Skip to content
/ ecto_anon Public

Data anonymization for your Ecto models !

License

Notifications You must be signed in to change notification settings

WTTJ/ecto_anon

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

47 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

ecto_anon

Module Version Hex Docs Total Download License Last Updated

Simple way to handle data anonymization directly in your Ecto schemas


Table of Contents

Installation

Add :ecto_anon to your mix.exs dependencies:

def deps do
  [
    {:ecto_anon, "~> 0.6.0"}
  ]
end

Usage

Define an anon_schema in your schema module and specify every fields you want to anonymize (regular fields, associations, embeds):

defmodule User do
  use Ecto.Schema
  use EctoAnon.Schema

  anon_schema [
    :name,
    :email
  ]

  schema "users" do
    field :name, :string
    field :age, :integer
    field :email, :string

    anonymized()
  end
end

Then use EctoAnon.run to apply anonymization on desired resource

user = Repo.get(User, id)
%User{name: "jane", age: 24, email: "jane@email.com"}

EctoAnon.run(user, Repo)
{:ok, %User{name: "redacted", age: 24, email: "redacted"}}

Options

cascade

When set to true, it allows ecto_anon to preload and anonymize all associations (and associations of these associations) automatically in cascade. Could be used to anonymize all data related to a struct in a single call.

Note that this won't traverse belongs_to associations.

Default: false

defmodule User do
  use Ecto.Schema
  use EctoAnon.Schema

  anon_schema [
    :lastname,
    :email,
    :followers,
    :favorite_quote,
    :quotes,
    last_sign_in_at: [:anonymized_date, options: [:only_year]]
  ]

  schema "users" do
    field(:firstname, :string)
    field(:lastname, :string)
    field(:email, :string)
    field(:last_sign_in_at, :utc_datetime)

    has_many(:comments, Comment, foreign_key: :author_id, references: :id)
    embeds_one(:favorite_quote, Quote)
    embeds_many(:quotes, Quote)

    many_to_many(
      :followers,
      __MODULE__,
      join_through: Follower,
      join_keys: [follower_id: :id, followee_id: :id]
    )

    anonymized()
  end
end

defmodule Quote do
  use Ecto.Schema
  use EctoAnon.Schema

  anon_schema([
    :quote,
    :author
  ])

  embedded_schema do
    field(:quote, :string)
    field(:author, :string)
  end
end

defmodule Follower do
  use Ecto.Schema

  schema "followers" do
    field(:follower_id, :id)
    field(:followee_id, :id)
    timestamps()
  end
end
Repo.get(User, id)
|> EctoAnon.run(Repo, cascade: true)

{:ok,
   %User{
     email: "redacted",
     firstname: "John",
     last_sign_in_at: ~U[2022-01-01 00:00:00Z],
     lastname: "redacted",
     favorite_quote: %Quote{
       quote: "redacted",
       author: "redacted"
     },
     quotes: [
       %Quote{
         quote: "redacted",
         author: "redacted"
       },
       %Quote{
         quote: "redacted",
         author: "redacted"
       }
     ]
   }
 }

log

When set to true, it will set anonymized field accordingly when EctoAnon.run is called on a ressource.

Default: true

Default values

By default, a field will be anonymized with those valuee, based on its type:

type value
integer 0
float 0.0
string redacted
map %{}
decimal Decimal . new ( " 0.0 " )
date ~D[ 1970-01-01 ]
datetime ~U[ 1970-01-01 00:00:00Z ]
datetime_usec ~U[ 1970-01-01 00:00:00.000000Z ]
naive_datetime ~N[ 1970-01-01 00:00:00 ]
naive_datetime_usec ~N[ 1970-01-01 00:00:00.000000 ]
time ~T[ 00:00:00 ]
time_usec ~T[ 00:00:00.000000 ]
boolean no change
id no change
binary_id no change
binary no change

Native functions

anon_schema([
  email: :anonymized_email,
  birthdate: [:anonymized_date, options: [:only_year]]
])

Natively, ecto_anon embeds differents functions to suit your needs

function role options
:anonymized_date Anonymizes partially or completely a date/datetime :only_year
:anonymized_email Anonymizes partially or completely an email :partial
:anonymized_phone Anonymizes a phone number (currently only FR)
:random_uuid Returns a random UUID

Custom functions

anon_schema([
    address: &__MODULE__.anonymized_address/3
])

def anonymized_address(:map, %{} = address, _opts \\ []) do
  address
  |> Map.drop(["street"])
end

You can also pass custom functions with the following signature: function(type, value, options)

Migrations

By importing EctoAnon.Migration in your ecto migration file, you can add an anonymized() macro that will generate an anonymized boolean field in your table:

defmodule MyApp.Repo.Migrations.CreateUser do
  use Ecto.Migration
  import EctoAnon.Migration

  def change do
    create table(:users) do
      add :firstname, :string
      add :lastname, :string
      timestamps()
      anonymized()
    end
  end
end

Combined with log option when executing the anonymization, it will allow you to identify anonymized rows and exclude them in your queries with EctoAnon.Query.not_anonymized/1.

Filtering

As you can create an anonymized field in your migration, you can add anonymized() in your schema, just like timestamps().

By adding this field, you can use it to filter your resources and exclude anonymized data easily:

import EctoAnon.Query
import Ecto.Query

from(u in User, select: u)
|> not_anonymized()
|> Repo.all()

Copyright and License

Copyright (c) 2022 CORUSCANT (Welcome to the Jungle) - https://www.welcometothejungle.com

This library is licensed under the MIT