sort.rs

static USAGE: &str = r#"
Sorts CSV data in alphabetical (with case-insensitive option), numerical,
reverse, unique or random (with optional seed) order.

The sort is done in lexicographical order.
https://en.wikipedia.org/wiki/Lexicographic_order

Note that this requires reading all of the CSV data into memory. If
you need to sort a large file that may not fit into memory, use the
extsort command instead.

For examples, see https://github.com/jqnatividad/qsv/blob/master/tests/test_sort.rs.

Usage:
    qsv sort [options] [<input>]
    qsv sort --help

sort options:
    -s, --select <arg>      Select a subset of columns to sort.
                            See 'qsv select --help' for the format details.
    -N, --numeric           Compare according to string numerical value
    -R, --reverse           Reverse order
    -i, --ignore-case       Compare strings disregarding case
    -u, --unique            When set, identical consecutive lines will be dropped
                            to keep only one line per sorted value.

    --random                Random order
    --seed <number>         Random Number Generator (RNG) seed to use if --random is set
    --rng <kind>            The RNG algorithm to use if --random is set.
                            Three RNGs are supported:
                            - standard: Use the standard RNG.
                              1.5 GB/s throughput.
                            - faster: Use faster RNG using the Xoshiro256Plus algorithm.
                              8 GB/s throughput.
                            - cryptosecure: Use cryptographically secure HC128 algorithm.
                              Recommended by eSTREAM (https://www.ecrypt.eu.org/stream/).
                              2.1 GB/s throughput though slow initialization.
                            [default: standard]
    

    -j, --jobs <arg>        The number of jobs to run in parallel.
                            When not set, the number of jobs is set to the
                            number of CPUs detected.
    --faster                When set, the sort will be faster. This is done by
                            using a faster sorting algorithm that is not stable
                            (i.e. the order of identical values is not guaranteed
                            to be preserved). It has the added side benefit that the
                            sort will also be in-place (i.e. does not allocate),
                            which is useful for sorting large files that will 
                            otherwise NOT fit in memory using the default allocating
                            stable sort.

Common options:
    -h, --help              Display this message
    -o, --output <file>     Write output to <file> instead of stdout.
    -n, --no-headers        When set, the first row will not be interpreted
                            as headers. Namely, it will be sorted with the rest
                            of the rows. Otherwise, the first row will always
                            appear as the header row in the output.
    -d, --delimiter <arg>   The field delimiter for reading CSV data.
                            Must be a single character. (default: ,)
    --memcheck              Check if there is enough memory to load the entire
                            CSV into memory using CONSERVATIVE heuristics.
"#;

use std::{cmp, str::FromStr};

// use fastrand; //DevSkim: ignore DS148264
use rand::{rngs::StdRng, seq::SliceRandom, Rng, SeedableRng};
use rand_hc::Hc128Rng;
use rand_xoshiro::Xoshiro256Plus;
use rayon::slice::ParallelSliceMut;
use serde::Deserialize;
use simdutf8::basic::from_utf8;
use strum_macros::EnumString;

use self::Number::{Float, Int};
use crate::{
    cmd::dedup::iter_cmp_ignore_case,
    config::{Config, Delimiter},
    select::SelectColumns,
    util, CliResult,
};

#[derive(Deserialize)]
struct Args {
    arg_input:        Option<String>,
    flag_select:      SelectColumns,
    flag_numeric:     bool,
    flag_reverse:     bool,
    flag_ignore_case: bool,
    flag_unique:      bool,
    flag_random:      bool,
    flag_seed:        Option<u64>,
    flag_rng:         String,
    flag_jobs:        Option<usize>,
    flag_faster:      bool,
    flag_output:      Option<String>,
    flag_no_headers:  bool,
    flag_delimiter:   Option<Delimiter>,
    flag_memcheck:    bool,
}

#[derive(Debug, EnumString, PartialEq)]
#[strum(ascii_case_insensitive)]
enum RngKind {
    Standard,
    Faster,
    Cryptosecure,
}

pub fn run(argv: &[&str]) -> CliResult<()> {
    let args: Args = util::get_args(USAGE, argv)?;
    let numeric = args.flag_numeric;
    let reverse = args.flag_reverse;
    let random = args.flag_random;
    let faster = args.flag_faster;
    let rconfig = Config::new(args.arg_input.as_ref())
        .delimiter(args.flag_delimiter)
        .no_headers(args.flag_no_headers)
        .select(args.flag_select);

    let Ok(rng_kind) = RngKind::from_str(&args.flag_rng) else {
        return fail_incorrectusage_clierror!(
            "Invalid RNG algorithm `{}`. Supported RNGs are: standard, faster, cryptosecure.",
            args.flag_rng
        );
    };

