UNPKG

clustering-tfjs

Version:

High-performance TypeScript clustering algorithms (K-Means, Spectral, Agglomerative) with TensorFlow.js acceleration and scikit-learn compatibility

43 lines (42 loc) 1.6 kB
/** * TensorFlow.js adapter module * * This module provides a platform-agnostic interface to TensorFlow.js, * allowing the library to work in both Node.js and browser environments. * * Phase 2 implementation: For now, we maintain backward compatibility * by still importing tfjs-node directly, but the infrastructure for * dynamic loading is ready in tf-backend.ts and loaders. */ // Handle Windows CI environment where native modules fail let tf; if (process.platform === 'win32' && process.env.CI) { // Use pure JS implementation on Windows CI try { tf = require('@tensorflow/tfjs'); } catch (error) { console.error('tf-adapter: Failed to load @tensorflow/tfjs on Windows CI'); throw new Error(`Failed to load TensorFlow.js: ${error instanceof Error ? error.message : String(error)}`); } } else { try { // Use Node.js backend for better performance tf = require('@tensorflow/tfjs-node'); } catch (error) { // Fallback to pure JS if tfjs-node fails to load console.warn('tf-adapter: Failed to load @tensorflow/tfjs-node, using pure JS fallback'); try { tf = require('@tensorflow/tfjs'); } catch (fallbackError) { console.error('tf-adapter: Failed to load @tensorflow/tfjs fallback'); throw new Error(`Failed to load TensorFlow.js: ${fallbackError instanceof Error ? fallbackError.message : String(fallbackError)}`); } } } export default tf; // Re-export everything from the loaded module export * from '@tensorflow/tfjs-core';