@openpond/evals 0.10.9 → 0.10.10

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 (74) hide show
  1. package/LEARNING.md +15 -0
  2. package/dist/graders.js +19 -4
  3. package/dist/learning/budgeted-reward-fixtures.js +34 -0
  4. package/dist/learning/contracts.js +3 -0
  5. package/dist/learning/grade-judge-budget.js +26 -0
  6. package/dist/learning/index.js +6 -0
  7. package/dist/learning/intake-history.js +2 -1
  8. package/dist/learning/intake-text.js +14 -0
  9. package/dist/learning/intake.js +5 -2
  10. package/dist/learning/judge-budget.js +59 -0
  11. package/dist/learning/model-judge.js +49 -0
  12. package/dist/learning/operations.js +3 -1
  13. package/dist/learning/reward-authoring.js +4 -0
  14. package/dist/learning/reward-calibration.js +66 -0
  15. package/dist/learning/reward-check-budget.js +27 -0
  16. package/dist/learning/reward-check-service.js +2 -2
  17. package/dist/learning/reward-check-worker.js +10 -2
  18. package/dist/learning/reward-checks.js +4 -1
  19. package/dist/learning/reward-fixture-executor.js +3 -2
  20. package/dist/learning/service.js +30 -5
  21. package/dist/rewards.js +15 -3
  22. package/dist/rollouts.js +2 -2
  23. package/dist/types/graders.d.ts +51 -0
  24. package/dist/types/graders.d.ts.map +1 -1
  25. package/dist/types/learning/admission.d.ts +5 -0
  26. package/dist/types/learning/admission.d.ts.map +1 -1
  27. package/dist/types/learning/budgeted-reward-fixtures.d.ts +27 -0
  28. package/dist/types/learning/budgeted-reward-fixtures.d.ts.map +1 -0
  29. package/dist/types/learning/contracts.d.ts +301 -0
  30. package/dist/types/learning/contracts.d.ts.map +1 -1
  31. package/dist/types/learning/grade-judge-budget.d.ts +10 -0
  32. package/dist/types/learning/grade-judge-budget.d.ts.map +1 -0
  33. package/dist/types/learning/index.d.ts +6 -0
  34. package/dist/types/learning/index.d.ts.map +1 -1
  35. package/dist/types/learning/intake-history.d.ts.map +1 -1
  36. package/dist/types/learning/intake-text.d.ts +6 -0
  37. package/dist/types/learning/intake-text.d.ts.map +1 -0
  38. package/dist/types/learning/intake.d.ts +1 -0
  39. package/dist/types/learning/intake.d.ts.map +1 -1
  40. package/dist/types/learning/iteration-dispatch-contracts.d.ts +1 -1
  41. package/dist/types/learning/iteration-reservation-contracts.d.ts +1 -1
  42. package/dist/types/learning/judge-budget.d.ts +43 -0
  43. package/dist/types/learning/judge-budget.d.ts.map +1 -0
  44. package/dist/types/learning/model-judge.d.ts +29 -0
  45. package/dist/types/learning/model-judge.d.ts.map +1 -0
  46. package/dist/types/learning/operations.d.ts +41 -0
  47. package/dist/types/learning/operations.d.ts.map +1 -1
  48. package/dist/types/learning/repository.d.ts +177 -4
  49. package/dist/types/learning/repository.d.ts.map +1 -1
  50. package/dist/types/learning/reward-authoring.d.ts.map +1 -1
  51. package/dist/types/learning/reward-calibration.d.ts +129 -0
  52. package/dist/types/learning/reward-calibration.d.ts.map +1 -0
  53. package/dist/types/learning/reward-check-budget.d.ts +12 -0
  54. package/dist/types/learning/reward-check-budget.d.ts.map +1 -0
  55. package/dist/types/learning/reward-check-service.d.ts +1 -1
