@ruvector/edge-full 0.1.0

package/onnx/ruvector_onnx_embeddings_wasm.js

@@ -0,0 +1,635 @@
+
+let imports = {};
+imports['__wbindgen_placeholder__'] = module.exports;
+
+function addToExternrefTable0(obj) {
+    const idx = wasm.__externref_table_alloc();
+    wasm.__wbindgen_externrefs.set(idx, obj);
+    return idx;
+}
+
+function _assertClass(instance, klass) {
+    if (!(instance instanceof klass)) {
+        throw new Error(`expected instance of ${klass.name}`);
+    }
+}
+
+function getArrayF32FromWasm0(ptr, len) {
+    ptr = ptr >>> 0;
+    return getFloat32ArrayMemory0().subarray(ptr / 4, ptr / 4 + len);
+}
+
+function getArrayU8FromWasm0(ptr, len) {
+    ptr = ptr >>> 0;
+    return getUint8ArrayMemory0().subarray(ptr / 1, ptr / 1 + len);
+}
+
+let cachedDataViewMemory0 = null;
+function getDataViewMemory0() {
+    if (cachedDataViewMemory0 === null || cachedDataViewMemory0.buffer.detached === true || (cachedDataViewMemory0.buffer.detached === undefined && cachedDataViewMemory0.buffer !== wasm.memory.buffer)) {
+        cachedDataViewMemory0 = new DataView(wasm.memory.buffer);
+    }
+    return cachedDataViewMemory0;
+}
+
+let cachedFloat32ArrayMemory0 = null;
+function getFloat32ArrayMemory0() {
+    if (cachedFloat32ArrayMemory0 === null || cachedFloat32ArrayMemory0.byteLength === 0) {
+        cachedFloat32ArrayMemory0 = new Float32Array(wasm.memory.buffer);
+    }
+    return cachedFloat32ArrayMemory0;
+}
+
+function getStringFromWasm0(ptr, len) {
+    ptr = ptr >>> 0;
+    return decodeText(ptr, len);
+}
+
+let cachedUint8ArrayMemory0 = null;
+function getUint8ArrayMemory0() {
+    if (cachedUint8ArrayMemory0 === null || cachedUint8ArrayMemory0.byteLength === 0) {
+        cachedUint8ArrayMemory0 = new Uint8Array(wasm.memory.buffer);
+    }
+    return cachedUint8ArrayMemory0;
+}
+
+function handleError(f, args) {
+    try {
+        return f.apply(this, args);
+    } catch (e) {
+        const idx = addToExternrefTable0(e);
+        wasm.__wbindgen_exn_store(idx);
+    }
+}
+
+function isLikeNone(x) {
+    return x === undefined || x === null;
+}
+
+function passArray8ToWasm0(arg, malloc) {
+    const ptr = malloc(arg.length * 1, 1) >>> 0;
+    getUint8ArrayMemory0().set(arg, ptr / 1);
+    WASM_VECTOR_LEN = arg.length;
+    return ptr;
+}
+
+function passArrayF32ToWasm0(arg, malloc) {
+    const ptr = malloc(arg.length * 4, 4) >>> 0;
+    getFloat32ArrayMemory0().set(arg, ptr / 4);
+    WASM_VECTOR_LEN = arg.length;
+    return ptr;
+}
+
+function passArrayJsValueToWasm0(array, malloc) {
+    const ptr = malloc(array.length * 4, 4) >>> 0;
+    for (let i = 0; i < array.length; i++) {
+        const add = addToExternrefTable0(array[i]);
+        getDataViewMemory0().setUint32(ptr + 4 * i, add, true);
+    }
+    WASM_VECTOR_LEN = array.length;
+    return ptr;
+}
+
+function passStringToWasm0(arg, malloc, realloc) {
+    if (realloc === undefined) {
+        const buf = cachedTextEncoder.encode(arg);
+        const ptr = malloc(buf.length, 1) >>> 0;
+        getUint8ArrayMemory0().subarray(ptr, ptr + buf.length).set(buf);
+        WASM_VECTOR_LEN = buf.length;
+        return ptr;
+    }
+
+    let len = arg.length;
+    let ptr = malloc(len, 1) >>> 0;
+
+    const mem = getUint8ArrayMemory0();
+
+    let offset = 0;
+
+    for (; offset < len; offset++) {
+        const code = arg.charCodeAt(offset);
+        if (code > 0x7F) break;
