perf(data): compact immutable snapshots with current execution rules
This commit is contained in:
@@ -0,0 +1,339 @@
|
||||
use std::borrow::Cow;
|
||||
use std::collections::BTreeMap;
|
||||
use std::fmt;
|
||||
use std::ops::Index;
|
||||
|
||||
use serde::de::{MapAccess, Visitor};
|
||||
use serde::ser::SerializeMap;
|
||||
use serde::{Deserialize, Deserializer, Serialize, Serializer};
|
||||
|
||||
/// Sorted numeric fields stored contiguously, without a tree node per snapshot.
|
||||
#[derive(Clone, Default, PartialEq)]
|
||||
pub struct NumericFactorMap {
|
||||
entries: Vec<(Cow<'static, str>, f64)>,
|
||||
}
|
||||
|
||||
impl NumericFactorMap {
|
||||
pub const fn new() -> Self {
|
||||
Self {
|
||||
entries: Vec::new(),
|
||||
}
|
||||
}
|
||||
|
||||
pub fn len(&self) -> usize {
|
||||
self.entries.len()
|
||||
}
|
||||
pub fn is_empty(&self) -> bool {
|
||||
self.entries.is_empty()
|
||||
}
|
||||
pub fn clear(&mut self) {
|
||||
self.entries.clear();
|
||||
}
|
||||
|
||||
pub fn get(&self, key: &str) -> Option<&f64> {
|
||||
self.entries
|
||||
.binary_search_by(|(name, _)| name.as_ref().cmp(key))
|
||||
.ok()
|
||||
.map(|index| &self.entries[index].1)
|
||||
}
|
||||
|
||||
pub fn get_mut(&mut self, key: &str) -> Option<&mut f64> {
|
||||
self.entries
|
||||
.binary_search_by(|(name, _)| name.as_ref().cmp(key))
|
||||
.ok()
|
||||
.map(|index| &mut self.entries[index].1)
|
||||
}
|
||||
|
||||
pub fn contains_key(&self, key: &str) -> bool {
|
||||
self.get(key).is_some()
|
||||
}
|
||||
|
||||
pub fn insert(&mut self, key: Cow<'static, str>, value: f64) -> Option<f64> {
|
||||
if self
|
||||
.entries
|
||||
.last()
|
||||
.is_none_or(|(last, _)| last.as_ref() < key.as_ref())
|
||||
{
|
||||
self.entries.push((key, value));
|
||||
return None;
|
||||
}
|
||||
match self
|
||||
.entries
|
||||
.binary_search_by(|(name, _)| name.as_ref().cmp(key.as_ref()))
|
||||
{
|
||||
Ok(index) => Some(std::mem::replace(&mut self.entries[index].1, value)),
|
||||
Err(index) => {
|
||||
self.entries.insert(index, (key, value));
|
||||
None
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
pub fn remove(&mut self, key: &str) -> Option<f64> {
|
||||
self.entries
|
||||
.binary_search_by(|(name, _)| name.as_ref().cmp(key))
|
||||
.ok()
|
||||
.map(|index| self.entries.remove(index).1)
|
||||
}
|
||||
|
||||
pub fn retain(&mut self, mut keep: impl FnMut(&Cow<'static, str>, &mut f64) -> bool) {
|
||||
self.entries.retain_mut(|(key, value)| keep(key, value));
|
||||
}
|
||||
|
||||
pub fn iter(&self) -> Iter<'_> {
|
||||
Iter(self.entries.iter())
|
||||
}
|
||||
pub fn keys(&self) -> impl DoubleEndedIterator<Item = &Cow<'static, str>> + ExactSizeIterator {
|
||||
self.entries.iter().map(|(key, _)| key)
|
||||
}
|
||||
pub fn values(&self) -> impl DoubleEndedIterator<Item = &f64> + ExactSizeIterator {
|
||||
self.entries.iter().map(|(_, value)| value)
|
||||
}
|
||||
}
|
||||
|
||||
impl fmt::Debug for NumericFactorMap {
|
||||
fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
|
||||
f.debug_map().entries(self).finish()
|
||||
}
|
||||
}
|
||||
|
||||
impl Index<&str> for NumericFactorMap {
|
||||
type Output = f64;
|
||||
fn index(&self, key: &str) -> &Self::Output {
|
||||
self.get(key).expect("numeric factor key not found")
|
||||
}
|
||||
}
|
||||
|
||||
pub struct Iter<'a>(std::slice::Iter<'a, (Cow<'static, str>, f64)>);
|
||||
impl<'a> Iterator for Iter<'a> {
|
||||
type Item = (&'a Cow<'static, str>, &'a f64);
|
||||
fn next(&mut self) -> Option<Self::Item> {
|
||||
self.0.next().map(|(k, v)| (k, v))
|
||||
}
|
||||
fn size_hint(&self) -> (usize, Option<usize>) {
|
||||
self.0.size_hint()
|
||||
}
|
||||
}
|
||||
impl DoubleEndedIterator for Iter<'_> {
|
||||
fn next_back(&mut self) -> Option<Self::Item> {
|
||||
self.0.next_back().map(|(k, v)| (k, v))
|
||||
}
|
||||
}
|
||||
impl ExactSizeIterator for Iter<'_> {}
|
||||
impl<'a> IntoIterator for &'a NumericFactorMap {
|
||||
type Item = (&'a Cow<'static, str>, &'a f64);
|
||||
type IntoIter = Iter<'a>;
|
||||
fn into_iter(self) -> Self::IntoIter {
|
||||
self.iter()
|
||||
}
|
||||
}
|
||||
impl IntoIterator for NumericFactorMap {
|
||||
type Item = (Cow<'static, str>, f64);
|
||||
type IntoIter = std::vec::IntoIter<Self::Item>;
|
||||
fn into_iter(self) -> Self::IntoIter {
|
||||
self.entries.into_iter()
|
||||
}
|
||||
}
|
||||
|
||||
impl FromIterator<(Cow<'static, str>, f64)> for NumericFactorMap {
|
||||
fn from_iter<T: IntoIterator<Item = (Cow<'static, str>, f64)>>(iter: T) -> Self {
|
||||
let mut entries: Vec<_> = iter.into_iter().collect();
|
||||
// Stable sorting preserves last-value-wins for repeated input keys.