  56. package/dist/types/learning/reward-check-service.d.ts.map +1 -1
  57. package/dist/types/learning/reward-check-worker.d.ts.map +1 -1
  58. package/dist/types/learning/reward-checks.d.ts +84 -0
  59. package/dist/types/learning/reward-checks.d.ts.map +1 -1
  60. package/dist/types/learning/reward-fixture-executor.d.ts +3 -0
  61. package/dist/types/learning/reward-fixture-executor.d.ts.map +1 -1
  62. package/dist/types/learning/schedule-contracts.d.ts +1 -1
  63. package/dist/types/learning/service.d.ts.map +1 -1
  64. package/dist/types/learning/transport.d.ts +198 -5
  65. package/dist/types/learning/transport.d.ts.map +1 -1
  66. package/dist/types/rewards.d.ts +107 -0
  67. package/dist/types/rewards.d.ts.map +1 -1
  68. package/dist/types/rollouts.d.ts.map +1 -1
  69. package/package.json +1 -1
  70. package/schemas/learning/v1/command-request.schema.json +225 -169
  71. package/schemas/learning/v1/decision.schema.json +192 -0
  72. package/schemas/learning/v1/grade.schema.json +335 -3
  73. package/schemas/learning/v1/reward.schema.json +22 -0
  74. package/schemas/learning/v1/reward_check.schema.json +335 -3
@@ -408,6 +408,194 @@
408
408
  "$ref": "#/$defs/__schema2"
409
409
  }
410
410
  },
411
+ "graderEvidence": {
412
+ "type": "object",
413
+ "properties": {
414
+ "schemaVersion": {
415
+ "type": "string",
416
+ "const": "openpond.graderEvidence.v1"
417
+ },
418
+ "graderId": {
419
+ "$ref": "#/$defs/__schema0"
420
+ },
421
+ "graderVersion": {
422
+ "type": "string",
423
+ "minLength": 1,
424
+ "maxLength": 100
425
+ },
426
+ "score": {
427
+ "anyOf": [
428
+ {
429
+ "type": "number",
430
+ "minimum": 0,
431
+ "maximum": 1
432
+ },
433
+ {
434
+ "type": "null"
435
+ }
436
+ ]
437
+ },
438
+ "passed": {
439
+ "type": "boolean"
440
+ },
441
+ "rewardEligible": {
442
+ "type": "boolean"
443
+ },
444
+ "failureClass": {
445
+ "anyOf": [
446
+ {
447
+ "type": "string",
448
+ "enum": [
449
+ "policy_failure",
450
+ "grader_failure",
451
+ "environment_failure",
452
+ "infrastructure_failure",
453
+ "timeout",
454
+ "cancelled"
455
+ ]
456
+ },
457
+ {
458
+ "type": "null"
459
+ }
460
+ ]
461
+ },
462
+ "feedback": {
463
+ "maxItems": 1000,
464
+ "type": "array",
465
+ "items": {
466
+ "$ref": "#/$defs/__schema6"
467
+ }
468
+ },
469
+ "visibleEvidenceRefs": {
470
+ "maxItems": 10000,
471
+ "type": "array",
472
+ "items": {
473
+ "$ref": "#/$defs/__schema0"
474
+ }
475
+ },
476
+ "privilegedEvidenceRefs": {
477
+ "maxItems": 10000,
478
+ "type": "array",
479
+ "items": {
480
+ "$ref": "#/$defs/__schema0"
481
+ }
482
+ },
483
+ "modelJudgeReceipt": {
484
+ "type": "object",
485
+ "properties": {
486
+ "schemaVersion": {
487
+ "type": "string",
488
+ "const": "openpond.modelJudgeReceipt.v1"
489
+ },
490
+ "providerId": {
491
+ "type": "string",
492
+ "minLength": 1,
493
+ "maxLength": 200
494
+ },
495
+ "modelId": {
496
+ "type": "string",
497
+ "minLength": 1,
498
+ "maxLength": 500
499
+ },
500
+ "modelRevision": {
501
+ "anyOf": [
502
+ {
503
+ "type": "string",
504
+ "minLength": 1,
505
+ "maxLength": 500
506
+ },
507
+ {
508
+ "type": "null"
509
+ }
510
+ ]
511
+ },
512
+ "responseId": {
513
+ "anyOf": [
514
+ {
515
+ "type": "string",
516
