Mask filter plugin for Embulk

Mask columns with asterisks in a variety of patterns (still in initial development phase and missing basic features to use in production).

Overview

• Plugin type: filter

Configuration

Caution : Now we use type to specify mask types such as all and email, instead of pattern which was used in version 0.1.1 or earlier.

• columns: target columns which would be replaced with asterisks (string, required)
  • name: name of the column (string, required)
  • type: mask type, all, email, regex or substring (string, default: all)
  • paths: list of JSON path and type, works if the column type is JSON
    • [{key: $.json_path1}, {key: $.json_path2}] would mask both $.json_path1 and $.json_path2 nodes
    • Elements under the nodes would be converted to string and then masked (e.g., [0,1,2] -> *******)
  • length: if specified, this filter replaces the column with fixed number of asterisks (integer, optional. supported only in all, email, substring.)
  • pattern: Regex pattern such as "[0-9]+" (string, required for regex type)
  • start: The beginning index for substring type. The value starts from 0 and inclusive (integer, default: 0)
  • end: The ending index for substring type. The value is exclusive (integer, default: length of the target column)

Example

If you have below data in csv or other format file,

first_name	last_name	gender	age	contact
Benjamin	Bell	male	30	bell.benjamin_dummy@example.com
Lucas	Duncan	male	20	lucas.duncan_dummy@example.com
Elizabeth	May	female	25	elizabeth.may_dummy@example.com
Christian	Reid	male	15	christian.reid_dummy@example.com
Amy	Avery	female	40	amy.avercy_dummy@example.com

below filter configuration

filters:
  - type: mask
    columns:
      - { name: last_name}
      - { name: age}
      - { name: contact, type: email, length: 5}

would produce

first_name	last_name	gender	age	contact
Benjamin	****	male	**	*****@example.com
Lucas	******	male	**	*****@example.com
Elizabeth	***	female	**	*****@example.com
Christian	****	male	**	*****@example.com
Amy	*****	female	**	*****@example.com

If you use regex and/or substring types,

filters:
  - type: mask
    columns:
      - { name: first_name, type: regex, pattern: "[a-z]"}
      - { name: contact, type: substring, start: 5, length: 5}

would produce

first_name	last_name	gender	age	contact
B*******	Bell	male	30	bell.*****
L****	Duncan	male	20	lucas*****
E*******	May	female	25	eliza*****
C********	Reid	male	15	chris*****
A**	Avery	female	40	amy.a*****

JSON type column is also partially supported.

If you have a user column with this JSON data structure

{
  "full_name": {
    "first_name": "Benjamin",
    "last_name": "Bell"
  },
  "gender": "male",
  "age": 30,
  "email": "[email protected]"
}

below filter configuration

filters:
  - type: mask
    columns:
      - { name: user, paths: [{key: $.full_name.first_name}, {key: $.email, type: email}]}    

would produce

{
  "full_name": {
    "first_name": "********",
    "last_name": "Bell"
  },
  "gender": "male",
  "age": 30,
  "email": "*********@example.com"
}

Build

$ ./gradlew gem  # -t to watch change of files and rebuild continuously