@superlinked/sie-sdk 0.6.10 → 0.6.12
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.cjs +10 -4
- package/dist/index.cjs.map +1 -1
- package/dist/index.d.cts +20 -12
- package/dist/index.d.ts +20 -12
- package/dist/index.js +10 -4
- package/dist/index.js.map +1 -1
- package/package.json +1 -1
package/dist/index.d.cts
CHANGED
|
@@ -360,17 +360,25 @@ interface CapacityInfo {
|
|
|
360
360
|
workers: WorkerInfo[];
|
|
361
361
|
}
|
|
362
362
|
/**
|
|
363
|
-
* Pool specification
|
|
363
|
+
* Pool specification returned by the gateway.
|
|
364
364
|
*/
|
|
365
365
|
interface PoolSpec {
|
|
366
366
|
/** Pool name (used in GPU param as "poolName/machineProfile") */
|
|
367
367
|
name: string;
|
|
368
|
+
/** Helm/NATS queue namespace backing this logical pool */
|
|
369
|
+
queue_pool?: string;
|
|
370
|
+
/** Optional bundle constraint for this pool */
|
|
371
|
+
bundle?: string | null;
|
|
368
372
|
/** Machine profile requirements for pool readiness, e.g., { l4: 2, "a100-40gb": 1 } */
|
|
369
373
|
gpus?: Record<string, number>;
|
|
370
|
-
/** Optional maximum assigned workers per machine profile */
|
|
371
|
-
gpuCaps?: Record<string, number>;
|
|
372
374
|
/** Optional maximum assigned workers per machine profile, as returned by the gateway */
|
|
373
375
|
gpu_caps?: Record<string, number>;
|
|
376
|
+
/** Optional TTL for dynamic pool leases */
|
|
377
|
+
ttl_seconds?: number | null;
|
|
378
|
+
/** Minimum workers kept warm for this pool */
|
|
379
|
+
minimum_worker_count?: number;
|
|
380
|
+
/** Models pinned for this pool */
|
|
381
|
+
pinned_models?: string[];
|
|
374
382
|
}
|
|
375
383
|
/**
|
|
376
384
|
* Pool status information.
|
|
@@ -396,10 +404,7 @@ interface PoolInfo {
|
|
|
396
404
|
/** Pool name */
|
|
397
405
|
name: string;
|
|
398
406
|
/** Pool specification */
|
|
399
|
-
spec:
|
|
400
|
-
gpus?: Record<string, number>;
|
|
401
|
-
gpu_caps?: Record<string, number>;
|
|
402
|
-
};
|
|
407
|
+
spec: PoolSpec;
|
|
403
408
|
/** Pool status */
|
|
404
409
|
status: PoolStatus;
|
|
405
410
|
}
|
|
@@ -924,7 +929,7 @@ declare function toFloat32Array(arr: number[]): Float32Array;
|
|
|
924
929
|
* ```typescript
|
|
925
930
|
* const client = new SIEClient("http://gateway:8080");
|
|
926
931
|
*
|
|
927
|
-
* // Create a
|
|
932
|
+
* // Create a logical pool backed by the cluster's default worker queue
|
|
928
933
|
* await client.createPool("eval-bench", { l4: 2 });
|
|
929
934
|
*
|
|
930
935
|
* // Use pool for requests
|
|
@@ -1228,12 +1233,15 @@ declare class SIEClient {
|
|
|
1228
1233
|
/**
|
|
1229
1234
|
* Create or update a resource pool for isolated capacity.
|
|
1230
1235
|
*
|
|
1231
|
-
* Pools provide
|
|
1232
|
-
*
|
|
1236
|
+
* Pools provide logical capacity isolation. By default they draw from the
|
|
1237
|
+
* cluster's `default` Helm/NATS queue; pass `queuePool` only when the cluster
|
|
1238
|
+
* has a dedicated physical worker queue declared under
|
|
1239
|
+
* `queueRouting.staticQueuePools` for this workload.
|
|
1233
1240
|
*
|
|
1234
1241
|
* @param name - Pool name (used in GPU param as "poolName/machineProfile")
|
|
1235
1242
|
* @param gpus - Optional machine profile requirements for pool readiness, e.g., { "l4": 2, "l4-spot": 1 }
|
|
1236
1243
|
* @param gpuCaps - Optional maximum assigned workers per machine profile
|
|
1244
|
+
* @param queuePool - Optional Helm/NATS queue namespace backing this logical pool. Defaults to "default".