+ "minLength": 1,
517
+ "maxLength": 500
518
+ },
519
+ {
520
+ "type": "null"
521
+ }
522
+ ]
523
+ },
524
+ "requestHash": {
525
+ "$ref": "#/$defs/__schema2"
526
+ },
527
+ "responseHash": {
528
+ "$ref": "#/$defs/__schema2"
529
+ },
530
+ "inputTokens": {
531
+ "anyOf": [
532
+ {
533
+ "type": "integer",
534
+ "minimum": 0,
535
+ "maximum": 9007199254740991
536
+ },
537
+ {
538
+ "type": "null"
539
+ }
540
+ ]
541
+ },
542
+ "outputTokens": {
543
+ "anyOf": [
544
+ {
545
+ "type": "integer",
546
+ "minimum": 0,
547
+ "maximum": 9007199254740991
548
+ },
549
+ {
550
+ "type": "null"
551
+ }
552
+ ]
553
+ },
554
+ "costUsd": {
555
+ "anyOf": [
556
+ {
557
+ "type": "number",
558
+ "minimum": 0
559
+ },
560
+ {
561
+ "type": "null"
562
+ }
563
+ ]
564
+ }
565
+ },
566
+ "required": [
567
+ "schemaVersion",
568
+ "providerId",
569
+ "modelId",
570
+ "modelRevision",
571
+ "responseId",
572
+ "requestHash",
573
+ "responseHash",
574
+ "inputTokens",
575
+ "outputTokens",
576
+ "costUsd"
577
+ ],
578
+ "additionalProperties": false
579
+ },
580
+ "contentHash": {
581
+ "$ref": "#/$defs/__schema2"
582
+ }
583
+ },
584
+ "required": [
585
+ "schemaVersion",
586
+ "graderId",
587
+ "graderVersion",
588
+ "score",
589
+ "passed",
590
+ "rewardEligible",
591
+ "failureClass",
592
+ "feedback",
593
+ "visibleEvidenceRefs",
594
+ "privilegedEvidenceRefs",
595
+ "contentHash"
596
+ ],
597
+ "additionalProperties": false
598
+ },
411
599
  "message": {
412
600
  "anyOf": [
413
601
  {
@@ -432,6 +620,10 @@
432
620
  "message"
433
621
  ],
434
622
  "additionalProperties": false
623
+ },
624
+ "__schema6": {
625
+ "type": "string",
626
+ "maxLength": 20000
435
627
  }
436
628
  },
437
629
  "$id": "urn:openpond:learning:v1:decision",
@@ -188,7 +188,7 @@
188
188
  "leaseExpiresAt": {
189
189
  "anyOf": [
190
190
  {
191
- "$ref": "#/$defs/__schema5"
191
+ "$ref": "#/$defs/__schema6"
192
192
  },
193
193
  {
194
194
  "type": "null"
@@ -210,6 +210,16 @@
210
210
  "minimum": 0,
211
211
  "maximum": 1000
212
212
  },
213
+ "judgeCalls": {
214
+ "maxItems": 10000,
215
+ "type": "array",
216
+ "items": {
217
+ "$ref": "#/$defs/__schema7"
218
+ }
219
+ },
220
+ "requestedBy": {
221
+ "$ref": "#/$defs/__schema0"
222
+ },
213
223
  "failure": {
214
224
  "anyOf": [
215
225
  {
@@ -222,10 +232,10 @@
222
232
  ]
223
233
  },
224
234
  "createdAt": {
225
- "$ref": "#/$defs/__schema5"
235
+ "$ref": "#/$defs/__schema6"
226
236
  },
227
237
  "updatedAt": {
228
- "$ref": "#/$defs/__schema5"
238
+ "$ref": "#/$defs/__schema6"
229
239
  }
230
240
  },
231
241
  "required": [
@@ -375,6 +385,194 @@
375
385
  "$ref": "#/$defs/__schema2"
376
386
  }
377
387
  },
388
+ "graderEvidence": {
389
+ "type": "object",
390
+ "properties": {
391
+ "schemaVersion": {
392
+ "type": "string",
393
+ "const": "openpond.graderEvidence.v1"
394
+ },
395
+ "graderId": {
396
+ "$ref": "#/$defs/__schema0"
397
+ },
398
+ "graderVersion": {
399
+ "type": "string",
400
+ "minLength": 1,
401
+ "maxLength": 100
402
+ },
403
+ "score": {
404
+ "anyOf": [
405
+ {
406
+ "type": "number",
407
+ "minimum": 0,
408
+ "maximum": 1
409
+ },
410
+ {
411
+ "type": "null"
412
+ }
413
+ ]
414
+ },
415
+ "passed": {
416