    // we're loading the entire file into memory, we need to check avail memory
    if let Some(path) = rconfig.path.clone() {
        // we only check if we're doing a stable sort and its not --random
        // coz with --faster option, the sort algorithm sorts in-place (non-allocating)
        if !faster && !random {
            util::mem_file_check(&path, false, args.flag_memcheck)?;
        }
    }

    let mut rdr = rconfig.reader()?;

    let headers = rdr.byte_headers()?.clone();
    let sel = rconfig.selection(&headers)?;

    util::njobs(args.flag_jobs);

    // Seeding RNG
    let seed = args.flag_seed;

    let ignore_case = args.flag_ignore_case;

    let mut all = rdr.byte_records().collect::<Result<Vec<_>, _>>()?;
    match (numeric, reverse, random, faster) {
        // --random sort
        (_, _, true, _) => {
            match rng_kind {
                RngKind::Standard => {
                    if let Some(val) = seed {
                        let mut rng = StdRng::seed_from_u64(val); //DevSkim: ignore DS148264
                        all.shuffle(&mut rng); //DevSkim: ignore DS148264
                    } else {
                        let mut rng = ::rand::thread_rng();
                        all.shuffle(&mut rng); //DevSkim: ignore DS148264
                    }
                },
                RngKind::Faster => {
                    let mut rng = match args.flag_seed {
                        None => Xoshiro256Plus::from_rng(rand::thread_rng()).unwrap(),
                        Some(sd) => Xoshiro256Plus::seed_from_u64(sd), // DevSkim: ignore DS148264
                    };
                    SliceRandom::shuffle(&mut *all, &mut rng); //DevSkim: ignore DS148264
                },
                RngKind::Cryptosecure => {
                    let seed_32 = match args.flag_seed {
                        None => rand::thread_rng().gen::<[u8; 32]>(),
                        Some(seed) => {
                            let seed_u8 = seed.to_le_bytes();
                            let mut seed_32 = [0u8; 32];
                            seed_32[..8].copy_from_slice(&seed_u8);
                            seed_32
                        },
                    };
                    let mut rng: Hc128Rng = match args.flag_seed {
                        None => Hc128Rng::from_rng(rand::thread_rng()).unwrap(),
                        Some(_) => Hc128Rng::from_seed(seed_32),
                    };
                    SliceRandom::shuffle(&mut *all, &mut rng);
                },
            }
        },

        // default stable parallel sort
        (false, false, false, false) => all.par_sort_by(|r1, r2| {
            let a = sel.select(r1);
            let b = sel.select(r2);
            if ignore_case {
                iter_cmp_ignore_case(a, b)
            } else {
                iter_cmp(a, b)
            }
        }),
        // default --faster unstable, non-allocating parallel sort
        (false, false, false, true) => all.par_sort_unstable_by(|r1, r2| {
            let a = sel.select(r1);
            let b = sel.select(r2);
            if ignore_case {
                iter_cmp_ignore_case(a, b)
            } else {
                iter_cmp(a, b)
            }
        }),

        // --numeric stable parallel numeric sort
        (true, false, false, false) => all.par_sort_by(|r1, r2| {
            let a = sel.select(r1);
            let b = sel.select(r2);
            iter_cmp_num(a, b)
        }),
        // --numeric --faster unstable, non-allocating, parallel numeric sort
        (true, false, false, true) => all.par_sort_unstable_by(|r1, r2| {
            let a = sel.select(r1);
            let b = sel.select(r2);
            iter_cmp_num(a, b)
        }),

        // --reverse stable parallel sort
        (false, true, false, false) => all.par_sort_by(|r1, r2| {
            let a = sel.select(r1);
            let b = sel.select(r2);
            if ignore_case {
                iter_cmp_ignore_case(b, a)
            } else {
                iter_cmp(b, a)
            }
        }),
        // --reverse --faster unstable parallel sort
        (false, true, false, true) => all.par_sort_unstable_by(|r1, r2| {
            let a = sel.select(r1);
            let b = sel.select(r2);
            if ignore_case {
                iter_cmp_ignore_case(b, a)
            } else {
                iter_cmp(b, a)
            }
        }),

