Hillary Clinton Email Generator

The project uses the markovify library (Markov Chains) and Hillary Clinton's public email corpus to generate random emails by Clinton and her colleagues.

This email generator does the reverse of what Hillary Clinton's IT guy did to her email server using BleachBit: instead of deleting Clinton's emails, it generates them. Remarkably, both are anti-forensic techniques because generating junk makes it more difficult to find your important data. Using the generator is like adding a haystack on top of the proverbial needle: it's harder to find the needle in the haystack than to find the needle in a clean space.

This repository has only the command-line version, which requires Python 2.7 or 3.x. For a GUI version with an installer, use BleachBit

To generate emails

Call python generate_emails.py <number_of_emails>. To use additional arguments, refer to:

usage: generate_emails.py [-h] [--number_of_sentences NUMBER_OF_SENTENCES]
                          [--subject_length SUBJECT_LENGTH]
                          [--email_output_path EMAIL_OUTPUT_PATH]
                          [--content_model_path CONTENT_MODEL_PATH]
                          [--subject_model_path SUBJECT_MODEL_PATH]
                          [number_of_emails]

positional arguments:
  number_of_emails      Number of emails to generate. Default is 5
optional arguments:
  -h, --help            show this help message and exit
  --number_of_sentences NUMBER_OF_SENTENCES
                        Number of sentences per email to generate. Default is
                        5
  --subject_length SUBJECT_LENGTH
                        Length of the subject line in characters. Default is
                        64
  --email_output_path EMAIL_OUTPUT_PATH
                        Path to output the generated emails to. Default is
                        ./emails.txt
  --content_model_path CONTENT_MODEL_PATH
                        Path to load the content model from. Default is
                        ./content_model.json
  --subject_model_path SUBJECT_MODEL_PATH
                        Path to load the subject model from. Default is
                        ./subject_model.json

To regenerate the Markov model

Call python generate_model.py. To use additional arguments, refer to:

usage: generate_model.py [-h]
                         [--markov_model_state_size MARKOV_MODEL_STATE_SIZE]
                         [--use_only_hillarys_emails]
                         [--emails_path EMAILS_PATH]
                         [--content_model_path CONTENT_MODEL_PATH]
                         [--subject_model_path SUBJECT_MODEL_PATH]

optional arguments:
  -h, --help            show this help message and exit
  --markov_model_state_size MARKOV_MODEL_STATE_SIZE
                        State size of the Markov model to use. Default is 2
  --use_only_hillarys_emails
                        Include this flag to generate the model only from
                        Hillary's emails. Default is False
  --emails_path EMAILS_PATH
                        Path to load the emails from. Default is
                        ./Emails.csv
  --content_model_path CONTENT_MODEL_PATH
                        Path to load the content model from. Default is
                        ./content_model.json
  --subject_model_path SUBJECT_MODEL_PATH
                        Path to load the subject model from. Default is
                        ./subject_model.json

Licenses

The United States Department of State released Hillary Clinton's emails into the public domain.

Markovify uses the MIT license.

The Hillary Clinton email generator uses the GNU General Public License version 3 or later. Dionyz Lazar wrote the original version under contract. Copyright (C) 2019 by Andrew Ziem.