Jamr

This is a little markov chain text generator I wrote a while ago. I was inspired to host it as a package hanging out with carykh's discord server.

Install by npm

npm install -g jamr

and its as simple as that, I still need to do the code to allow you to access the test data provided though, but the above commands will work as expected with any arbitrary text.

usage

To get it to run on the sample inputs:

jamr -i ./input.txt -o ./output.txt -n 5 -s 100

where n is the amount of context to give the robot, and s is how many words to generate once the model has been built. the larger the value of n, the more data should be provided, otherwise it will generate nearly exact quotes from the corpus.

the script can also take multi file inputs, for example

jamr -i ./input.txt,./other-input.txt -o ./big-output.txt -n 5 -s 100

loading and saving language models

in addition to just generating text from other text, you can save the language models used. Saving them, then using them again later, will greatly reduce the amount of time it takes to generate text

here is how you save them:

jamr -w ./model.json -i ./text-file.txt -o ./where-ever.txt -n 5 -s 5000

here is how you load models:

jamr -l ./model.json -s 5000 -n 4