Breaking Vigenère Ciphers /w Genetic Algorithm

Implementation of a genetic algorithm to break Vigenère ciphers

Generated Data

See data/ directory

Build

1. Activate virtual environment

$ python -m venv .venv
$ source ./.venv/bin/activate

2. Build project

$ pip install build
$ python -m build

3. Install project

$ pip install dist/geneticalgorithm-0.1.0-py3-none-any.whl
$ pip install tqdm tabulate # (optional) for progress bar and formatting output

Usage

Run Genetic Algorithm

$ geneticalgorithm 8 -f attachments/sample.txt -s 3

Run experiment

$ experiment config/sample.json -o csv
$ experiment config/sample.json -v     # display all outputs (including decrypted text)
$ experiment config/sample.json -o tbl # requires tabulate package to be installed

Reproducing data/data1.csv and data/data2.csv

$ experiment config/data1.json -o csv > data/data1.csv
$ experiment config/data2.json -o csv > data/data2.csv

Changing Parameters (geneticalgorithm)

To get the list of parameters that can be changed, use:

$ geneticalgorithm --help

For example, to change the crossover algorithm to Order Crossover (OX):

$ geneticalgorithm 8 -f attachments/sample.txt -s 3 --crossover-alg=ox

Or the crossover and mutation rates:

$ geneticalgorithm 8 -f attachments/sample.txt -s 3 -c 0.8 -m 0.05

Changing Parameters (experiment)

To change the experiment configurations, edit one of data1.json, data2.json, or sample.json. A sample configuration file looks like the following:

{
  "pop_size": 50,
  "max_gen": 20,
  "elites": 2,
  "seeds": [2, 3],
  "runs": [
    {
      "file": "attachments/sample.txt",
      "key_length": 8,
      "crossover_algorithms": ["ux", "ox"],
      "mutation_algorithms": ["rx", "rc"],
      "selection_algorithms": ["tour3"],
      "rates": [
        {"crossover": 1.0, "mutation": 0.0},
        {"crossover": 0.9, "mutation": 0.1}
      ]
    }
  ]
}

You can run the experiment with this configuration via:

$ experiment /path/to/config.json

For the sample configuration above, this will run the GA for every combination of random seeds, crossover algorithms, mutation algorithms, selection algorithms, and crossover rates.