superset_map

superset_map is a library for Sets that have an order defined on them. Its main data structure is SupersetMap which is a specialized version of BTreeMap where only supersets are stored. This can be useful when the keys don't fit well or at all with the concept of RangeBounds.

Nightly Rust is required. Once BTreeMap cursors are stabilized, stable Rust will work.

Example

use core::{borrow::Borrow, cmp::Ordering};
use superset_map::{zfc::{num_bigint::BigUint, BoundedCardinality, Cardinality, Set}, SetOrd, SupersetSet};
#[derive(Clone, Copy, Eq, PartialEq)]
struct ShortAscii<'a> {
    val: &'a [u8],
}
impl<'a> ShortAscii<'a> {
    fn new(val: &'a [u8]) -> Option<ShortAscii<'a>> {
        (val.len() <= 255 && val.is_ascii()).then_some(Self { val })
    }
    fn len(self) -> u8 {
        self.val.len().try_into().expect("The ShortAscii instance was not constructed properly and contains more than 255 bytes.")
    }
}
#[derive(Clone, Copy, Eq, PartialEq)]
enum WildcardAscii<'a> {
    Plain(ShortAscii<'a>),
    // Represents a ShortAscii<'a> with an implicit wildcard at the end
    // meaning it's all ShortAscii<'a>s that begin with the contained ShortAscii<'a>.val.
    Wildcard(ShortAscii<'a>),
}
impl<'a> WildcardAscii<'a> {
    const fn val(self) -> ShortAscii<'a> {
        match self {
            WildcardAscii::Plain(s) | WildcardAscii::Wildcard(s) => s,
        }
    }
    const fn is_plain(self) -> bool {
        match self {
            WildcardAscii::Plain(_) => true,
            WildcardAscii::Wildcard(_) => false,
        }
    }
    const fn is_wildcard(self) -> bool {
        !self.is_plain()
    }
}
impl<'a> PartialOrd<Self> for WildcardAscii<'a> {
    fn partial_cmp(&self, other: &Self) -> Option<Ordering> {
        Some(self.cmp(other))
    }
}
impl<'a> Ord for WildcardAscii<'a> {
    fn cmp(&self, other: &Self) -> Ordering {
        let len = u8::min(self.val().len(), other.val().len()) as usize;
        match self.val().val[..len].cmp(&other.val().val[..len]) {
            Ordering::Less => Ordering::Less,
            Ordering::Equal => {
                if self.is_wildcard() {
                    if other.is_wildcard() {
                        self.val().len().cmp(&other.val().len()).reverse()
                    } else {
                        Ordering::Greater
                    }
                } else if other.is_wildcard() {
                    Ordering::Less
                } else {
                    self.val().len().cmp(&other.val().len())
                }
            }
            Ordering::Greater => Ordering::Greater,
        }
    }
}
impl<'a> Set for WildcardAscii<'a> {
    type Elem = ShortAscii<'a>;
    fn bounded_cardinality(&self) -> BoundedCardinality {
        BoundedCardinality::new_exact(self.cardinality().unwrap())
    }
    fn cardinality(&self) -> Option<Cardinality> {
        Some(Cardinality::Finite(match *self {
            WildcardAscii::Plain(_) => BigUint::new(vec![1]),
            // Geometric series.
            WildcardAscii::Wildcard(v) => {
                (BigUint::new(vec![128]).pow((u8::MAX - v.len()) as u32 + 1)
                    - BigUint::new(vec![1]))
                    / BigUint::new(vec![127])
            }
        }))
    }
    fn contains<Q>(&self, elem: &Q) -> bool
    where
        Q: Borrow<Self::Elem> + Eq + ?Sized,
    {
        match *self {
            WildcardAscii::Plain(v) => v == *elem.borrow(),
            WildcardAscii::Wildcard(v) => {
                v.len() <= elem.borrow().len() && *v.val == elem.borrow().val[..v.len() as usize]
            }
        }
    }
    fn is_proper_subset(&self, val: &Self) -> bool {
        val.is_wildcard()
            && match val.val().len().cmp(&self.val().len()) {
                Ordering::Less => val.val().val == &self.val().val[..val.val().len() as usize],
                Ordering::Equal => self.is_plain() && self.val() == val.val(),
                Ordering::Greater => false,
            }
    }
    fn is_subset(&self, val: &Self) -> bool {
        self == val || self.is_proper_subset(val)
    }
}
impl<'a> SetOrd for WildcardAscii<'a> {}
fn main() {
    let mut set = SupersetSet::new();
    set.insert(WildcardAscii::Plain(ShortAscii::new(b"foo").unwrap()));
    set.insert(WildcardAscii::Plain(ShortAscii::new(b"bar").unwrap()));
    set.insert(WildcardAscii::Wildcard(ShortAscii::new(b"b").unwrap()));
    set.insert(WildcardAscii::Wildcard(ShortAscii::new(b"bar").unwrap()));
    let mut iter = set.into_iter();
    assert!(iter.next().map_or(false, |s| s
        == WildcardAscii::Wildcard(ShortAscii::new(b"b").unwrap())));
    assert!(iter.next().map_or(false, |s| s
        == WildcardAscii::Plain(ShortAscii::new(b"foo").unwrap())));
    assert!(iter.next().is_none());
}

Minimum Supported Rust Version (MSRV)

This will frequently be updated to be the same as stable. Specifically, any time stable is updated and that update has "useful" features or compilation no longer succeeds (e.g., due to new compiler lints), then MSRV will be updated.

MSRV changes will correspond to a SemVer patch version bump pre-1.0.0; otherwise a minor version bump.

SemVer Policy

• All on-by-default features of this library are covered by SemVer
• MSRV is considered exempt from SemVer as noted above

License

Licensed under either of

• Apache License, Version 2.0 (LICENSE-APACHE)
• MIT license (LICENSE-MIT)

at your option.

Contribution

Unless you explicitly state otherwise, any contribution intentionally submitted for inclusion in the work by you, as defined in the Apache-2.0 license, shall be dual licensed as above, without any additional terms or conditions.

Before any PR is sent, cargo clippy and cargo t should be run for both. Additionally RUSTDOCFLAGS="--cfg docsrs" cargo +nightly doc --all-features should be run to ensure documentation can be built.

Status

The crate is only tested on the x86_64-unknown-linux-gnu, x86_64-unknown-openbsd, and aarch64-apple-darwin targets; but it should work on most platforms.