about
a markov chain implementation in zig. this tool is useful for generating pseudo-random text based on some input files. outputs to stdout.
usage
$ zig build run -Dblock-len=8 -- --start-block "pub fn" --maxlen 1000 infile1 infile2 ...args
- -Dblock-len : how many bytes to consider when predicting the next character. defaults to 8.
- --start-block : initial seed to start generation. this string should be present somewhere in input. generated text will start with this string. if not provided, a random start-block is chosen. if provided string is longer than -Dblock-len, only it's last block-len characters will be used.
- --maxlen : how many characters to generate. does not include length of --start-block. generation will stop after this many characters.
usage examples
$ zig build run -- --start-block "test \"" --maxlen 1000 $(ls ../zig/test/behavior/*.zig)$ zig build run -- --start-block "pub fn" --maxlen 1000 $(find ../zig/test/behavior/ -name "*.zig")TODOs
- gen - fix repetitive spaces
- gen - recover when current block not found
- add a block-len build param
- add usage text
- gen - write possibly skipped portion of start-block