A powerful, headless data ingestion engine for tabular data (CSV and Excel) in JavaScript and TypeScript.

Ingestx is designed to handle massive files smoothly by parsing in chunks, offering built-in data validation, schema mapping, and execution control (pause, resume, cancel). Since it's completely headless, you bring your own UI and we handle the heavy lifting!

• 📊 Format Support: Seamlessly process CSV and Excel (.xlsx, .xls) files.
• ⚡ Chunk-based Processing: Prevents browser freezes and memory limits by processing huge files in manageable chunks.
• 🔍 Robust Validation: Define schemas with strict types (string, number, boolean), regex rules, min/max limits, and custom validation logic.
• 🔀 Smart Column Mapping: Automatically map variations of column headers (e.g., email, Email Address, User_Email) to a single key.
• ⏯️ Execution Control: Pause, resume, and cancel the ingestion process on the fly.
• ⚛️ React Ready: Comes with a built-in useIngestion hook for effortless React integration.

npm install @parallelbytes/ingestx papaparse xlsx
# or
yarn add @parallelbytes/ingestx papaparse xlsx

(Note: papaparse and xlsx are peer dependencies used under the hood for parsing).

Using the useIngestion hook is the fastest way to get started in a React application.

import { useIngestion } from 'ingestx/react';
import type { ColumnConfig } from 'ingestx';

const columnConfigs: ColumnConfig[] = [
  { key: 'id', displayNames: ['id', 'user id'], type: 'number', validationRequired: true },
  { key: 'email', displayNames: ['email', 'email address'], type: 'string', validationRequired: true },
  { 
    key: 'isActive', 
    displayNames: ['active', 'is active'], 
    type: 'boolean', 
    trueValues: ['yes', 'true'], 
    falseValues: ['no', 'false'] 
  }
];

export default function Uploader() {
  const {
    isProcessing,
    progress,
    result,
    startIngestion,
    pause,
    resume,
    cancel
  } = useIngestion({
    columnConfigs,
    chunkSize: 1000,
  });

  const handleFileUpload = (e: React.ChangeEvent<HTMLInputElement>) => {
    const file = e.target.files?.[0];
    if (file) startIngestion(file);
  };

  return (
    <div>
      <input type="file" onChange={handleFileUpload} accept=".csv, .xlsx" />
      
      {isProcessing && <p>Processing... {progress.toFixed(0)}%</p>}
      
      {result && (
        <div>
          <p>✅ Valid Rows: {result.validRowsCount}</p>
          <p>❌ Invalid Rows: {result.invalidRowsCount}</p>
        </div>
      )}
    </div>
  );
}

The heart of Ingestx is the schema definition. You define exactly what your data should look like.

• String: regex, allowedValues
• Number: min, max, allowedValues
• Boolean: trueValues, falseValues (Strictly maps specific strings to booleans).

You can configure global behavior when initializing the ingestion:

const options = {
  trimValues: true,             // Trims whitespace from all cell values
  trimHeaders: true,            // Trims whitespace from column headers
  caseInsensitiveHeaders: true, // Matches headers ignoring case
  shouldAccumulateResult: true  // If false, results are flushed per chunk (useful for massive datasets to save memory)
}

When ingestion completes (or pauses), the result object contains:

• validRows: Array of cleanly parsed and mapped objects.
• invalidRows: Array of raw objects that failed validation.
• errorsData: Detailed row-wise and column-wise error messages indicating exactly why a row failed.

MIT