+ "type": "boolean"
417
+ },
418
+ "rewardEligible": {
419
+ "type": "boolean"
420
+ },
421
+ "failureClass": {
422
+ "anyOf": [
423
+ {
424
+ "type": "string",
425
+ "enum": [
426
+ "policy_failure",
427
+ "grader_failure",
428
+ "environment_failure",
429
+ "infrastructure_failure",
430
+ "timeout",
431
+ "cancelled"
432
+ ]
433
+ },
434
+ {
435
+ "type": "null"
436
+ }
437
+ ]
438
+ },
439
+ "feedback": {
440
+ "maxItems": 1000,
441
+ "type": "array",
442
+ "items": {
443
+ "$ref": "#/$defs/__schema5"
444
+ }
445
+ },
446
+ "visibleEvidenceRefs": {
447
+ "maxItems": 10000,
448
+ "type": "array",
449
+ "items": {
450
+ "$ref": "#/$defs/__schema0"
451
+ }
452
+ },
453
+ "privilegedEvidenceRefs": {
454
+ "maxItems": 10000,
455
+ "type": "array",
456
+ "items": {
457
+ "$ref": "#/$defs/__schema0"
458
+ }
459
+ },
460
+ "modelJudgeReceipt": {
461
+ "type": "object",
462
+ "properties": {
463
+ "schemaVersion": {
464
+ "type": "string",
465
+ "const": "openpond.modelJudgeReceipt.v1"
466
+ },
467
+ "providerId": {
468
+ "type": "string",
469
+ "minLength": 1,
470
+ "maxLength": 200
471
+ },
472
+ "modelId": {
473
+ "type": "string",
474
+ "minLength": 1,
475
+ "maxLength": 500
476
+ },
477
+ "modelRevision": {
478
+ "anyOf": [
479
+ {
480
+ "type": "string",
481
+ "minLength": 1,
482
+ "maxLength": 500
483
+ },
484
+ {
485
+ "type": "null"
486
+ }
487
+ ]
488
+ },
489
+ "responseId": {
490
+ "anyOf": [
491
+ {
492
+ "type": "string",
493
+ "minLength": 1,
494
+ "maxLength": 500
495
+ },
496
+ {
497
+ "type": "null"
498
+ }
499
+ ]
500
+ },
501
+ "requestHash": {
502
+ "$ref": "#/$defs/__schema2"
503
+ },
504
+ "responseHash": {
505
+ "$ref": "#/$defs/__schema2"
506
+ },
507
+ "inputTokens": {
508
+ "anyOf": [
509
+ {
510
+ "type": "integer",
511
+ "minimum": 0,
512
+ "maximum": 9007199254740991
513
+ },
514
+ {
515
+ "type": "null"
516
+ }
517
+ ]
518
+ },
519
+ "outputTokens": {
520
+ "anyOf": [
521
+ {
522
+ "type": "integer",
523
+ "minimum": 0,
524
+ "maximum": 9007199254740991
525
+ },
526
+ {
527
+ "type": "null"
528
+ }
529
+ ]
530
+ },
531
+ "costUsd": {
532
+ "anyOf": [
533
+ {
534
+ "type": "number",
535
+ "minimum": 0
536
+ },
537
+ {
538
+ "type": "null"
539
+ }
540
+ ]
541
+ }
542
+ },
543
+ "required": [
544
+ "schemaVersion",
545
+ "providerId",
546
+ "modelId",
547
+ "modelRevision",
548
+ "responseId",
549
+ "requestHash",
550
+ "responseHash",
551
+ "inputTokens",
552
+ "outputTokens",
553
+ "costUsd"
554
+ ],
555
+ "additionalProperties": false
556
+ },
557
+ "contentHash": {
558
+ "$ref": "#/$defs/__schema2"
559
+ }
560
+ },
561
+ "required": [
562
+ "schemaVersion",
563
+ "graderId",
564
+ "graderVersion",
565
+ "score",
566
+ "passed",
567
+ "rewardEligible",
568
+ "failureClass",
569
+ "feedback",
570
+ "visibleEvidenceRefs",
571
+ "privilegedEvidenceRefs",
572
+ "contentHash"
573
+ ],
574
+ "additionalProperties": false
575
+ },
378
576
  "message": {
379
577
  "anyOf": [
380
578
  {
@@ -401,9 +599,143 @@
401
599
  "additionalProperties": false
402
600
  },
403
601
  "__schema5": {
602
+ "type": "string",
603
+ "maxLength": 20000
604
+ },
605
+ "__schema6": {
404
606
  "type": "string",
405
607
  "format": "date-time",
406
608
