runbios-sdk 0.2.6-dev.196 → 0.2.6-dev.198
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.
- package/dist/index.d.ts +1 -1
- package/dist/index.js +1 -1
- package/dist/resources/datasets.d.ts +25 -1
- package/dist/resources/datasets.js +39 -0
- package/dist/types.d.ts +15 -0
- package/package.json +1 -1
package/dist/index.d.ts
CHANGED
|
@@ -36,7 +36,7 @@ import { Loop } from './resources/loop.js';
|
|
|
36
36
|
* SDK version. Sent as part of the User-Agent header.
|
|
37
37
|
* Must match package.json "version" -- enforced by a contract test.
|
|
38
38
|
*/
|
|
39
|
-
export declare const VERSION = "0.2.6-dev.
|
|
39
|
+
export declare const VERSION = "0.2.6-dev.198";
|
|
40
40
|
export declare class RunBiOS {
|
|
41
41
|
/** Search models, fetch configs, check adapter compatibility. */
|
|
42
42
|
readonly models: Models;
|
package/dist/index.js
CHANGED
|
@@ -36,7 +36,7 @@ import { Loop } from './resources/loop.js';
|
|
|
36
36
|
* SDK version. Sent as part of the User-Agent header.
|
|
37
37
|
* Must match package.json "version" -- enforced by a contract test.
|
|
38
38
|
*/
|
|
39
|
-
export const VERSION = '0.2.6-dev.
|
|
39
|
+
export const VERSION = '0.2.6-dev.198';
|
|
40
40
|
export class RunBiOS {
|
|
41
41
|
/** Search models, fetch configs, check adapter compatibility. */
|
|
42
42
|
models;
|
|
@@ -1,5 +1,5 @@
|
|
|
1
1
|
import type { HttpClient } from '../client.js';
|
|
2
|
-
import type { Dataset, DatasetListParams, DatasetListResponse, DatasetUploadParams, DatasetPreview, DatasetPreviewParams, DatasetImportHFParams, DatasetRegisterHFParams, DatasetHubSearchParams, DatasetHubPreviewParams, DatasetValidation, DatasetFormatSpecs, DatasetStorageUsage } from '../types.js';
|
|
2
|
+
import type { Dataset, DatasetListParams, DatasetListResponse, DatasetUploadParams, DatasetPreview, DatasetPreviewParams, DatasetImportHFParams, DatasetRegisterHFParams, DatasetRegisterLoopParams, DatasetHubSearchParams, DatasetHubPreviewParams, DatasetValidation, DatasetFormatSpecs, DatasetStorageUsage } from '../types.js';
|
|
3
3
|
/**
|
|
4
4
|
* Manage training datasets -- upload files, import from HuggingFace,
|
|
5
5
|
* preview contents, and manage storage.
|
|
@@ -118,6 +118,30 @@ export declare class Datasets {
|
|
|
118
118
|
* ```
|
|
119
119
|
*/
|
|
120
120
|
registerHuggingFace(params: DatasetRegisterHFParams | Record<string, unknown>): Promise<Dataset>;
|
|
121
|
+
/**
|
|
122
|
+
* Turn a curated Conscious Loop set into a dataset training can use.
|
|
123
|
+
*
|
|
124
|
+
* Without this, using a loop set from code meant downloading the JSONL and
|
|
125
|
+
* uploading it again as an unrelated dataset. That works, and it throws away
|
|
126
|
+
* the one thing that made the set worth having: which conversations, which
|
|
127
|
+
* verdicts and which judges produced it. Registering keeps that link.
|
|
128
|
+
*
|
|
129
|
+
* The set is then put through the SAME validation every uploaded dataset
|
|
130
|
+
* gets, so it comes back `processing` rather than ready -- poll
|
|
131
|
+
* {@link getStatus} before training on it. Registering the same set and split
|
|
132
|
+
* twice is refused rather than billed twice, and an empty set is refused
|
|
133
|
+
* rather than registered as something that would fail inside the trainer.
|
|
134
|
+
*
|
|
135
|
+
* @example
|
|
136
|
+
* ```ts
|
|
137
|
+
* const train = await client.datasets.registerLoop({
|
|
138
|
+
* loopDatasetId: 'd244894c-7ab4-4b3d-bfda-e0bb4c563ceb',
|
|
139
|
+
* name: 'extract-v1 distill SFT',
|
|
140
|
+
* split: 'train',
|
|
141
|
+
* });
|
|
142
|
+
* ```
|
|
143
|
+
*/
|
|
144
|
+
registerLoop(params: DatasetRegisterLoopParams | Record<string, unknown>): Promise<Dataset>;
|
|
121
145
|
/**
|
|
122
146
|
* Get the processing status of a dataset.