+        mem[ptr + offset] = code;
+    }
+    if (offset !== len) {
+        if (offset !== 0) {
+            arg = arg.slice(offset);
+        }
+        ptr = realloc(ptr, len, len = offset + arg.length * 3, 1) >>> 0;
+        const view = getUint8ArrayMemory0().subarray(ptr + offset, ptr + len);
+        const ret = cachedTextEncoder.encodeInto(arg, view);
+
+        offset += ret.written;
+        ptr = realloc(ptr, len, offset, 1) >>> 0;
+    }
+
+    WASM_VECTOR_LEN = offset;
+    return ptr;
+}
+
+function takeFromExternrefTable0(idx) {
+    const value = wasm.__wbindgen_externrefs.get(idx);
+    wasm.__externref_table_dealloc(idx);
+    return value;
+}
+
+let cachedTextDecoder = new TextDecoder('utf-8', { ignoreBOM: true, fatal: true });
+cachedTextDecoder.decode();
+function decodeText(ptr, len) {
+    return cachedTextDecoder.decode(getUint8ArrayMemory0().subarray(ptr, ptr + len));
+}
+
+const cachedTextEncoder = new TextEncoder();
+
+if (!('encodeInto' in cachedTextEncoder)) {
+    cachedTextEncoder.encodeInto = function (arg, view) {
+        const buf = cachedTextEncoder.encode(arg);
+        view.set(buf);
+        return {
+            read: arg.length,
+            written: buf.length
+        };
+    }
+}
+
+let WASM_VECTOR_LEN = 0;
+
+const WasmEmbedderFinalization = (typeof FinalizationRegistry === 'undefined')
+    ? { register: () => {}, unregister: () => {} }
+    : new FinalizationRegistry(ptr => wasm.__wbg_wasmembedder_free(ptr >>> 0, 1));
+
+const WasmEmbedderConfigFinalization = (typeof FinalizationRegistry === 'undefined')
+    ? { register: () => {}, unregister: () => {} }
+    : new FinalizationRegistry(ptr => wasm.__wbg_wasmembedderconfig_free(ptr >>> 0, 1));
+
+/**
+ * Strategy for pooling token embeddings into a single sentence embedding
+ * @enum {0 | 1 | 2 | 3 | 4}
+ */
+const PoolingStrategy = Object.freeze({
+    /**
+     * Average all token embeddings (most common)
+     */
+    Mean: 0, "0": "Mean",
+    /**
+     * Use only the [CLS] token embedding
+     */
+    Cls: 1, "1": "Cls",
+    /**
+     * Take the maximum value across all tokens for each dimension
+     */
+    Max: 2, "2": "Max",
+    /**
+     * Mean pooling normalized by sqrt of sequence length
+     */
+    MeanSqrtLen: 3, "3": "MeanSqrtLen",
+    /**
+     * Use the last token embedding (for decoder models)
+     */
+    LastToken: 4, "4": "LastToken",
+});
+exports.PoolingStrategy = PoolingStrategy;
+
+/**
+ * WASM-compatible embedder using Tract for inference
+ */
+class WasmEmbedder {
+    static __wrap(ptr) {
+        ptr = ptr >>> 0;
+        const obj = Object.create(WasmEmbedder.prototype);
+        obj.__wbg_ptr = ptr;
+        WasmEmbedderFinalization.register(obj, obj.__wbg_ptr, obj);
+        return obj;
+    }
+    __destroy_into_raw() {
+        const ptr = this.__wbg_ptr;
+        this.__wbg_ptr = 0;
+        WasmEmbedderFinalization.unregister(this);
+        return ptr;
+    }
+    free() {
+        const ptr = this.__destroy_into_raw();
+        wasm.__wbg_wasmembedder_free(ptr, 0);
+    }
+    /**
+     * Get maximum sequence length
+     * @returns {number}
+     */
+    maxLength() {
+        const ret = wasm.wasmembedder_maxLength(this.__wbg_ptr);
+        return ret >>> 0;
+    }
+    /**
+     * Compute similarity between two texts
+     * @param {string} text1
+     * @param {string} text2