  "pattern": "^(?:(?:\\d\\d[2468][048]|\\d\\d[13579][26]|\\d\\d0[48]|[02468][048]00|[13579][26]00)-02-29|\\d{4}-(?:(?:0[13578]|1[02])-(?:0[1-9]|[12]\\d|3[01])|(?:0[469]|11)-(?:0[1-9]|[12]\\d|30)|(?:02)-(?:0[1-9]|1\\d|2[0-8])))T(?:(?:[01]\\d|2[0-3]):[0-5]\\d(?::[0-5]\\d(?:\\.\\d+)?)?(?:Z|([+-](?:[01]\\d|2[0-3]):[0-5]\\d)))$"
609
+ },
610
+ "__schema7": {
611
+ "type": "object",
612
+ "properties": {
613
+ "id": {
614
+ "$ref": "#/$defs/__schema0"
615
+ },
616
+ "requestHash": {
617
+ "$ref": "#/$defs/__schema2"
618
+ },
619
+ "reservedUsd": {
620
+ "$ref": "#/$defs/__schema8"
621
+ },
622
+ "status": {
623
+ "$ref": "#/$defs/__schema9"
624
+ },
625
+ "response": {
626
+ "$ref": "#/$defs/__schema10"
627
+ }
628
+ },
629
+ "required": [
630
+ "id",
631
+ "requestHash",
632
+ "reservedUsd",
633
+ "status",
634
+ "response"
635
+ ],
636
+ "additionalProperties": false
637
+ },
638
+ "__schema8": {
639
+ "type": "number",
640
+ "exclusiveMinimum": 0
641
+ },
642
+ "__schema9": {
643
+ "type": "string",
644
+ "enum": [
645
+ "reserved",
646
+ "settled",
647
+ "not_dispatched"
648
+ ]
649
+ },
650
+ "__schema10": {
651
+ "anyOf": [
652
+ {
653
+ "type": "object",
654
+ "properties": {
655
+ "text": {
656
+ "type": "string",
657
+ "maxLength": 100000
658
+ },
659
+ "modelId": {
660
+ "type": "string",
661
+ "minLength": 1,
662
+ "maxLength": 500
663
+ },
664
+ "modelRevision": {
665
+ "anyOf": [
666
+ {
667
+ "type": "string",
668
+ "minLength": 1,
669
+ "maxLength": 500
670
+ },
671
+ {
672
+ "type": "null"
673
+ }
674
+ ]
675
+ },
676
+ "responseId": {
677
+ "anyOf": [
678
+ {
679
+ "type": "string",
680
+ "minLength": 1,
681
+ "maxLength": 500
682
+ },
683
+ {
684
+ "type": "null"
685
+ }
686
+ ]
687
+ },
688
+ "inputTokens": {
689
+ "anyOf": [
690
+ {
691
+ "type": "integer",
692
+ "minimum": 0,
693
+ "maximum": 9007199254740991
694
+ },
695
+ {
696
+ "type": "null"
697
+ }
698
+ ]
699
+ },
700
+ "outputTokens": {
701
+ "anyOf": [
702
+ {
703
+ "type": "integer",
704
+ "minimum": 0,
705
+ "maximum": 9007199254740991
706
+ },
707
+ {
708
+ "type": "null"
709
+ }
710
+ ]
711
+ },
712
+ "costUsd": {
713
+ "anyOf": [
714
+ {
715
+ "type": "number",
716
+ "minimum": 0
717
+ },
718
+ {
719
+ "type": "null"
720
+ }
721
+ ]
722
+ }
723
+ },
724
+ "required": [
725
+ "text",
726
+ "modelId",
727
+ "modelRevision",
728
+ "responseId",
729
+ "inputTokens",
730
+ "outputTokens",
731
+ "costUsd"
732
+ ],
733
+ "additionalProperties": false
734
+ },
735
+ {
736
+ "type": "null"
737
+ }
738
+ ]
407
739
  }
408
740
  },
409
741
  "$id": "urn:openpond:learning:v1:grade",
@@ -230,6 +230,28 @@
230
230
  "fixtureSetRef": {
231
231
  "$ref": "#/$defs/__schema1"
232
232
  },
233
+ "calibrationCheckRef": {
234
+ "type": "object",
235
+ "properties": {
236
+ "id": {
237
+ "$ref": "#/$defs/__schema0"
238
+ },
239
+ "contentHash": {
240
+ "$ref": "#/$defs/__schema2"
241
+ },
242
+ "revision": {
243
+ "type": "integer",
244
+ "exclusiveMinimum": 0,
245
+ "maximum": 9007199254740991
246
+ }
247
+ },
248
+ "required": [
249
+ "id",
250
+ "contentHash",
251
+ "revision"
252
+ ],
253
+ "additionalProperties": false
254
+ },
233
255
  "contentHash": {
234
256
  "$ref": "#/$defs/__schema2"
235
257
  }