        // --numeric --reverse stable sort
        (true, true, false, false) => all.par_sort_by(|r1, r2| {
            let a = sel.select(r1);
            let b = sel.select(r2);
            iter_cmp_num(b, a)
        }),
        // --numeric --reverse --faster unstable sort
        (true, true, false, true) => all.par_sort_unstable_by(|r1, r2| {
            let a = sel.select(r1);
            let b = sel.select(r2);
            iter_cmp_num(b, a)
        }),
    }

    let mut wtr = Config::new(args.flag_output.as_ref()).writer()?;
    let mut prev: Option<csv::ByteRecord> = None;
    rconfig.write_headers(&mut rdr, &mut wtr)?;
    for r in all {
        if args.flag_unique {
            match prev {
                Some(other_r) => match iter_cmp(sel.select(&r), sel.select(&other_r)) {
                    cmp::Ordering::Equal => (),
                    _ => {
                        wtr.write_byte_record(&r)?;
                    },
                },
                None => {
                    wtr.write_byte_record(&r)?;
                },
            }

            prev = Some(r);
        } else {
            wtr.write_byte_record(&r)?;
        }
    }
    Ok(wtr.flush()?)
}

/// Order `a` and `b` lexicographically using `Ord`
#[inline]
pub fn iter_cmp<A, L, R>(mut a: L, mut b: R) -> cmp::Ordering
where
    A: Ord,
    L: Iterator<Item = A>,
    R: Iterator<Item = A>,
{
    loop {
        match (a.next(), b.next()) {
            (None, None) => return cmp::Ordering::Equal,
            (None, _) => return cmp::Ordering::Less,
            (_, None) => return cmp::Ordering::Greater,
            (Some(x), Some(y)) => match x.cmp(&y) {
                cmp::Ordering::Equal => (),
                non_eq => return non_eq,
            },
        }
    }
}

/// Try parsing `a` and `b` as numbers when ordering
#[inline]
pub fn iter_cmp_num<'a, L, R>(mut a: L, mut b: R) -> cmp::Ordering
where
    L: Iterator<Item = &'a [u8]>,
    R: Iterator<Item = &'a [u8]>,
{
    loop {
        match (next_num(&mut a), next_num(&mut b)) {
            (None, None) => return cmp::Ordering::Equal,
            (None, _) => return cmp::Ordering::Less,
            (_, None) => return cmp::Ordering::Greater,
            (Some(x), Some(y)) => match compare_num(x, y) {
                cmp::Ordering::Equal => (),
                non_eq => return non_eq,
            },
        }
    }
}

#[derive(Clone, Copy, PartialEq)]
enum Number {
    Int(i64),
    Float(f64),
}

#[inline]
fn compare_num(n1: Number, n2: Number) -> cmp::Ordering {
    match (n1, n2) {
        (Int(i1), Int(i2)) => i1.cmp(&i2),
        #[allow(clippy::cast_precision_loss)]
        (Int(i1), Float(f2)) => compare_float(i1 as f64, f2),
        #[allow(clippy::cast_precision_loss)]
        (Float(f1), Int(i2)) => compare_float(f1, i2 as f64),
        (Float(f1), Float(f2)) => compare_float(f1, f2),
    }
}

#[inline]
fn compare_float(f1: f64, f2: f64) -> cmp::Ordering {
    f1.partial_cmp(&f2).unwrap_or(cmp::Ordering::Equal)
}

#[inline]
fn next_num<'a, X>(xs: &mut X) -> Option<Number>
where
    X: Iterator<Item = &'a [u8]>,
{
    match xs.next() {
        Some(bytes) => {
            if let Ok(i) = atoi_simd::parse::<i64>(bytes) {
                Some(Number::Int(i))
            } else {
                // If parsing as i64 failed, try parsing as f64
                if let Ok(f) = from_utf8(bytes).unwrap().parse::<f64>() {
                    Some(Number::Float(f))
                } else {
                    None
                }
            }
        },
        None => None,
    }
}