+     * @returns {number}
+     */
+    similarity(text1, text2) {
+        const ptr0 = passStringToWasm0(text1, wasm.__wbindgen_malloc, wasm.__wbindgen_realloc);
+        const len0 = WASM_VECTOR_LEN;
+        const ptr1 = passStringToWasm0(text2, wasm.__wbindgen_malloc, wasm.__wbindgen_realloc);
+        const len1 = WASM_VECTOR_LEN;
+        const ret = wasm.wasmembedder_similarity(this.__wbg_ptr, ptr0, len0, ptr1, len1);
+        if (ret[2]) {
+            throw takeFromExternrefTable0(ret[1]);
+        }
+        return ret[0];
+    }
+    /**
+     * Generate embeddings for multiple texts
+     * @param {string[]} texts
+     * @returns {Float32Array}
+     */
+    embedBatch(texts) {
+        const ptr0 = passArrayJsValueToWasm0(texts, wasm.__wbindgen_malloc);
+        const len0 = WASM_VECTOR_LEN;
+        const ret = wasm.wasmembedder_embedBatch(this.__wbg_ptr, ptr0, len0);
+        if (ret[3]) {
+            throw takeFromExternrefTable0(ret[2]);
+        }
+        var v2 = getArrayF32FromWasm0(ret[0], ret[1]).slice();
+        wasm.__wbindgen_free(ret[0], ret[1] * 4, 4);
+        return v2;
+    }
+    /**
+     * Create embedder with custom configuration
+     * @param {Uint8Array} model_bytes
+     * @param {string} tokenizer_json
+     * @param {WasmEmbedderConfig} config
+     * @returns {WasmEmbedder}
+     */
+    static withConfig(model_bytes, tokenizer_json, config) {
+        const ptr0 = passArray8ToWasm0(model_bytes, wasm.__wbindgen_malloc);
+        const len0 = WASM_VECTOR_LEN;
+        const ptr1 = passStringToWasm0(tokenizer_json, wasm.__wbindgen_malloc, wasm.__wbindgen_realloc);
+        const len1 = WASM_VECTOR_LEN;
+        _assertClass(config, WasmEmbedderConfig);
+        var ptr2 = config.__destroy_into_raw();
+        const ret = wasm.wasmembedder_withConfig(ptr0, len0, ptr1, len1, ptr2);
+        if (ret[2]) {
+            throw takeFromExternrefTable0(ret[1]);
+        }
+        return WasmEmbedder.__wrap(ret[0]);
+    }
+    /**
+     * Create a new embedder from model and tokenizer bytes
+     *
+     * # Arguments
+     * * `model_bytes` - ONNX model file bytes
+     * * `tokenizer_json` - Tokenizer JSON configuration
+     * @param {Uint8Array} model_bytes
+     * @param {string} tokenizer_json
+     */
+    constructor(model_bytes, tokenizer_json) {
+        const ptr0 = passArray8ToWasm0(model_bytes, wasm.__wbindgen_malloc);
+        const len0 = WASM_VECTOR_LEN;
+        const ptr1 = passStringToWasm0(tokenizer_json, wasm.__wbindgen_malloc, wasm.__wbindgen_realloc);
+        const len1 = WASM_VECTOR_LEN;
+        const ret = wasm.wasmembedder_new(ptr0, len0, ptr1, len1);
+        if (ret[2]) {
+            throw takeFromExternrefTable0(ret[1]);
+        }
+        this.__wbg_ptr = ret[0] >>> 0;
+        WasmEmbedderFinalization.register(this, this.__wbg_ptr, this);
+        return this;
+    }
+    /**
+     * Get the embedding dimension
+     * @returns {number}
+     */
+    dimension() {
+        const ret = wasm.wasmembedder_dimension(this.__wbg_ptr);
+        return ret >>> 0;
+    }
+    /**
+     * Generate embedding for a single text
+     * @param {string} text
+     * @returns {Float32Array}
+     */
+    embedOne(text) {
+        const ptr0 = passStringToWasm0(text, wasm.__wbindgen_malloc, wasm.__wbindgen_realloc);
+        const len0 = WASM_VECTOR_LEN;
+        const ret = wasm.wasmembedder_embedOne(this.__wbg_ptr, ptr0, len0);
