Computes the Shannon entropy for a given input file

shannon-entropy-for-a-file.js

const fs = require('fs');

const updateOccurenceTable = (byteOccurenceTable, inputBuffer) => {
  inputBuffer.forEach(character => {
    const byteOccurenceAccumulator = byteOccurenceTable[character];
    byteOccurenceTable[character] = byteOccurenceAccumulator !== undefined ? byteOccurenceAccumulator + 1 : 1;
  });

  const { total } = byteOccurenceTable;
  byteOccurenceTable.total = total ? total + inputBuffer.length : inputBuffer.length;
}

let filePath = process.argv[2];
if (!filePath) {
  console.log('No input file path specified. Please provide a file path.');
  process.exit(-1);
}

fs.readFile(filePath, {}, (err, chunk) => {
  let occurenceTable = { total: 0 };
  updateOccurenceTable(occurenceTable, chunk);

  let entropy = 0.0;
  const { total } = occurenceTable;
  Object.keys(occurenceTable).forEach(item => {
    if (item !== 'total') {
      const itemCount = occurenceTable[item];
      const probability = itemCount / total;
      if (probability > 0) {
        const entropyComponent = probability * Math.log(probability, 2);
        entropy = entropy + entropyComponent;
      }
    }
  })
  console.log(-entropy);
});