base-data-manager/util/scrub_primitive.ts

/**Scrubs a primitive the same as scrub.jq
 * @todo Keep in sync with scrub.jq, as ideally they'd just be the same code... but diff
 * languages ugh*/
export function scrubPrimitive(value: unknown): unknown {
  if (typeof value === "string") {
    if (/^(([0-9]{1,3}\.){3}[0-9]{1,3})$/.test(value)) {
      // IPv4
      return "1.1.1.1";
    }
    else if (/^([0-9a-fA-F]{0,4}:){2,7}[0-9a-fA-F]{0,4}$/.test(value)) {
      // IPv6
      return "2000:0000:0000:0000:0000:0000:0000:0000";
    }
    else if (/^[a-zA-Z0-9._%+-]+@[a-zA-Z0-9.-]+\.[a-zA-Z]{2,}$/.test(value)) {
      // Email-like
      return "not_a_real_email@example.com";
    }
    else if (/\.(jpg|jpeg|png|gif|bmp|webp|svg|ico|tiff|mp3|wav|flac|aac|ogg|wma|m4a|mp4|avi|mkv|mov|wmv|flv|webm)$/i.test(value)) {
      // Leave these alone, you will have to manually go through these later and replace with
      // placeholders
      console.error(`MANUAL REPAIR NEEDED: ${value}`);
      return value;
    }
    else if (/^[0-9]{4}-[0-9]{2}-[0-9]{2}T[0-9]{2}:[0-9]{2}:[0-9]{2}[+-][0-9]{2}:[0-9]{2}$/.test(value)) {
      // iso date time without millis with timezone
      return "2020-04-13T10:09:08+00:00";
    }
    else if (/^[0-9]{4}-[0-9]{2}-[0-9]{2}T[0-9]{2}:[0-9]{2}:[0-9]{2}(\.[0-9]{1,6})?[+-][0-9]{2}:[0-9]{2}$/.test(value)) {
      // iso date time with millis with timezone
      return "2020-04-13T10:09:08.000000+00:00";
    }
    else if (/^[0-9]{4}-[0-9]{2}-[0-9]{2}T[0-9]{2}:[0-9]{2}:[0-9]{2}(\.[0-9]{1,6})?$/.test(value)) {
      // iso date time with millis with timezone
      return "2020-04-13T10:09:08.000";
    }
    else if (/^[0-9]{4}-[0-9]{2}-[0-9]{2}T[0-9]{2}:[0-9]{2}:[0-9]{2}Z$/.test(value)) {
      // iso date time with z and the end (from fitbit export)
      return "2020-04-13T10:09:08Z";
    }
    else if (/^[0-9]{4}-[0-9]{2}-[0-9]{2}T[0-9]{2}:[0-9]{2}:[0-9]{2}\.[0-9]{1,6}Z$/.test(value)) {
      // iso date time with z and the end (from fitbit export)
      return "2020-04-13T10:09:08.000000Z";
    }
    else if (/^[0-9]{4}-[0-9]{2}-[0-9]{2}T[0-9]{2}:[0-9]{2}$/.test(value)) {
      // iso date time no seconds (from fitbit export)
      return "2020-04-13T10:09";
    }
    else if (/^[0-9]{4}-[0-9]{2}-[0-9]{2} [0-9]{2}:[0-9]{2}:[0-9]{2} - [0-9]{4}-[0-9]{2}-[0-9]{2} [0-9]{2}:[0-9]{2}:[0-9]{2}$/.test(value)) {
      // iso date time range (from fitbit export)
      return "2020-04-13 10:09:08 - 2020-04-13 10:09:08";
    }
    else if (/^[0-9]{4}-[0-9]{2}-[0-9]{2} [0-9]{2}:[0-9]{2}:[0-9]{2}(\.[0-9]{1,6})?[+-][0-9]{4}$/.test(value)) {
      // iso date time with millis with timezone no colon, no T (from fitbit export)
      return "2020-04-13 10:09:08+0000";
    }
    else if (/^[0-9]{4}-[0-9]{2}-[0-9]{2}$/.test(value)) {
      //  just date
      return "2020-04-13";
    }
    else if (/^[0-9]{4}-[0-9]{2}-[0-9]{2} [0-9]{2}:[0-9]{2}:[0-9]{2} UTC$/.test(value)) {
      // Date format from snapchat export
      return "2020-04-13 10:09:08 UTC";
    }
    else if (/^[0-9]{4}-[0-9]{2}-[0-9]{2} [0-9]{2}:[0-9]{2}:[0-9]{2}$/.test(value)) {
      // Date format from snapchat export
      return "2020-04-13 10:09:08";
    }
    else if (/^[0-9]{2}\/[0-9]{2}\/[0-9]{2} [0-9]{2}:[0-9]{2}:[0-9]{2}$/.test(value)) {
      // Date format from fitbit export (MM/DD/YY)
      return "04/13/20 10:09:08";
    }
    else if (/^\w{3} \w{3} \d{2} \d{2}:\d{2}:\d{2} UTC \d{4}$/.test(value)) {
      // Date format from fitbit export (uughhh)
      return "Mon Apr 13 10:09:08 UTC 2020";
    }
    else if (/^\+[0-9]{2}:[0-2]{2}$/.test(value)) {
      // UTC offset (from fitbit export)
      return "+00:00";
    }
    else if (/^-[0-9]{2}:[0-2]{2}$/.test(value)) {
      // UTC offset (from fitbit export)
      return "-00:00"
    }
    else if (/^[0-9]+$/.test(value)) {
      // preserve length of the string
      return "1".repeat(value.length);
    }
    else if (/^-?[0-9]*(\.[0-9]*)?$/.test(value) && /[0-9]/.test(value)) {
      // Decimal string - preserve shape
      return value.replace(/[0-9]/g, "1");
    }
    else if (/^[0-9a-fA-F]+$/.test(value)) {
      // hexadecimal string - repeat the hex pattern and truncate to original length
      return "a1".repeat(value.length)
        .slice(0, value.length);
    }
    else if (value === "") {
      // prevents empty string from just returning null instead of empty string
      return "";
    }
    else if (value === "true" || value === "false") {
      return "false";
    }
    else if (/:\/\//.test(value)) {
      return "url://somewhere";
    }
    else if (/\//.test(value)) {
      return "some/path";
    }
    else if (value === "null") {
      return "null";
    }
    else {
      // Preserve string length for other strings
      return "x".repeat(value.length);
    }
  }
  else if (typeof value === "number") {
    if (946702800 <= value && value <= 1893474000) {
      // Take modulo 1 year to get variance in the output, then add offset to bring to ~2024
      return Math.floor(((value % 31557600) + 1704067200) / 5000) * 5000;
    }
    else if (Number.isInteger(value)) {
      // Integer - preserve digit count
      const len = value.toString()
        .length;
      return Number("1".repeat(len));
    }
    else {
      // Decimal - preserve shape, sign, and leading zero
      const sign = value < 0 ? "-" : "";
      const parts = Math.abs(value).toString().split(".");
      const intPart = parts[0] === "0" ? "0" : "1".repeat(parts[0].length);
      const fracPart = "1".repeat(parts[1].length);
      return Number(`${sign}${intPart}.${fracPart}`);
    }
  }
  else if (typeof value === "boolean") {
    // Replace all booleans with false, this can give sensitive info away based
    // on what the key was in the data
    return false;
  }
  else {
    return value;
  }
}