+        if (ret[3]) {
+            throw takeFromExternrefTable0(ret[2]);
+        }
+        var v2 = getArrayF32FromWasm0(ret[0], ret[1]).slice();
+        wasm.__wbindgen_free(ret[0], ret[1] * 4, 4);
+        return v2;
+    }
+}
+if (Symbol.dispose) WasmEmbedder.prototype[Symbol.dispose] = WasmEmbedder.prototype.free;
+exports.WasmEmbedder = WasmEmbedder;
+
+/**
+ * Configuration for the WASM embedder
+ */
+class WasmEmbedderConfig {
+    static __wrap(ptr) {
+        ptr = ptr >>> 0;
+        const obj = Object.create(WasmEmbedderConfig.prototype);
+        obj.__wbg_ptr = ptr;
+        WasmEmbedderConfigFinalization.register(obj, obj.__wbg_ptr, obj);
+        return obj;
+    }
+    __destroy_into_raw() {
+        const ptr = this.__wbg_ptr;
+        this.__wbg_ptr = 0;
+        WasmEmbedderConfigFinalization.unregister(this);
+        return ptr;
+    }
+    free() {
+        const ptr = this.__destroy_into_raw();
+        wasm.__wbg_wasmembedderconfig_free(ptr, 0);
+    }
+    /**
+     * Set pooling strategy (0=Mean, 1=Cls, 2=Max, 3=MeanSqrtLen, 4=LastToken)
+     * @param {number} pooling
+     * @returns {WasmEmbedderConfig}
+     */
+    setPooling(pooling) {
+        const ptr = this.__destroy_into_raw();
+        const ret = wasm.wasmembedderconfig_setPooling(ptr, pooling);
+        return WasmEmbedderConfig.__wrap(ret);
+    }
+    /**
+     * Set whether to normalize embeddings
+     * @param {boolean} normalize
+     * @returns {WasmEmbedderConfig}
+     */
+    setNormalize(normalize) {
+        const ptr = this.__destroy_into_raw();
+        const ret = wasm.wasmembedderconfig_setNormalize(ptr, normalize);
+        return WasmEmbedderConfig.__wrap(ret);
+    }
+    /**
+     * Set maximum sequence length
+     * @param {number} max_length
+     * @returns {WasmEmbedderConfig}
+     */
+    setMaxLength(max_length) {
+        const ptr = this.__destroy_into_raw();
+        const ret = wasm.wasmembedderconfig_setMaxLength(ptr, max_length);
+        return WasmEmbedderConfig.__wrap(ret);
+    }
+    /**
+     * Create a new configuration
+     */
+    constructor() {
+        const ret = wasm.wasmembedderconfig_new();
+        this.__wbg_ptr = ret >>> 0;
+        WasmEmbedderConfigFinalization.register(this, this.__wbg_ptr, this);
+        return this;
+    }
+}
+if (Symbol.dispose) WasmEmbedderConfig.prototype[Symbol.dispose] = WasmEmbedderConfig.prototype.free;
+exports.WasmEmbedderConfig = WasmEmbedderConfig;
+
+/**
+ * Compute cosine similarity between two embedding vectors (JS-friendly)
+ * @param {Float32Array} a
+ * @param {Float32Array} b
+ * @returns {number}
+ */
+function cosineSimilarity(a, b) {
+    const ptr0 = passArrayF32ToWasm0(a, wasm.__wbindgen_malloc);
+    const len0 = WASM_VECTOR_LEN;
+    const ptr1 = passArrayF32ToWasm0(b, wasm.__wbindgen_malloc);
+    const len1 = WASM_VECTOR_LEN;
+    const ret = wasm.cosineSimilarity(ptr0, len0, ptr1, len1);
+    return ret;
+}
+exports.cosineSimilarity = cosineSimilarity;
+
+/**
+ * Initialize panic hook for better error messages in WASM
+ */
+function init() {
+    wasm.init();
+}
+exports.init = init;
+
+/**
+ * L2 normalize an embedding vector (JS-friendly)
+ * @param {Float32Array} embedding
+ * @returns {Float32Array}
+ */
+function normalizeL2(embedding) {
+    const ptr0 = passArrayF32ToWasm0(embedding, wasm.__wbindgen_malloc);
