@epfml/discojs 3.0.1-p20250729132444.0 → 3.0.1-p20250924113522.0

package/dist/validator.js

@@ -7,20 +7,21 @@ export class Validator {
         this.#model = model;
     }
     /** infer every line of the dataset and check that it is as labelled */
-    async test(dataset) {
-        const preprocessed = await processing.preprocess(this.task, dataset);
+    test(dataset) {
+        const preprocessed = processing.preprocess(this.task, dataset);
         const batched = preprocessed.batch(this.task.trainingInformation.batchSize);
         const predictionWithTruth = batched
             .map(async (batch) => (await this.#model.predict(batch.map(([inputs, _]) => inputs))).zip(batch.map(([_, outputs]) => outputs)))
             .flatten();
-        return predictionWithTruth.map(async ([predicted, truth]) => ({
-            predicted: await processing.postprocess(this.task, predicted),
-            truth: await processing.postprocess(this.task, truth),
+        return predictionWithTruth.map(([predicted, truth]) => ({
+            predicted: processing.postprocess(this.task, predicted),
+            truth: processing.postprocess(this.task, truth),
         }));
     }
     /** use the model to predict every line of the dataset */
     async *infer(dataset) {
-        const modelPredictions = (await processing.preprocessWithoutLabel(this.task, dataset))
+        const modelPredictions = processing
+            .preprocessWithoutLabel(this.task, dataset)
             .batch(this.task.trainingInformation.batchSize)
             .map((batch) => this.#model.predict(batch))
             .flatten();

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@epfml/discojs",
-  "version": "3.0.1-p20250729132444.0",
+  "version": "3.0.1-p20250924113522.0",
   "type": "module",
   "main": "dist/index.js",
   "types": "dist/index.d.ts",
@@ -8,7 +8,7 @@
     "watch": "nodemon --ext ts --ignore dist --exec npm run",
     "build": "tsc",
     "lint": "npx eslint .",
-    "test": "mocha"
+    "test": "cd .. && vitest --run --project=discojs"
   },
   "repository": {
     "type": "git",
@@ -28,15 +28,12 @@
     "isomorphic-ws": "5",
     "simple-peer": "9",
     "tslib": "2",
-    "ws": "8"
+    "ws": "8",
+    "zod": "3"
   },
   "devDependencies": {
     "@tensorflow/tfjs-node": "4",
-    "@types/chai": "5",
-    "@types/mocha": "10",
     "@types/simple-peer": "9",
-    "chai": "5",
-    "mocha": "10",
     "nodemon": "3",
     "ts-node": "10"
   }

package/dist/processing/text.d.ts

@@ -1,21 +0,0 @@
-import { PreTrainedTokenizer } from "@xenova/transformers";
-import type { Text, TokenizedText } from '../index.js';
-interface TokenizingConfig {
-    padding?: boolean;
-    padding_side?: 'left' | 'right';
-    truncation?: boolean;
-    max_length?: number;
-}
-/**
- * Tokenize one line of text.
- * Wrapper around Transformers.js tokenizer to handle type checking and format the output.
- * Note that Transformers.js's tokenizer can tokenize multiple lines of text at once
- * but we are currently not making use of it. Can be useful when padding a batch
- *
- * @param tokenizer the tokenizer object
- * @param text the text to tokenize
- * @param config TokenizingConfig, the tokenizing parameters when using `tokenizer`
- * @returns List<number> the tokenized text
- */
-export declare function tokenize(tokenizer: PreTrainedTokenizer, text: Text, config?: TokenizingConfig): TokenizedText;
-export {};

package/dist/processing/text.js

@@ -1,36 +0,0 @@
-import { List } from "immutable";
-function isArrayOfNumber(raw) {
-    return Array.isArray(raw) && raw.every((e) => typeof e === "number");
-}
-/**
- * Tokenize one line of text.
- * Wrapper around Transformers.js tokenizer to handle type checking and format the output.
- * Note that Transformers.js's tokenizer can tokenize multiple lines of text at once
- * but we are currently not making use of it. Can be useful when padding a batch
- *
- * @param tokenizer the tokenizer object
- * @param text the text to tokenize
- * @param config TokenizingConfig, the tokenizing parameters when using `tokenizer`
- * @returns List<number> the tokenized text
- */
-export function tokenize(tokenizer, text, config) {
-    config = { ...config }; // create a config if undefined
-    if (config.padding || config.truncation) {
-        if (config.max_length === undefined)
-            throw new Error("max_length needs to be specified to use padding or truncation");
-        if (!Number.isInteger(config.max_length))
-            throw new Error("max_length should be an integer");
-    }
-    if (config.padding) {
-        // The padding side is set as an attribute, not in the config
-        tokenizer.padding_side = config.padding_side ?? 'left';
-        config.truncation = true; // for a single sequence, padding implies truncation to max_length
-    }
-    const tokenizerResult = tokenizer(text, { ...config, return_tensor: false });
-    if (typeof tokenizerResult !== "object" ||
-        tokenizerResult === null ||
-        !("input_ids" in tokenizerResult) ||
-        !isArrayOfNumber(tokenizerResult.input_ids))
-        throw new Error("tokenizer returned unexpected type");
-    return List(tokenizerResult.input_ids);
-}

package/dist/task/data_example.d.ts

@@ -1,5 +0,0 @@
-export interface DataExample {
-    columnName: string;
-    columnData: string | number;
-}
-export declare function isDataExample(raw: unknown): raw is DataExample;

package/dist/task/data_example.js

@@ -1,14 +0,0 @@
-export function isDataExample(raw) {
-    if (typeof raw !== 'object' || raw === null) {
-        return false;
-    }
-    const { columnName, columnData } = raw;
-    if (typeof columnName !== 'string' ||
-        (typeof columnData !== 'string' && typeof columnData !== 'number')) {
-        return false;
-    }
-    const repack = { columnName, columnData };
-    const _correct = repack;
-    const _total = repack;
-    return true;
-}

package/dist/task/summary.d.ts

@@ -1,5 +0,0 @@
-export interface Summary {
-    preview: string;
-    overview: string;
-}
-export declare function isSummary(raw: unknown): raw is Summary;

package/dist/task/summary.js

@@ -1,13 +0,0 @@
-export function isSummary(raw) {
-    if (typeof raw !== 'object' || raw === null) {
-        return false;
-    }
-    const { preview, overview } = raw;
-    if (!(typeof preview === 'string' && typeof overview === 'string')) {
-        return false;
-    }
-    const repack = { preview, overview };
-    const _correct = repack;
-    const _total = repack;
-    return true;
-}