
  98. package/schemas/learning/v1/definition.schema.json +397 -0
  99. package/schemas/learning/v1/evidence.schema.json +317 -0
  100. package/schemas/learning/v1/example-submission.schema.json +253 -0
  101. package/schemas/learning/v1/feedback-submission.schema.json +114 -0
  102. package/schemas/learning/v1/feedback.schema.json +233 -0
  103. package/schemas/learning/v1/grade.schema.json +412 -0
  104. package/schemas/learning/v1/iteration.schema.json +236 -0
  105. package/schemas/learning/v1/package.schema.json +776 -0
  106. package/schemas/learning/v1/policy.schema.json +373 -0
  107. package/schemas/learning/v1/read-request.schema.json +101 -0
  108. package/schemas/learning/v1/reward.schema.json +302 -0
  109. package/schemas/learning/v1/source.schema.json +198 -0
  110. package/schemas/telemetry/v1/evidence-completeness.schema.json +82 -0
  111. package/schemas/telemetry/v1/evidence-reference.schema.json +41 -0
  112. package/schemas/telemetry/v1/metric-definition.schema.json +96 -0
  113. package/schemas/telemetry/v1/metric-observation.schema.json +182 -0
  114. package/schemas/telemetry/v1/run-metric-summary.schema.json +98 -0
  115. package/schemas/telemetry/v1/run-telemetry-batch.schema.json +405 -0
  116. package/schemas/telemetry/v1/run-telemetry-event.schema.json +201 -0
  117. package/schemas/telemetry/v1/telemetry-cohort.schema.json +100 -0
  118. package/schemas/telemetry/v1/telemetry-export-bundle.schema.json +655 -0
@@ -0,0 +1,302 @@
1
+ {
2
+ "$schema": "https://json-schema.org/draft/2020-12/schema",
3
+ "type": "object",
4
+ "properties": {
5
+ "schemaVersion": {
6
+ "type": "string",
7
+ "const": "openpond.rewardRelease.v1"
8
+ },
9
+ "id": {
10
+ "$ref": "#/$defs/__schema0"
11
+ },
12
+ "revision": {
13
+ "type": "integer",
14
+ "exclusiveMinimum": 0,
15
+ "maximum": 9007199254740991
16
+ },
17
+ "name": {
18
+ "type": "string",
19
+ "minLength": 1,
20
+ "maxLength": 500
21
+ },
22
+ "description": {
23
+ "type": "string",
24
+ "maxLength": 10000
25
+ },
26
+ "implementation": {
27
+ "anyOf": [
28
+ {
29
+ "type": "object",
30
+ "properties": {
31
+ "kind": {
32
+ "type": "string",
33
+ "enum": [
34
+ "content",
35
+ "schema",
36
+ "artifact",
37
+ "runtime_event",
38
+ "state"
39
+ ]
40
+ },
41
+ "config": {
42
+ "type": "object",
43
+ "propertyNames": {
44
+ "type": "string"
45
+ },
46
+ "additionalProperties": {}
47
+ }
48
+ },
49
+ "required": [
50
+ "kind",
51
+ "config"
52
+ ],
53
+ "additionalProperties": false
54
+ },
55
+ {
56
+ "type": "object",
57
+ "properties": {
58
+ "kind": {
59
+ "type": "string",
60
+ "const": "model_judge"
61
+ },
62
+ "rubricRef": {
63
+ "$ref": "#/$defs/__schema1"
64
+ },
65
+ "calibrationStatus": {
66
+ "type": "string",
67
+ "enum": [
68
+ "pending",
69
+ "passed",
70
+ "failed"
71
+ ]
72
+ },
73
+ "model": {
74
+ "type": "object",
75
+ "properties": {
76
+ "providerId": {
77
+ "$ref": "#/$defs/__schema0"
78
+ },
79
+ "modelId": {
80
+ "$ref": "#/$defs/__schema0"
81
+ },
82
+ "revision": {
83
+ "anyOf": [
84
+ {
85
+ "type": "string",
86
+ "minLength": 1,
87
+ "maxLength": 500
88
+ },
89
+ {
90
+ "type": "null"
91
+ }
92
+ ]
93
+ }
94
+ },
95
+ "required": [
96
+ "providerId",
97
+ "modelId",
98
+ "revision"
99
+ ],
100
+ "additionalProperties": false
101
+ },
102
+ "temperature": {
103
+ "type": "number",
104
+ "minimum": 0,
105
+ "maximum": 2
106
+ }
107
+ },
108
+ "required": [
109
+ "kind",
110
+ "rubricRef",
111
+ "calibrationStatus"
112
+ ],
113
+ "additionalProperties": false
114
+ },
115
+ {
116
+ "type": "object",
117
+ "properties": {
118
+ "kind": {
119
+ "type": "string",
120
+ "const": "custom_verifier"
