Moss SDK Documentation

Appearance

@inferedge/moss-clustering v1.0.0-beta.3

@inferedge/moss-clustering / MossClusteringClient

Class: MossClusteringClient ​

Main client for interacting with the Moss Clustering API

Example ​

ts
1. Instantiate with your project ID and API key.
2. Call `generateClusters(5)` to create clusters.
3. Access results via `getClusters()`.

Constructors ​

Constructor ​

new MossClusteringClient(options): MossClusteringClient

Creates a new MossClusteringClient instance

Parameters ​

options ​

MossClusteringOptions

Configuration options

Returns ​

MossClusteringClient

Methods ​

getServiceHealth() ​

getServiceHealth(endpoint): Promise<HealthStatus>

Retrieves service health information.

Parameters ​

endpoint ​

Health endpoint to query (health, healthz, or readyz)

"health" | "healthz" | "readyz"

Returns ​

Promise<HealthStatus>

Health status payload

startClusterGeneration() ​

startClusterGeneration(k, options): Promise<GenerateClusterJobResponse>

Starts cluster generation job (async operation)

This method starts an async job to perform k-means++ clustering on indexes in the project. The operation runs in the background. Use the returned jobId to poll for status.

Parameters ​

k ​

number

Number of clusters to generate (must be > 0)

options ​

StartClusterOptions = {}

Optional settings such as a subset of index names to cluster

Returns ​

Promise<GenerateClusterJobResponse>

Job information with jobId to poll for status

Throws ​

ClusteringError If k is invalid or the API request fails

Example ​

typescript
const jobResponse = await clustering.startClusterGeneration(5);
console.log('Job started:', jobResponse.jobId);
// Pass StartClusterOptions.indexNames to limit specific indexes

getJobStatus() ​

getJobStatus(jobId): Promise<GetJobResponse>

Gets the status of a clustering job

Parameters ​

jobId ​

string

Job ID returned from startClusterGeneration

Returns ​

Promise<GetJobResponse>

Job status with progress and results (if completed)

Example ​

typescript
const jobStatus = await clustering.getJobStatus(jobId);
console.log('Status:', jobStatus.job.status);
console.log('Progress:', jobStatus.job.progress);
console.log('Phase:', jobStatus.job.currentPhase);

waitForJobCompletion() ​

waitForJobCompletion(jobId, pollInterval, maxWaitTime): Promise<ClusteringResult>

Waits for a job to complete by polling

This is a convenience method that polls getJobStatus until the job completes or fails.

Parameters ​

jobId ​

string

Job ID to wait for

pollInterval ​

number = 2000

Milliseconds between polls (default: 2000)

maxWaitTime ​

number = 300000

Maximum time to wait in milliseconds (default: 300000 = 5 minutes)

Returns ​

Promise<ClusteringResult>

Clustering result when job completes

Throws ​

ClusteringError If job fails or times out

Example ​

typescript
const jobResponse = await clustering.startClusterGeneration(5);
const result = await clustering.waitForJobCompletion(jobResponse.jobId, 2000, 600000);
console.log('Completed job with clusters:', result.clusters.length);

generateClusters() ​

generateClusters(k, optionsOrPollInterval?, legacyMaxWaitTime?): Promise<ClusteringResult>

Generates clusters and waits for completion (convenience method)

This method combines startClusterGeneration and waitForJobCompletion for a simpler API when you want to wait for the result.

Parameters ​

k ​

number

Number of clusters to generate (must be > 0)

optionsOrPollInterval? ​

Either an options object or the polling interval in milliseconds

number | GenerateClustersOptions

legacyMaxWaitTime? ​

number

Maximum wait time when using the legacy signature

Returns ​

Promise<ClusteringResult>

Clustering result with cluster information

Throws ​

ClusteringError If k is invalid, job fails, or times out

Example ​

typescript
// Simple usage - waits for completion
const result = await clustering.generateClusters(5);
console.log('Created clusters:', result.clusters.length);

Advanced usage: supply GenerateClustersOptions with indexNames, pollInterval, and maxWaitTime to process subsets or override polling behavior.

getClusters() ​

getClusters(): Promise<ClusterInfo[]>

Retrieves all clusters for the project

Returns current clustering state with index counts and assignments. Returns empty array if no clustering exists.

Returns ​

Promise<ClusterInfo[]>

Array of cluster information

Example ​

typescript
const clusters = await clustering.getClusters();
clusters.forEach(cluster =>
  console.log('Cluster label:', cluster.label, 'indexes:', cluster.indexCount)
);

getIndexCluster() ​

getIndexCluster(indexName): Promise<ClusterAssignment | null>

Gets the cluster assignment for a specific index

Parameters ​

indexName ​

string

Name of the index

Returns ​

Promise<ClusterAssignment | null>

Cluster assignment or null if not classified

Example ​

typescript
const assignment = await clustering.getIndexCluster('my-index');
console.log('Belongs to:', assignment?.label);

clusterIndexDocuments() ​

clusterIndexDocuments(indexName, k, options?): Promise<GenerateIndexDocumentClustersResponse>

Clusters documents stored within a specific index

Parameters ​

indexName ​

string

Name of the index whose documents should be clustered

k ​

number

Number of clusters to generate (must be > 0)

options? ​

DocumentClusteringOptions

Optional tuning parameters such as sampling strategy

Returns ​

Promise<GenerateIndexDocumentClustersResponse>

Document clustering payload including per-phase progress

getIndexDocumentClusters() ​

getIndexDocumentClusters(indexName): Promise<GetIndexDocumentClustersResponse | null>

Retrieves the most recently generated document clusters for an index

Returns null if document clustering has not been generated yet.

Parameters ​

indexName ​

string

Name of the index to query

Returns ​

Promise<GetIndexDocumentClustersResponse | null>

Document clustering snapshot or null when unavailable

hasActiveClustering() ​

hasActiveClustering(): Promise<boolean>

Checks if clustering exists for the project

Quick check to determine if generateClusters() has been called.

Returns ​

Promise<boolean>

True if clustering exists, false otherwise

Example ​

typescript
if (!(await clustering.hasActiveClustering())) {
  await clustering.generateClusters(5);
}

deleteClustering() ​

deleteClustering(): Promise<void>

Deletes all clustering data for the project

This removes all clusters and assignments. This action cannot be undone. You will need to call generateClusters() again to re-cluster.

Returns ​

Promise<void>

Example ​

typescript
await clustering.deleteClustering();
console.log('All clustering data deleted');