|
|
1237
1245
|
*
|
|
1238
1246
|
* @example
|
|
1239
1247
|
* ```typescript
|
|
@@ -1247,7 +1255,7 @@ declare class SIEClient {
|
|
|
1247
1255
|
* await client.deletePool("eval-bench");
|
|
1248
1256
|
* ```
|
|
1249
1257
|
*/
|
|
1250
|
-
createPool(name: string, gpus?: Record<string, number>, gpuCaps?: Record<string, number
|
|
1258
|
+
createPool(name: string, gpus?: Record<string, number>, gpuCaps?: Record<string, number>, queuePool?: string): Promise<void>;
|
|
1251
1259
|
/**
|
|
1252
1260
|
* Get information about a pool.
|
|
1253
1261
|
*
|
|
@@ -1360,7 +1368,7 @@ declare class SIEClient {
|
|
|
1360
1368
|
private detectEndpointType;
|
|
1361
1369
|
}
|
|
1362
1370
|
|
|
1363
|
-
declare const SDK_VERSION = "0.6.
|
|
1371
|
+
declare const SDK_VERSION = "0.6.12";
|
|
1364
1372
|
|
|
1365
1373
|
/**
|
|
1366
1374
|
* Helpers for converting SIE encode results to plain JavaScript types.
|
package/dist/index.d.ts
CHANGED
|
@@ -360,17 +360,25 @@ interface CapacityInfo {
|
|
|
360
360
|
workers: WorkerInfo[];
|
|
361
361
|
}
|
|
362
362
|
/**
|
|
363
|
-
* Pool specification
|
|
363
|
+
* Pool specification returned by the gateway.
|
|
364
364
|
*/
|
|
365
365
|
interface PoolSpec {
|
|
366
366
|
/** Pool name (used in GPU param as "poolName/machineProfile") */
|
|
367
367
|
name: string;
|
|
368
|
+
/** Helm/NATS queue namespace backing this logical pool */
|
|
369
|
+
queue_pool?: string;
|
|
370
|
+
/** Optional bundle constraint for this pool */
|
|
371
|
+
bundle?: string | null;
|
|
368
372
|
/** Machine profile requirements for pool readiness, e.g., { l4: 2, "a100-40gb": 1 } */
|
|
369
373
|
gpus?: Record<string, number>;
|
|
370
|
-
/** Optional maximum assigned workers per machine profile */
|
|
371
|
-
gpuCaps?: Record<string, number>;
|
|
372
374
|
/** Optional maximum assigned workers per machine profile, as returned by the gateway */
|
|
373
375
|
gpu_caps?: Record<string, number>;
|
|
376
|
+
/** Optional TTL for dynamic pool leases */
|
|
377
|
+
ttl_seconds?: number | null;
|
|
378
|
+
/** Minimum workers kept warm for this pool */
|
|
379
|
+
minimum_worker_count?: number;
|
|
380
|
+
/** Models pinned for this pool */
|
|
381
|
+
pinned_models?: string[];
|
|
374
382
|
}
|
|
375
383
|
/**
|
|
376
384
|
* Pool status information.
|
|
@@ -396,10 +404,7 @@ interface PoolInfo {
|
|
|
396
404
|
/** Pool name */
|
|
397
405
|
name: string;
|
|
398
406
|
/** Pool specification */
|
|
399
|
-
spec:
|
|
400
|
-
gpus?: Record<string, number>;
|
|
401
|
-
gpu_caps?: Record<string, number>;
|
|
402
|
-
};
|
|
407
|
+
spec: PoolSpec;
|
|
403
408
|
/** Pool status */
|
|
404
409
|
status: PoolStatus;
|
|
405
410
|
}
|
|
@@ -924,7 +929,7 @@ declare function toFloat32Array(arr: number[]): Float32Array;
|
|
|
924
929
|
* ```typescript
|
|
925
930
|
* const client = new SIEClient("http://gateway:8080");
|
|
926
931
|
*
|
|
927
|
-
* // Create a
|
|
932
|
+
* // Create a logical pool backed by the cluster's default worker queue
|
|
928
933
|
* await client.createPool("eval-bench", { l4: 2 });
|
|
929
934
|
*
|
|
930
935
|
* // Use pool for requests
|
|
@@ -1228,12 +1233,15 @@ declare class SIEClient {
|
|
|
1228
1233
|
/**
|
|
1229
1234
|
* Create or update a resource pool for isolated capacity.
|
|
1230
1235
|
*
|
|
1231
|
-
* Pools provide
|
|
1232
|
-
*
|
|
1236
|
+
* Pools provide logical capacity isolation. By default they draw from the
|
|
1237
|
+
* cluster's `default` Helm/NATS queue; pass `queuePool` only when the cluster
|
|
1238
|
+
* has a dedicated physical worker queue declared under
|
|
1239
|
+
* `queueRouting.staticQueuePools` for this workload.
