redact-phi

redact-PHI

A command-line utility to redact and fabricate PHI, PII or protected data (CSV, JSON, XLSX).

Installation

Install globally using npm i -g redact-phi

Usage

% redact --help
Usage: redact [options] <infile> [outfile]

Arguments:
  infile                   File to redact
  outfile                  Output file, defaults to {input_filename_redacted.csv}

Options:
  -V, --version            output the version number
  --delimiter <delimiter>  sets delimiter type (default: ",")
  --override <override>    Provide the location of your redaction specification
  -h, --help               display help for command

Redaction Example

redact-PHI includes an example (./example/) :

• example.csv - CSV source file (data to be redacted)
• example.json - JSON redaction strategy (defines how columns will be redacted with: faker, a custom redactor or a constant)
• example.js - JavaScript custom redactors (this file is optional)

The .csv, .json and .js files must all be named the same (e.g. data.csv, data.json, data.js)

---

example/example.csv :

id,email,ssn,first name,last name,full address,order date,order amount
783,[e-mail],706-25-4558,Caroline,Goodwin,82703 Yasmeen Corner Apt. 379,"10/9/2020, 4:48:52 AM",252.94
784,[e-mail],282-57-6226,Izabella,Bosco,2930 Alisa Heights Apt. 572,"3/27/2021, 11:51:26 PM",689.40
784,[e-mail],282-57-6226,Izabella,Bosco,2930 Alisa Heights Apt. 572,"6/17/2021, 9:44:54 PM",446.29
786,[e-mail],677-86-2303,Celine,Dicki,68305 Labadie Shoal Suite 608,"4/23/2021, 6:49:34 PM",889.37
...

---

example/example.json :

{
  "columns": [
    {
      "redactWith": "random.number",
      "columnNum": "",
      "columnKey": "order id",
      "tracked": false
    },
    {
      "redactWith": "customGenerateId",
      "columnNum": 0,
      "columnKey": "",
      "tracked": true
    },
...

The columns array contains objects with properties that describe how each column should be redacted:

• columnNum - (integer) Identifies the column to be redacted with an integer index (zero-based). (columnNum or columnKey must be set)

• columnKey - (string) Identifies the column to be redacted using the file's header. (columnNum or columnKey must be set)

• redactWith - (string) The strategy used to redact a value, there are four options:

  • faker function - Available functions : name.firstName or internet.email
  • faker template - Faker methods in a mustache template : {{address.streetAddress(true)}} or {{address.city}}, {{address.state}} {{address.zip}}
  • custom JavaScript - For more complicated redacted values you can call a JavaScript function defined in your .js file (see example.js)
  • constant value - If you wish to replace every value in this column with a constant, e.g. John Doe

• tracked - Tracking preserves the relationships within your data while de-identifying. With tracking enabled, the redaction engine tracks a column's original value and reuses the redacted value if the original is re-encountered. To see this in-action, run redact /example/example.csv and note how columns with tracking enabled (id,ssn,email) are redacted with the same value in example_redacted.csv.

---

example/example.js :

module.exports = {
    customGenerateId: () => {
        return ++currentId;
    },
...

Custom redactors provide more control over your data. A custom redactor is referenced from your .json file using the redactWith property. For example: redactWith: "customGenerateId"

JSON schema for creating a redaction strategy

JSON Schema for creating a spec