ttv 0.4.0

Create train, test and validation sets from text files
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113

use std::path::PathBuf;

use structopt::StructOpt;

use crate::split::{ProportionSplit, RowSplit};

#[derive(Debug, StructOpt)]
#[structopt(
    name = "ttv",
    about = "Flexibly create test, train and validation sets"
)]
pub struct Opt {
    #[structopt(
        parse(from_occurrences),
        short = "v",
        help = "Set the level of verbosity"
    )]
    pub verbose: u64,

    #[structopt(subcommand)]
    pub cmd: Command,
}

#[derive(Debug, StructOpt)]
pub enum Command {
    #[structopt(
        name = "split",
        about = "Split dataset into two or more files for test/train/validation sets"
    )]
    Split(Split),
}

#[derive(Debug, StructOpt)]
pub struct Split {
    #[structopt(
        short = "r",
        long = "rows",
        required_unless = "prop",
        conflicts_with = "prop",
        help = "Specify splits by number of rows",
        use_delimiter = true
    )]
    pub rows: Vec<RowSplit>,

    #[structopt(
        short = "p",
        long = "prop",
        required_unless = "rows",
        conflicts_with = "rows",
        help = "Specify splits by proportion of rows",
        use_delimiter = true
    )]
    pub prop: Vec<ProportionSplit>,

    #[structopt(
        short = "n",
        long = "no-header",
        help = "Don't treat the first row as a header"
    )]
    pub no_header: bool,

    #[structopt(
        short = "c",
        long = "chunk-size",
        help = "Maximum number of rows per output chunk"
    )]
    pub chunk_size: Option<u64>,

    #[structopt(
        short = "t",
        long = "total-rows",
        help = "Number of rows in input file. Used for progress when using proportion splits"
    )]
    pub total_rows: Option<u64>,

    #[structopt(short = "s", long = "seed", help = "RNG seed, for reproducibility")]
    pub seed: Option<u64>,

    #[structopt(
        long = "csv",
        help = "Parse input as CSV. Only needed if rows contain embedded newlines - will impact performance."
    )]
    pub csv: bool,

    #[structopt(
        parse(from_os_str),
        help = "Data to split, optionally gzip compressed. If '-', read from stdin"
    )]
    pub input: PathBuf,

    #[structopt(
        short = "o",
        long = "output-prefix",
        parse(from_os_str),
        required_if("input", "-"),
        help = "Output filename prefix. Only used if reading from stdin"
    )]
    pub output_prefix: Option<PathBuf>,

    #[structopt(
        short = "d",
        long = "decompress-input",
        help = "Decompress input from gzip format"
    )]
    pub decompress_input: bool,

    #[structopt(
        short = "C",
        long = "compressed-output",
        help = "Compress output files using gzip"
    )]
    pub compress_output: bool,
}