121
+ },
122
+ "verifierRef": {
123
+ "$ref": "#/$defs/__schema1"
124
+ },
125
+ "exportName": {
126
+ "type": "string",
127
+ "minLength": 1,
128
+ "maxLength": 240
129
+ },
130
+ "timeoutMs": {
131
+ "type": "integer",
132
+ "exclusiveMinimum": 0,
133
+ "maximum": 300000
134
+ },
135
+ "networkPolicy": {
136
+ "type": "string",
137
+ "const": "none"
138
+ }
139
+ },
140
+ "required": [
141
+ "kind",
142
+ "verifierRef",
143
+ "timeoutMs",
144
+ "networkPolicy"
145
+ ],
146
+ "additionalProperties": false
147
+ },
148
+ {
149
+ "type": "object",
150
+ "properties": {
151
+ "kind": {
152
+ "type": "string",
153
+ "const": "human"
154
+ },
155
+ "rubricRef": {
156
+ "$ref": "#/$defs/__schema1"
157
+ },
158
+ "reviewerRole": {
159
+ "type": "string",
160
+ "minLength": 1,
161
+ "maxLength": 500
162
+ }
163
+ },
164
+ "required": [
165
+ "kind",
166
+ "rubricRef",
167
+ "reviewerRole"
168
+ ],
169
+ "additionalProperties": false
170
+ },
171
+ {
172
+ "type": "object",
173
+ "properties": {
174
+ "kind": {
175
+ "type": "string",
176
+ "const": "learned_model"
177
+ },
178
+ "modelVersion": {
179
+ "type": "object",
180
+ "properties": {
181
+ "id": {
182
+ "$ref": "#/$defs/__schema0"
183
+ },
184
+ "contentHash": {
185
+ "$ref": "#/$defs/__schema2"
186
+ }
187
+ },
188
+ "required": [
189
+ "id",
190
+ "contentHash"
191
+ ],
192
+ "additionalProperties": false
193
+ },
194
+ "inputContract": {
195
+ "$ref": "#/$defs/__schema1"
196
+ }
197
+ },
198
+ "required": [
199
+ "kind",
200
+ "modelVersion",
201
+ "inputContract"
202
+ ],
203
+ "additionalProperties": false
204
+ }
205
+ ]
206
+ },
207
+ "rawScore": {
208
+ "type": "object",
209
+ "properties": {
210
+ "minimum": {
211
+ "type": "number"
212
+ },
213
+ "maximum": {
214
+ "type": "number"
215
+ }
216
+ },
217
+ "required": [
218
+ "minimum",
219
+ "maximum"
220
+ ],
221
+ "additionalProperties": false
222
+ },
223
+ "assets": {
224
+ "maxItems": 1000,
225
+ "type": "array",
226
+ "items": {
227
+ "$ref": "#/$defs/__schema1"
228
+ }
229
+ },
230
+ "contentHash": {
231
+ "$ref": "#/$defs/__schema2"
232
+ }
233
+ },
234
+ "required": [
235
+ "schemaVersion",
236
+ "id",
237
+ "revision",
238
+ "name",
239
+ "description",
240
+ "implementation",
241
+ "rawScore",
242
+ "assets",
243
+ "contentHash"
244
+ ],
245
+ "additionalProperties": false,
246
+ "$defs": {
247
+ "__schema0": {
248
+ "type": "string",
249
+ "minLength": 1,
250
+ "maxLength": 240
251
+ },
252
+ "__schema1": {
253
+ "type": "object",
254
+ "properties": {
255
+ "id": {
256
+ "$ref": "#/$defs/__schema0"
257
+ },
258
+ "path": {
259
+ "type": "string",
260
+ "minLength": 1,
261
+ "maxLength": 2000
262
+ },
263
+ "contentHash": {
264
+ "$ref": "#/$defs/__schema2"
265
+ },
266
+ "sizeBytes": {
267
+ "type": "integer",
268
+ "minimum": 0,
269
+ "maximum": 250000000
270
+ },
271
+ "mediaType": {
272
+ "type": "string",
273
+ "minLength": 1,
274
+ "maxLength": 200
275
+ },
276
+ "visibility": {
277
+ "type": "string",
278
+ "enum": [
279
+ "policy",
280
+ "verifier",
281
+ "host_private"
282
+ ]
283
+ }
284
+ },
285
+ "required": [
286
+ "id",
287
+ "path",
288
+ "contentHash",
289
+ "sizeBytes",
290
+ "mediaType",
291
+ "visibility"
292
+ ],
293
+ "additionalProperties": false
294
+ },
295
+ "__schema2": {
296
+ "type": "string",
297
+ "pattern": "^[a-f0-9]{64}$"
298
+ }
299
+ },
300
+ "$id": "urn:openpond:learning:v1:reward",
301
+ "$comment": "Generated from @openpond/evals. Cross-resource identity, authorization, idempotency, admission, grading and split constraints are enforced by the shared domain service in addition to this structural schema."