+    const len0 = WASM_VECTOR_LEN;
+    const ret = wasm.normalizeL2(ptr0, len0);
+    var v2 = getArrayF32FromWasm0(ret[0], ret[1]).slice();
+    wasm.__wbindgen_free(ret[0], ret[1] * 4, 4);
+    return v2;
+}
+exports.normalizeL2 = normalizeL2;
+
+/**
+ * Check if SIMD is available (for performance info)
+ * Returns true if compiled with WASM SIMD128 support
+ * @returns {boolean}
+ */
+function simd_available() {
+    const ret = wasm.simd_available();
+    return ret !== 0;
+}
+exports.simd_available = simd_available;
+
+/**
+ * Get the library version
+ * @returns {string}
+ */
+function version() {
+    let deferred1_0;
+    let deferred1_1;
+    try {
+        const ret = wasm.version();
+        deferred1_0 = ret[0];
+        deferred1_1 = ret[1];
+        return getStringFromWasm0(ret[0], ret[1]);
+    } finally {
+        wasm.__wbindgen_free(deferred1_0, deferred1_1, 1);
+    }
+}
+exports.version = version;
+
+exports.__wbg___wbindgen_is_function_8d400b8b1af978cd = function(arg0) {
+    const ret = typeof(arg0) === 'function';
+    return ret;
+};
+
+exports.__wbg___wbindgen_is_object_ce774f3490692386 = function(arg0) {
+    const val = arg0;
+    const ret = typeof(val) === 'object' && val !== null;
+    return ret;
+};
+
+exports.__wbg___wbindgen_is_string_704ef9c8fc131030 = function(arg0) {
+    const ret = typeof(arg0) === 'string';
+    return ret;
+};
+
+exports.__wbg___wbindgen_is_undefined_f6b95eab589e0269 = function(arg0) {
+    const ret = arg0 === undefined;
+    return ret;
+};
+
+exports.__wbg___wbindgen_string_get_a2a31e16edf96e42 = function(arg0, arg1) {
+    const obj = arg1;
+    const ret = typeof(obj) === 'string' ? obj : undefined;
+    var ptr1 = isLikeNone(ret) ? 0 : passStringToWasm0(ret, wasm.__wbindgen_malloc, wasm.__wbindgen_realloc);
+    var len1 = WASM_VECTOR_LEN;
+    getDataViewMemory0().setInt32(arg0 + 4 * 1, len1, true);
+    getDataViewMemory0().setInt32(arg0 + 4 * 0, ptr1, true);
+};
+
+exports.__wbg___wbindgen_throw_dd24417ed36fc46e = function(arg0, arg1) {
+    throw new Error(getStringFromWasm0(arg0, arg1));
+};
+
+exports.__wbg_call_3020136f7a2d6e44 = function() { return handleError(function (arg0, arg1, arg2) {
+    const ret = arg0.call(arg1, arg2);
+    return ret;
+}, arguments) };
+
+exports.__wbg_call_abb4ff46ce38be40 = function() { return handleError(function (arg0, arg1) {
+    const ret = arg0.call(arg1);
+    return ret;
+}, arguments) };
+
+exports.__wbg_crypto_574e78ad8b13b65f = function(arg0) {
+    const ret = arg0.crypto;
+    return ret;
+};
+
+exports.__wbg_error_7534b8e9a36f1ab4 = function(arg0, arg1) {
+    let deferred0_0;
+    let deferred0_1;
+    try {
+        deferred0_0 = arg0;
+        deferred0_1 = arg1;
+        console.error(getStringFromWasm0(arg0, arg1));
+    } finally {
+        wasm.__wbindgen_free(deferred0_0, deferred0_1, 1);
+    }
+};
+
+exports.__wbg_getRandomValues_b8f5dbd5f3995a9e = function() { return handleError(function (arg0, arg1) {
+    arg0.getRandomValues(arg1);
+}, arguments) };
+
+exports.__wbg_length_22ac23eaec9d8053 = function(arg0) {
+    const ret = arg0.length;
+    return ret;
+};
+
+exports.__wbg_msCrypto_a61aeb35a24c1329 = function(arg0) {
+    const ret = arg0.msCrypto;
+    return ret;
+};
+
+exports.__wbg_new_8a6f238a6ece86ea = function() {
+    const ret = new Error();
+    return ret;
+};
+
