Anonymise CKAN DB

A script to replace user information in the CKAN database with fake data. The script will:

• Connect to the CKAN database in a local PostgreSQL.
• Select all entries from the user table.
• Replace the each entry's fullname with a name generated by the faker tool.
• Repleace each entry's name and email with a values generated from the new fullname.
• Dump the modified and anomymised db to a file using pg_dump.
• Roll back the db to its initial state.

Usage

% python anonymise_db.py --help    
usage: anonymise_db.py [-h] [--db DB] [--user USER] [--host HOST] [--port PORT] --pw PW

Connect to a CKAN DB in Postgres and replace `fullname`, `name` and `email` in the `user` table with random names.
Also delete `about` and `image_url` (set to '').

options:
  -h, --help           show this help message and exit
  --db DB              Name of the CKAN database in Postgres. Default: ckan
  --user USER          Username to access the database. Default: ckan
  --host HOST          Postgres database host. Default: localhost.
  --port PORT          Postgres database port. Default: 5432.

required named arguments:
  --pw PW              Password to access the database

Requirements

• The script has been tested with Python 3.12.
• It uses faker, python-slugify and psycopg2 (see requirements.txt)

Installation

• Clone this repository

git clone https://github.com/berlinonline/anonymise_ckan_db

• Optionally create and activate a virtual environment

cd anonymise_ckan_db
python -m venv venv
. venv/bin/activate

• Install dependencies:

pip install -r requirements.txt

• Before running, you'll need a file exclude.json with a JSON list of usernames to exclude from the anonymisation (function accounts etc.).

License

All code in this repository is published under the MIT License.