audit-labs/audit-tools
A collection of scripts, queries, and other goodies you can use in an audit.
clone: git clone https://gitbay.org/audit-labs/audit-tools.git
60c4938f3825f0a1253f1e6eb9a55fb330ed9d80
verified · cmc
author: Christian Cleberg <hello@cleberg.net> · 2025-01-16T20:21:04Z
sampling/README.org | 32 +++++++++ sampling/sample-html.png | Bin 0 -> 217862 bytes sampling/sample.html | 167 +++++++++++++++++++++++++++++++++++++++++++++++ 3 files changed, 199 insertions(+) new file mode 100644 @@ -0,0 +1,32 @@ +#+title: Sampling Tools + +* =sample.py= + +#+begin_src sh +python ./sample.py +#+end_src + +#+begin_src text +Dataframe size (rows, columns): (100, 9) +Sample size: 5 +Sample: + Index Organization Id ... Industry Number of employees +79 80 cBa7EFe5D05Adaf ... Online Publishing 7805 +97 98 E7df80C60Abd7f9 ... Broadcast Media 236 +3 4 2bFC1Be8a4ce42f ... Automotive 921 +42 43 A2D89Ab9bCcAd4e ... Capital Markets / Hedge Fund / Private Equity 3816 +70 71 32BB9Ff4d939788 ... Wireless 6146 + +[5 rows x 9 columns] +#+end_src + +* =sample.html= + +This is an interactive web page that allows users to submit their population +size, sample size(s), and generate a psuedo-random sample list of numbers to use +when sampling against their population. + +Samples can be re-generated and validated using the seed numbers provided during +the original generation. + +[[sample-html.png]] new file mode 100644 Binary files /dev/null and b/sampling/sample-html.png differ new file mode 100644 @@ -0,0 +1,167 @@ +<!doctype html> +<html lang="en"> +<head> + <meta charset="utf-8"> + <meta name="viewport" content="width=device-width, initial-scale=1"> + <title>Sampling Tool</title> + <style> + form { + display: table; + } + form div { + display: table-row; + } + label, input { + display: table-cell; + margin-bottom: 10px; + } + label { + padding-right: 10px; + } + table { + width: 100%; + border-collapse: collapse; + margin-top: 10px; + } + table, th, td { + border: 1px solid black; + } + th, td { + padding: 5px; + text-align: center; + } + </style> +</head> +<body> +<h1>Sampling Tool</h1> +<p>This sampling tool provides a quick and easy way to generate +psuedo-random samples within a defined range. Simply enter the size of your +population, enter your desired number of samples, and generate!</p> +<p>To reproduce and validate a previously-generated sample, please ensure +you have entered the seed value correctly when submitting the form. All +inputs must match the previously-generated sample's inputs to generate the +same output.</p> +<form id="sampleForm" onsubmit="return handleFormSubmit(event)"> + <div> + <label for="populationSize">Population Size*:</label> + <input type="number" id="populationSize" name="populationSize" min="1" autofocus required> + </div> + <div> + <label for="sampleSize">Sample Size*:</label> + <input type="number" id="sampleSize" name="sampleSize" min="1" required> + </div> + <div> + <label for="replacementSize">Replacement Sample Size:</label> + <input type="number" id="replacementSize" name="replacementSize" min="0"> + </div> + <div> + <label for="customSeed">Custom Seed (optional):</label> + <input type="number" id="customSeed" name="customSeed" min="0"> + </div> + <div> + <input type="submit" value="Calculate"> + </div> + <p><i>* Indicates a required field</i></p> +</form> +<hr> +<div id="results"></div> +<script> +function seededRandom(seed) { + let s = seed % 2147483647; + return function() { + s = (s * 16807) % 2147483647; + return (s - 1) / 2147483646; + }; +} + +function handleFormSubmit(event) { + event.preventDefault(); // Prevent the default form submission behavior + const customSeedInput = document.getElementById('customSeed').value; + const seed = customSeedInput ? parseInt(customSeedInput) : Math.floor(Math.random() * 1000000); // Use custom seed if provided + generateSamples(seed); // Call the function with the seed +} + +function generateSamples(seed) { + const populationSize = parseInt(document.getElementById('populationSize').value); + const sampleSize = parseInt(document.getElementById('sampleSize').value); + const replacementSize = parseInt(document.getElementById('replacementSize').value || 0); + const resultsDiv = document.getElementById('results'); + + // Clear previous results + resultsDiv.innerHTML = ''; + + // Validate inputs + if (isNaN(populationSize) || isNaN(sampleSize) || populationSize <= 0 || sampleSize <= 0) { + alert("Please enter valid numbers for required fields."); + return; + } + if (sampleSize + replacementSize > populationSize) { + alert("Not enough unique samples available. Reduce the sample size or replacement size."); + return; + } + + // Create seeded random function + const random = seededRandom(seed); + + // Generate an array of population numbers + const population = Array.from({ length: populationSize }, (_, i) => i + 1); + + // Shuffle the population array using the seeded random function + const shuffledPopulation = population + .map(value => ({ value, sort: random() })) + .sort((a, b) => a.sort - b.sort) + .map(({ value }) => value); + + // Select original samples + const originalSamples = shuffledPopulation.slice(0, sampleSize); + + // Select replacement samples + const replacementSamples = shuffledPopulation.slice(sampleSize, sampleSize + replacementSize); + + // Display the seed + const seedInfo = document.createElement('p'); + seedInfo.textContent = `Seed: ${seed}`; + resultsDiv.appendChild(seedInfo); + + // Generate tables + const originalTable = createTable(originalSamples, "Original Samples"); + const replacementTable = createTable(replacementSamples, "Replacement Samples"); + + // Append tables to the results div + resultsDiv.appendChild(originalTable); + if (replacementSamples.length > 0) { + resultsDiv.appendChild(replacementTable); + } +} + +function createTable(dataArray, tableTitle) { + const table = document.createElement('table'); + const caption = document.createElement('caption'); + caption.textContent = tableTitle; + table.appendChild(caption); + + const headerRow = document.createElement('tr'); + const indexHeader = document.createElement('th'); + indexHeader.textContent = 'Index'; + const sampleHeader = document.createElement('th'); + sampleHeader.textContent = 'Sample'; + headerRow.appendChild(indexHeader); + headerRow.appendChild(sampleHeader); + table.appendChild(headerRow); + + dataArray.forEach((sample, index) => { + const row = document.createElement('tr'); + const indexCell = document.createElement('td'); + const sampleCell = document.createElement('td'); + indexCell.textContent = index + 1; + sampleCell.textContent = sample; + row.appendChild(indexCell); + row.appendChild(sampleCell); + table.appendChild(row); + }); + + return table; +} +</script> +</body> +</html>