302
+ }
@@ -0,0 +1,198 @@
1
+ {
2
+ "$schema": "https://json-schema.org/draft/2020-12/schema",
3
+ "type": "object",
4
+ "properties": {
5
+ "schemaVersion": {
6
+ "type": "string",
7
+ "const": "openpond.learningSource.v1"
8
+ },
9
+ "id": {
10
+ "$ref": "#/$defs/__schema0"
11
+ },
12
+ "revision": {
13
+ "type": "integer",
14
+ "exclusiveMinimum": 0,
15
+ "maximum": 9007199254740991
16
+ },
17
+ "name": {
18
+ "type": "string",
19
+ "minLength": 1,
20
+ "maxLength": 500
21
+ },
22
+ "kind": {
23
+ "type": "string",
24
+ "enum": [
25
+ "direct",
26
+ "benchmark",
27
+ "work",
28
+ "opentelemetry",
29
+ "provider"
30
+ ]
31
+ },
32
+ "taskDefinition": {
33
+ "type": "object",
34
+ "properties": {
35
+ "id": {
36
+ "$ref": "#/$defs/__schema0"
37
+ },
38
+ "contentHash": {
39
+ "$ref": "#/$defs/__schema1"
40
+ },
41
+ "revision": {
42
+ "type": "integer",
43
+ "exclusiveMinimum": 0,
44
+ "maximum": 9007199254740991
45
+ }
46
+ },
47
+ "required": [
48
+ "id",
49
+ "contentHash",
50
+ "revision"
51
+ ],
52
+ "additionalProperties": false
53
+ },
54
+ "enabled": {
55
+ "type": "boolean"
56
+ },
57
+ "allowedSplits": {
58
+ "minItems": 1,
59
+ "maxItems": 4,
60
+ "type": "array",
61
+ "items": {
62
+ "$ref": "#/$defs/__schema2"
63
+ }
64
+ },
65
+ "mapping": {
66
+ "anyOf": [
67
+ {
68
+ "type": "object",
69
+ "properties": {
70
+ "schemaVersion": {
71
+ "type": "string",
72
+ "const": "openpond.taskSourceMapping.v1"
73
+ },
74
+ "exampleId": {
75
+ "$ref": "#/$defs/__schema3"
76
+ },
77
+ "attemptId": {
78
+ "$ref": "#/$defs/__schema3"
79
+ },
80
+ "occurredAt": {
81
+ "$ref": "#/$defs/__schema3"
82
+ },
83
+ "input": {
84
+ "$ref": "#/$defs/__schema3"
85
+ },
86
+ "observedOutput": {
87
+ "anyOf": [
88
+ {
89
+ "$ref": "#/$defs/__schema3"
90
+ },
91
+ {
92
+ "type": "null"
93
+ }
94
+ ]
95
+ },
96
+ "expected": {
97
+ "anyOf": [
98
+ {
99
+ "$ref": "#/$defs/__schema3"
100
+ },
101
+ {
102
+ "type": "null"
103
+ }
104
+ ]
105
+ },
106
+ "evaluatorContext": {
107
+ "anyOf": [
108
+ {
109
+ "$ref": "#/$defs/__schema3"
110
+ },
111
+ {
112
+ "type": "null"
113
+ }
114
+ ]
115
+ },
116
+ "familyKey": {
117
+ "$ref": "#/$defs/__schema3"
118
+ },
119
+ "split": {
120
+ "$ref": "#/$defs/__schema2"
121
+ }
122
+ },
123
+ "required": [
124
+ "schemaVersion",
125
+ "exampleId",
126
+ "attemptId",
127
+ "occurredAt",
128
+ "input",
129
+ "observedOutput",
130
+ "expected",
131
+ "evaluatorContext",
132
+ "familyKey",
133
+ "split"
134
+ ],
135
+ "additionalProperties": false
136
+ },
137
+ {
138
+ "type": "null"
139
+ }
140
+ ]
141
+ },
142
+ "adapterVersion": {
143
+ "anyOf": [
144
+ {
145
+ "type": "string",
146
+ "minLength": 1,
147
+ "maxLength": 200
148
+ },
149
+ {
150
+ "type": "null"
151
+ }
152
+ ]
153
+ },
154
+ "contentHash": {
155
+ "$ref": "#/$defs/__schema1"
156
+ }
157
+ },
158
+ "required": [
159
+ "schemaVersion",
160
+ "id",
161
+ "revision",
162
+ "name",
163
+ "kind",
164
+ "taskDefinition",
165
+ "enabled",
166
+ "allowedSplits",
167
+ "mapping",
168
+ "adapterVersion",
169
+ "contentHash"
170
+ ],
171
+ "additionalProperties": false,
172
+ "$defs": {
173
+ "__schema0": {
174
+ "type": "string",
175
+ "minLength": 1,
176
+ "maxLength": 240
177
+ },
178
+ "__schema1": {
179
+ "type": "string",
180
+ "pattern": "^[a-f0-9]{64}$"
181
+ },
182
+ "__schema2": {
183
+ "type": "string",
184
+ "enum": [
185
+ "train",
186
+ "validation",
187
+ "test",
188
+ "frozen_eval"
189
+ ]
190
+ },
191
+ "__schema3": {
192
+ "type": "string",
193
+ "maxLength": 2000
194
+ }
195
+ },
196
+ "$id": "urn:openpond:learning:v1:source",
197
+ "$comment": "Generated from @openpond/evals. Cross-resource identity, authorization, idempotency, admission, grading and split constraints are enforced by the shared domain service in addition to this structural schema."