|
|
1233
1240
|
*
|
|
1234
1241
|
* @param name - Pool name (used in GPU param as "poolName/machineProfile")
|
|
1235
1242
|
* @param gpus - Optional machine profile requirements for pool readiness, e.g., { "l4": 2, "l4-spot": 1 }
|
|
1236
1243
|
* @param gpuCaps - Optional maximum assigned workers per machine profile
|
|
1244
|
+
* @param queuePool - Optional Helm/NATS queue namespace backing this logical pool. Defaults to "default".
|
|
1237
1245
|
*
|
|
1238
1246
|
* @example
|
|
1239
1247
|
* ```typescript
|
|
@@ -1247,7 +1255,7 @@ declare class SIEClient {
|
|
|
1247
1255
|
* await client.deletePool("eval-bench");
|
|
1248
1256
|
* ```
|
|
1249
1257
|
*/
|
|
1250
|
-
createPool(name: string, gpus?: Record<string, number>, gpuCaps?: Record<string, number
|
|
1258
|
+
createPool(name: string, gpus?: Record<string, number>, gpuCaps?: Record<string, number>, queuePool?: string): Promise<void>;
|
|
1251
1259
|
/**
|
|
1252
1260
|
* Get information about a pool.
|
|
1253
1261
|
*
|
|
@@ -1360,7 +1368,7 @@ declare class SIEClient {
|
|
|
1360
1368
|
private detectEndpointType;
|
|
1361
1369
|
}
|
|
1362
1370
|
|
|
1363
|
-
declare const SDK_VERSION = "0.6.
|
|
1371
|
+
declare const SDK_VERSION = "0.6.12";
|
|
1364
1372
|
|
|
1365
1373
|
/**
|
|
1366
1374
|
* Helpers for converting SIE encode results to plain JavaScript types.
|
package/dist/index.js
CHANGED
|
@@ -854,7 +854,7 @@ function extractDataPayload(block) {
|
|
|
854
854
|
}
|
|
855
855
|
|
|
856
856
|
// src/version.ts
|
|
857
|
-
var SDK_VERSION = "0.6.
|
|
857
|
+
var SDK_VERSION = "0.6.12";
|
|
858
858
|
|
|
859
859
|
// src/client.ts
|
|
860
860
|
function sleep2(ms) {
|
|
@@ -1631,12 +1631,15 @@ var SIEClient = class {
|
|
|
1631
1631
|
/**
|
|
1632
1632
|
* Create or update a resource pool for isolated capacity.
|
|
1633
1633
|
*
|
|
1634
|
-
* Pools provide
|
|
1635
|
-
*
|
|
1634
|
+
* Pools provide logical capacity isolation. By default they draw from the
|
|
1635
|
+
* cluster's `default` Helm/NATS queue; pass `queuePool` only when the cluster
|
|
1636
|
+
* has a dedicated physical worker queue declared under
|
|
1637
|
+
* `queueRouting.staticQueuePools` for this workload.
|
|
1636
1638
|
*
|
|
1637
1639
|
* @param name - Pool name (used in GPU param as "poolName/machineProfile")
|
|
1638
1640
|
* @param gpus - Optional machine profile requirements for pool readiness, e.g., { "l4": 2, "l4-spot": 1 }
|
|
1639
1641
|
* @param gpuCaps - Optional maximum assigned workers per machine profile
|
|
1642
|
+
* @param queuePool - Optional Helm/NATS queue namespace backing this logical pool. Defaults to "default".
|
|
1640
1643
|
*
|
|
1641
1644
|
* @example
|
|
1642
1645
|
* ```typescript
|
|
@@ -1650,7 +1653,7 @@ var SIEClient = class {
|
|
|
1650
1653
|
* await client.deletePool("eval-bench");
|
|
1651
1654
|
* ```
|
|
1652
1655
|
*/
|
|
1653
|
-
async createPool(name, gpus, gpuCaps) {
|
|
1656
|
+
async createPool(name, gpus, gpuCaps, queuePool) {
|
|
1654
1657
|
const alreadyTracking = this.pools.has(name);
|
|
1655
1658
|
const requestBody = {
|
|
1656
1659
|
name
|
|
@@ -1661,6 +1664,9 @@ var SIEClient = class {
|
|
|
1661
1664
|
if (gpuCaps) {
|
|
1662
1665
|
requestBody.gpu_caps = gpuCaps;
|
|
1663
1666
|
}
|
|
1667
|
+
if (queuePool) {
|
|
1668
|
+
requestBody.queue_pool = queuePool;
|
|
1669
|
+
}
|
|
1664
1670
|
const url = `${this.baseUrl}/v1/pools`;
|
|
1665
1671
|
const headers = {
|
|
1666
1672
|
"Content-Type": JSON_CONTENT_TYPE,
|