|
|
123
147
|
*
|
|
@@ -260,6 +260,45 @@ export class Datasets {
|
|
|
260
260
|
});
|
|
261
261
|
return normalizeDataset(result);
|
|
262
262
|
}
|
|
263
|
+
/**
|
|
264
|
+
* Turn a curated Conscious Loop set into a dataset training can use.
|
|
265
|
+
*
|
|
266
|
+
* Without this, using a loop set from code meant downloading the JSONL and
|
|
267
|
+
* uploading it again as an unrelated dataset. That works, and it throws away
|
|
268
|
+
* the one thing that made the set worth having: which conversations, which
|
|
269
|
+
* verdicts and which judges produced it. Registering keeps that link.
|
|
270
|
+
*
|
|
271
|
+
* The set is then put through the SAME validation every uploaded dataset
|
|
272
|
+
* gets, so it comes back `processing` rather than ready -- poll
|
|
273
|
+
* {@link getStatus} before training on it. Registering the same set and split
|
|
274
|
+
* twice is refused rather than billed twice, and an empty set is refused
|
|
275
|
+
* rather than registered as something that would fail inside the trainer.
|
|
276
|
+
*
|
|
277
|
+
* @example
|
|
278
|
+
* ```ts
|
|
279
|
+
* const train = await client.datasets.registerLoop({
|
|
280
|
+
* loopDatasetId: 'd244894c-7ab4-4b3d-bfda-e0bb4c563ceb',
|
|
281
|
+
* name: 'extract-v1 distill SFT',
|
|
282
|
+
* split: 'train',
|
|
283
|
+
* });
|
|
284
|
+
* ```
|
|
285
|
+
*/
|
|
286
|
+
async registerLoop(params) {
|
|
287
|
+
const raw = params;
|
|
288
|
+
const loopDatasetId = String(raw.loopDatasetId || raw.loop_dataset_id || '');
|
|
289
|
+
if (!loopDatasetId)
|
|
290
|
+
throw new Error('RunBiOS: loopDatasetId is required');
|
|
291
|
+
const workspaceId = String(raw.workspaceId || raw.workspace_id || '')
|
|
292
|
+
|| this._http.workspaceId
|
|
293
|
+
|| await this._http.resolveWorkspaceId();
|
|
294
|
+
const result = await this._http.fetchPost('/api/datasets/register-loop', {
|
|
295
|
+
loop_dataset_id: loopDatasetId,
|
|
296
|
+
workspace_id: workspaceId,
|
|
297
|
+
split: raw.split ?? 'all',
|
|
298
|
+
...(raw.name !== undefined && { name: raw.name }),
|
|
299
|
+
});
|
|
300
|
+
return normalizeDataset(result);
|
|
301
|
+
}
|
|
263
302
|
/**
|
|
264
303
|
* Get the processing status of a dataset.
|
|
265
304
|
*
|
package/dist/types.d.ts
CHANGED
|
@@ -444,6 +444,21 @@ export interface DatasetRegisterHFParams {
|
|
|
444
444
|
columnMapping?: Record<string, string>;
|
|
445
445
|
importMode?: 'auto' | 'reference' | 'materialize';
|
|
446
446
|
}
|
|
447
|
+
/** Parameters for registering a curated Conscious Loop set as a dataset. */
|
|
448
|
+
export interface DatasetRegisterLoopParams {
|
|
449
|
+
/** The Conscious Loop dataset to register. */
|
|
450
|
+
loopDatasetId: string;
|
|
451
|
+
/** Defaults to `loop-<first 8 of the loop id>`. */
|
|
452
|
+
name?: string;
|
|
453
|
+
/**
|
|
454
|
+
* Which rows to take. Defaults to `all`, because the loop already split the
|
|
455
|
+
* set by TIME and re-splitting a deliberate split throws that decision away.
|
|
456
|
+
* Register `train` and `holdout` separately to score a training job against
|
|
457
|
+
* the loop's own held-back rows; both can exist at once.
|
|
458
|
+
*/
|
|
459
|
+
split?: 'all' | 'train' | 'holdout';
|
|
460
|
+
workspaceId?: string;
|
|
461
|
+
}
|
|
447
462
|
/** Parameters for searching HuggingFace Hub datasets. */
|
|
448
463
|
export interface DatasetHubSearchParams {
|
|
449
464
|
/** Search query. */
|