198
+ }
@@ -0,0 +1,82 @@
1
+ {
2
+ "$schema": "http://json-schema.org/draft-07/schema#",
3
+ "type": "object",
4
+ "properties": {
5
+ "schemaVersion": {
6
+ "type": "string",
7
+ "const": "openpond.telemetryEvidenceCompleteness.v1"
8
+ },
9
+ "runId": {
10
+ "type": "string",
11
+ "minLength": 1,
12
+ "maxLength": 240
13
+ },
14
+ "status": {
15
+ "type": "string",
16
+ "enum": [
17
+ "complete",
18
+ "partial",
19
+ "missing"
20
+ ]
21
+ },
22
+ "expectedEventTypes": {
23
+ "maxItems": 100,
24
+ "type": "array",
25
+ "items": {
26
+ "type": "string",
27
+ "minLength": 1,
28
+ "maxLength": 100
29
+ }
30
+ },
31
+ "observedEventTypes": {
32
+ "maxItems": 100,
33
+ "type": "array",
34
+ "items": {
35
+ "type": "string",
36
+ "minLength": 1,
37
+ "maxLength": 100
38
+ }
39
+ },
40
+ "missingEventTypes": {
41
+ "maxItems": 100,
42
+ "type": "array",
43
+ "items": {
44
+ "type": "string",
45
+ "minLength": 1,
46
+ "maxLength": 100
47
+ }
48
+ },
49
+ "lastSequence": {
50
+ "anyOf": [
51
+ {
52
+ "type": "integer",
53
+ "minimum": 0,
54
+ "maximum": 9007199254740991
55
+ },
56
+ {
57
+ "type": "null"
58
+ }
59
+ ]
60
+ },
61
+ "sequenceGaps": {
62
+ "maxItems": 10000,
63
+ "type": "array",
64
+ "items": {
65
+ "type": "integer",
66
+ "minimum": 0,
67
+ "maximum": 9007199254740991
68
+ }
69
+ }
70
+ },
71
+ "required": [
72
+ "schemaVersion",
73
+ "runId",
74
+ "status",
75
+ "expectedEventTypes",
76
+ "observedEventTypes",
77
+ "missingEventTypes",
78
+ "lastSequence",
79
+ "sequenceGaps"
80
+ ],
81
+ "additionalProperties": false
82
+ }
@@ -0,0 +1,41 @@
1
+ {
2
+ "$schema": "http://json-schema.org/draft-07/schema#",
3
+ "type": "object",
4
+ "properties": {
5
+ "id": {
6
+ "type": "string",
7
+ "minLength": 1,
8
+ "maxLength": 240
9
+ },
10
+ "contentHash": {
11
+ "type": "string",
12
+ "pattern": "^[a-f0-9]{64}$"
13
+ },
14
+ "kind": {
15
+ "type": "string",
16
+ "enum": [
17
+ "rollout",
18
+ "attempt",
19
+ "trace",
20
+ "grader",
21
+ "checkpoint",
22
+ "artifact"
23
+ ]
24
+ },
25
+ "visibility": {
26
+ "type": "string",
27
+ "enum": [
28
+ "policy_visible",
29
+ "team_visible",
30
+ "host_private"
31
+ ]
32
+ }
33
+ },
34
+ "required": [
35
+ "id",
36
+ "contentHash",
37
+ "kind",
38
+ "visibility"
39
+ ],
40
+ "additionalProperties": false
41
+ }