Store settings in output directory

For the `generate_models` script, it would be useful to make a file in the output directory that records the date + the settings that were used in the training script. (I.e. corpus settings + command line arguments.)

Start by locating the generate_models script and inspect how corpus settings and command-line arguments reach the training step and output directory. Add an output file containing the run date and the settings used; done means both corpus settings and command-line arguments are recorded alongside each generated model.

Written by the indexing model from the issue text.