mkv-this/README.md

95 lines
5.1 KiB
Markdown
Raw Normal View History

2020-04-19 19:57:13 +02:00
2020-04-21 02:42:00 +02:00
## disclaimer
2020-04-20 18:02:13 +02:00
2020-04-20 18:36:27 +02:00
i wrote this cli rapper for the `markovify` python module because i wanted its features to be available as a cli tool.
2020-04-20 18:02:13 +02:00
2020-04-20 18:36:27 +02:00
i only published it in case someone who actually knows what they're doing felt like picking it up and improving on it. (and to share it with friends.)
2020-04-20 18:02:13 +02:00
2020-04-20 19:55:34 +02:00
maybe this functionality already exists somewhere, but i couldn't find it.
2020-04-20 18:36:27 +02:00
so if unlike me you are actually a programmer and are interested in expanding/correcting/fixing my non-programmer's code, then by all means!
2020-04-20 18:02:13 +02:00
2020-04-19 19:57:13 +02:00
## mkv-this
2020-04-20 18:24:36 +02:00
`mkv-this` is a little script that outputs a bunch of bot-like sentences based on a bank of text that you feed it. the results are saved to a text file. if you run it again with the same output file, the new results are appended after the old ones.
2020-04-19 19:57:13 +02:00
2020-04-20 18:24:36 +02:00
a second command, `mkv-this-dir` (see below) allows you to input a directory and it will read all text files within it as the input.
2020-04-19 19:57:13 +02:00
2020-04-20 18:24:36 +02:00
`mkv-this` simply makes some of the features of the excellent `markovify` python module available as a command line tool. it was written by a total novice, so you probably shouldnt download it. i only learned about `argparser` yesterday, and pypi.org today, no matter what day it is. tomorrow i might learn about `os` and `sys`. and then maybe even `cookiecutter`!
2020-04-19 19:57:13 +02:00
2020-04-21 02:42:00 +02:00
### installing
2020-04-19 19:57:13 +02:00
install it with `pip`, the python package manager:
`python3 -m pip install mkv-this`
or
`pip install mkv-this`
to do this you need `python3` and `pip`. if you don't have them, install them through your system's package manager. on debian (+ derivatives), for example, you'd run:
`sudo apt install python3 python3-pip`
`markovify` is also a dependency, but it should install along with `mkv-this`.
2020-04-21 02:42:00 +02:00
if you get sth like `ModuleNotFound error: No module named 'modulename'`, just run `pip install modulename` to get it.
2020-04-20 18:42:45 +02:00
### repository
if you are reading this on pypi.org, the repo is here:
https://git.disroot.org/mousebot/mkv-this
### macos
2020-04-19 19:57:13 +02:00
it seems to run on macos too.
you may already have python installed. if not, you first need to install [homebrew](https://brew.sh/#install), edit your PATH so that it works, then install `python3` with `brew install python3`. if you are already running an old version of `homebrew` you might need to run `brew install python3 && brew postinstall python3` to get `python3` and `pip` running right.
2020-04-19 19:57:13 +02:00
you can check if `pip` is installed with `pip --version`, or `pip3 --version`.
2020-04-19 19:57:13 +02:00
i know nothing about macs so if you ask me for help i'll just send you random copypasta from the interwebs.
2020-04-21 02:42:00 +02:00
### options
2020-04-19 19:57:13 +02:00
the script implements a few of the basic `markovify` options, so you can:
* specify output file (default = "./mkv-output.txt")
* specify a maximum sentence length, in characters.
* specify how many sentences to output (default = 5)
* specify state size, i.e. the number of preceeding words to be used in calculating the probability of the next word (default = 2).
* specify the amount of (verbatim) overlap allowed between your input text and your output text.
* specify that your text's sentences end with newlines rather than full-stops.
2020-04-19 19:57:13 +02:00
* specify an additional file to use for text input. you can add only one. if you want to feed a stack of files into your bank, use `mkv-this-dir`.
* if a second file is added, you can also specify the relative weight to give to the two files.
run `mkv-this -h` to see how to use these options.
### mkv-this-dir: markovify a directory of text files
2020-04-19 19:57:13 +02:00
`mkv-this` can only take two files as input material each time. if you want to input a stack of files, use `mkv-this-dir`. it allows you to specify a directory and all text files in it will be used as input material.
if for some reason you want to get a similar funtionality with `mkv-this`, you can easily concatenate some files yourself in bash, then process them:
* copy all your text files into a directory
* cd into the directory
* run `cat * > outputfile.txt`
* run mkv-this on your newly created file: `mkv-this outputfile.txt`
* this approach has the benefit of creating a file with encoding that mkv-this can certainly handle.
2020-04-19 19:57:13 +02:00
2020-04-21 02:42:00 +02:00
### file types
2020-04-19 19:57:13 +02:00
you need to input plain text files. currently accepted file extensions are `.txt`, `.org` and `.md`. it is trivial to add others, so if you want one included just ask.
2020-04-21 02:42:00 +02:00
### for best results
2020-04-19 19:57:13 +02:00
feed `mkv-this` large-ish amounts of well punctuated text. it works best if you bulk replace/remove as much mess as possible (URLs, metadata, stars, bullets, etc.), unless you want mashed versions of those things in your output.
if your input text doesnt use full-stops to mark the ends of sentences, try putting each 'sentence' on a newline, and then write to the maintainer of this package to complain about how that option isn't yet implemented. then the parser wont read your entire file as one big sentence and output nothing.
youll probably want to edit the output. it is very much supposed to be a kind of raw material rather than print-ready boilerplate bosh, although many bots are happily publishing such output directly. you might find that it prompts you to edit it like a bot yourself.
for a few further tips, see https://github.com/jsvine/markovify#basic-usage.
happy zaning.