polars-arrow 0.26.1

Arrow interfaces for Polars DataFrame library
Documentation
use std::ops::BitAnd;

use arrow::array::PrimitiveArray;
use arrow::bitmap::Bitmap;
use arrow::types::NativeType;

use crate::trusted_len::{FromIteratorReversed, PushUnchecked, TrustedLen};

#[derive(Clone)]
pub struct TrustMyLength<I: Iterator<Item = J>, J> {
    iter: I,
    len: usize,
}

impl<I, J> TrustMyLength<I, J>
where
    I: Iterator<Item = J>,
{
    #[inline]
    pub fn new(iter: I, len: usize) -> Self {
        Self { iter, len }
    }
}

impl<I, J> Iterator for TrustMyLength<I, J>
where
    I: Iterator<Item = J>,
{
    type Item = J;

    #[inline]
    fn next(&mut self) -> Option<Self::Item> {
        self.iter.next()
    }

    fn size_hint(&self) -> (usize, Option<usize>) {
        (self.len, Some(self.len))
    }
}

impl<I, J> ExactSizeIterator for TrustMyLength<I, J> where I: Iterator<Item = J> {}

impl<I, J> DoubleEndedIterator for TrustMyLength<I, J>
where
    I: Iterator<Item = J> + DoubleEndedIterator,
{
    #[inline]
    fn next_back(&mut self) -> Option<Self::Item> {
        self.iter.next_back()
    }
}

pub fn combine_validities(opt_l: Option<&Bitmap>, opt_r: Option<&Bitmap>) -> Option<Bitmap> {
    match (opt_l, opt_r) {
        (Some(l), Some(r)) => Some(l.bitand(r)),
        (None, Some(r)) => Some(r.clone()),
        (Some(l), None) => Some(l.clone()),
        (None, None) => None,
    }
}
unsafe impl<I, J> arrow::trusted_len::TrustedLen for TrustMyLength<I, J> where I: Iterator<Item = J> {}

pub trait CustomIterTools: Iterator {
    fn fold_first_<F>(mut self, f: F) -> Option<Self::Item>
    where
        Self: Sized,
        F: FnMut(Self::Item, Self::Item) -> Self::Item,
    {
        let first = self.next()?;
        Some(self.fold(first, f))
    }

    /// Turn any iterator in a trusted length iterator
    ///
    /// # Safety
    /// The given length must be correct.
    unsafe fn trust_my_length(self, length: usize) -> TrustMyLength<Self, Self::Item>
    where
        Self: Sized,
    {
        TrustMyLength::new(self, length)
    }

    fn collect_trusted<T: FromTrustedLenIterator<Self::Item>>(self) -> T
    where
        Self: Sized + TrustedLen,
    {
        FromTrustedLenIterator::from_iter_trusted_length(self)
    }

    fn collect_reversed<T: FromIteratorReversed<Self::Item>>(self) -> T
    where
        Self: Sized + TrustedLen,
    {
        FromIteratorReversed::from_trusted_len_iter_rev(self)
    }

    fn all_equal(&mut self) -> bool
    where
        Self: Sized,
        Self::Item: PartialEq,
    {
        match self.next() {
            None => true,
            Some(a) => self.all(|x| a == x),
        }
    }

    fn fold_options<A, B, F>(&mut self, mut start: B, mut f: F) -> Option<B>
    where
        Self: Iterator<Item = Option<A>>,
        F: FnMut(B, A) -> B,
    {
        for elt in self {
            match elt {
                Some(v) => start = f(start, v),
                None => return None,
            }
        }
        Some(start)
    }
}

pub trait CustomIterToolsSized: Iterator + Sized {}

impl<T: ?Sized> CustomIterTools for T where T: Iterator {}

pub trait FromTrustedLenIterator<A>: Sized {
    fn from_iter_trusted_length<T: IntoIterator<Item = A>>(iter: T) -> Self
    where
        T::IntoIter: TrustedLen;
}

impl<T> FromTrustedLenIterator<T> for Vec<T> {
    fn from_iter_trusted_length<I: IntoIterator<Item = T>>(iter: I) -> Self
    where
        I::IntoIter: TrustedLen,
    {
        let iter = iter.into_iter();
        let len = iter.size_hint().0;
        let mut v = Vec::with_capacity(len);
        v.extend_trusted_len(iter);
        v
    }
}

impl<T: NativeType> FromTrustedLenIterator<Option<T>> for PrimitiveArray<T> {
    fn from_iter_trusted_length<I: IntoIterator<Item = Option<T>>>(iter: I) -> Self
    where
        I::IntoIter: TrustedLen,
    {
        let iter = iter.into_iter();
        unsafe { PrimitiveArray::from_trusted_len_iter_unchecked(iter) }
    }
}

impl<T: NativeType> FromTrustedLenIterator<T> for PrimitiveArray<T> {
    fn from_iter_trusted_length<I: IntoIterator<Item = T>>(iter: I) -> Self
    where
        I::IntoIter: TrustedLen,
    {
        let iter = iter.into_iter();
        unsafe { PrimitiveArray::from_trusted_len_values_iter_unchecked(iter) }
    }
}

macro_rules! with_match_primitive_type {(
    $key_type:expr, | $_:tt $T:ident | $($body:tt)*
) => ({
    macro_rules! __with_ty__ {( $_ $T:ident ) => ( $($body)* )}
    use arrow::datatypes::PrimitiveType::*;
    use arrow::types::{days_ms, months_days_ns};
    match $key_type {
        Int8 => __with_ty__! { i8 },
        Int16 => __with_ty__! { i16 },
        Int32 => __with_ty__! { i32 },
        Int64 => __with_ty__! { i64 },
        Int128 => __with_ty__! { i128 },
        DaysMs => __with_ty__! { days_ms },
        MonthDayNano => __with_ty__! { months_days_ns },
        UInt8 => __with_ty__! { u8 },
        UInt16 => __with_ty__! { u16 },
        UInt32 => __with_ty__! { u32 },
        UInt64 => __with_ty__! { u64 },
        Float32 => __with_ty__! { f32 },
        Float64 => __with_ty__! { f64 },
        Float16 | Int256 => unimplemented!(),
    }
})}
pub(crate) use with_match_primitive_type;