+exports.__wbg_new_no_args_cb138f77cf6151ee = function(arg0, arg1) {
+    const ret = new Function(getStringFromWasm0(arg0, arg1));
+    return ret;
+};
+
+exports.__wbg_new_with_length_aa5eaf41d35235e5 = function(arg0) {
+    const ret = new Uint8Array(arg0 >>> 0);
+    return ret;
+};
+
+exports.__wbg_node_905d3e251edff8a2 = function(arg0) {
+    const ret = arg0.node;
+    return ret;
+};
+
+exports.__wbg_process_dc0fbacc7c1c06f7 = function(arg0) {
+    const ret = arg0.process;
+    return ret;
+};
+
+exports.__wbg_prototypesetcall_dfe9b766cdc1f1fd = function(arg0, arg1, arg2) {
+    Uint8Array.prototype.set.call(getArrayU8FromWasm0(arg0, arg1), arg2);
+};
+
+exports.__wbg_randomFillSync_ac0988aba3254290 = function() { return handleError(function (arg0, arg1) {
+    arg0.randomFillSync(arg1);
+}, arguments) };
+
+exports.__wbg_require_60cc747a6bc5215a = function() { return handleError(function () {
+    const ret = module.require;
+    return ret;
+}, arguments) };
+
+exports.__wbg_stack_0ed75d68575b0f3c = function(arg0, arg1) {
+    const ret = arg1.stack;
+    const ptr1 = passStringToWasm0(ret, wasm.__wbindgen_malloc, wasm.__wbindgen_realloc);
+    const len1 = WASM_VECTOR_LEN;
+    getDataViewMemory0().setInt32(arg0 + 4 * 1, len1, true);
+    getDataViewMemory0().setInt32(arg0 + 4 * 0, ptr1, true);
+};
+
+exports.__wbg_static_accessor_GLOBAL_769e6b65d6557335 = function() {
+    const ret = typeof global === 'undefined' ? null : global;
+    return isLikeNone(ret) ? 0 : addToExternrefTable0(ret);
+};
+
+exports.__wbg_static_accessor_GLOBAL_THIS_60cf02db4de8e1c1 = function() {
+    const ret = typeof globalThis === 'undefined' ? null : globalThis;
+    return isLikeNone(ret) ? 0 : addToExternrefTable0(ret);
+};
+
+exports.__wbg_static_accessor_SELF_08f5a74c69739274 = function() {
+    const ret = typeof self === 'undefined' ? null : self;
+    return isLikeNone(ret) ? 0 : addToExternrefTable0(ret);
+};
+
+exports.__wbg_static_accessor_WINDOW_a8924b26aa92d024 = function() {
+    const ret = typeof window === 'undefined' ? null : window;
+    return isLikeNone(ret) ? 0 : addToExternrefTable0(ret);
+};
+
+exports.__wbg_subarray_845f2f5bce7d061a = function(arg0, arg1, arg2) {
+    const ret = arg0.subarray(arg1 >>> 0, arg2 >>> 0);
+    return ret;
+};
+
+exports.__wbg_versions_c01dfd4722a88165 = function(arg0) {
+    const ret = arg0.versions;
+    return ret;
+};
+
+exports.__wbindgen_cast_2241b6af4c4b2941 = function(arg0, arg1) {
+    // Cast intrinsic for `Ref(String) -> Externref`.
+    const ret = getStringFromWasm0(arg0, arg1);
+    return ret;
+};
+
+exports.__wbindgen_cast_cb9088102bce6b30 = function(arg0, arg1) {
+    // Cast intrinsic for `Ref(Slice(U8)) -> NamedExternref("Uint8Array")`.
+    const ret = getArrayU8FromWasm0(arg0, arg1);
+    return ret;
+};
+
+exports.__wbindgen_init_externref_table = function() {
+    const table = wasm.__wbindgen_externrefs;
+    const offset = table.grow(4);
+    table.set(0, undefined);
+    table.set(offset + 0, undefined);
+    table.set(offset + 1, null);
+    table.set(offset + 2, true);
+    table.set(offset + 3, false);
+};
+
+const wasmPath = `${__dirname}/ruvector_onnx_embeddings_wasm_bg.wasm`;
+const wasmBytes = require('fs').readFileSync(wasmPath);
+const wasmModule = new WebAssembly.Module(wasmBytes);
+const wasm = exports.__wasm = new WebAssembly.Instance(wasmModule, imports).exports;
+
+wasm.__wbindgen_start();