|
||||
if !entries.windows(2).all(|pair| pair[0].0 <= pair[1].0) {
|
||||
entries.sort_by(|left, right| left.0.cmp(&right.0));
|
||||
}
|
||||
entries.dedup_by(|later, earlier| {
|
||||
if later.0 == earlier.0 {
|
||||
earlier.1 = later.1;
|
||||
true
|
||||
} else {
|
||||
false
|
||||
}
|
||||
});
|
||||
Self { entries }
|
||||
}
|
||||
}
|
||||
impl Extend<(Cow<'static, str>, f64)> for NumericFactorMap {
|
||||
fn extend<T: IntoIterator<Item = (Cow<'static, str>, f64)>>(&mut self, iter: T) {
|
||||
let mut incoming: Self = iter.into_iter().collect();
|
||||
if incoming.is_empty() {
|
||||
return;
|
||||
}
|
||||
if self.is_empty() {
|
||||
*self = incoming;
|
||||
return;
|
||||
}
|
||||
if self.entries.last().unwrap().0 < incoming.entries[0].0 {
|
||||
self.entries.append(&mut incoming.entries);
|
||||
return;
|
||||
}
|
||||
// Merge sorted sets in linear time; wide factor batches must not shift
|
||||
// the existing vector once per field. Existing keys keep their identity.
|
||||
let mut merged = Vec::with_capacity(self.len() + incoming.len());
|
||||
let mut old = std::mem::take(&mut self.entries).into_iter().peekable();
|
||||
let mut new = incoming.entries.into_iter().peekable();
|
||||
while let (Some(left), Some(right)) = (old.peek(), new.peek()) {
|
||||
match left.0.cmp(&right.0) {
|
||||
std::cmp::Ordering::Less => merged.push(old.next().unwrap()),
|
||||
std::cmp::Ordering::Greater => merged.push(new.next().unwrap()),
|
||||
std::cmp::Ordering::Equal => {
|
||||
let (key, _) = old.next().unwrap();
|
||||
merged.push((key, new.next().unwrap().1));
|
||||
}
|
||||
}
|
||||
}
|
||||
merged.extend(old);
|
||||
merged.extend(new);
|
||||
self.entries = merged;
|
||||
}
|
||||
}
|
||||
impl<const N: usize> From<[(Cow<'static, str>, f64); N]> for NumericFactorMap {
|
||||
fn from(entries: [(Cow<'static, str>, f64); N]) -> Self {
|
||||
entries.into_iter().collect()
|
||||
}
|
||||
}
|
||||
impl From<BTreeMap<Cow<'static, str>, f64>> for NumericFactorMap {
|
||||
fn from(entries: BTreeMap<Cow<'static, str>, f64>) -> Self {
|
||||
Self {
|
||||
entries: entries.into_iter().collect(),
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
impl Serialize for NumericFactorMap {
|
||||
fn serialize<S: Serializer>(&self, serializer: S) -> Result<S::Ok, S::Error> {
|
||||
let mut map = serializer.serialize_map(Some(self.len()))?;
|
||||
for (key, value) in self {
|
||||
map.serialize_entry(key, value)?;
|
||||
}
|
||||
map.end()
|
||||
}
|
||||
}
|
||||
impl<'de> Deserialize<'de> for NumericFactorMap {
|
||||
fn deserialize<D: Deserializer<'de>>(deserializer: D) -> Result<Self, D::Error> {
|
||||
struct Fields;
|
||||
impl<'de> Visitor<'de> for Fields {
|
||||
type Value = NumericFactorMap;
|
||||
fn expecting(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
|
||||
f.write_str("a numeric factor map")
|
||||
}
|
||||
fn visit_map<A: MapAccess<'de>>(self, mut map: A) -> Result<Self::Value, A::Error> {
|
||||
let mut entries = Vec::new();
|
||||
while let Some((key, value)) = map.next_entry::<String, f64>()? {
|
||||
entries.push((Cow::Owned(key), value));
|
||||
}
|
||||
Ok(entries.into_iter().collect())
|
||||
}
|
||||
}
|
||||
deserializer.deserialize_map(Fields)
|
||||
}
|
||||
}
|
||||
|
||||
#[cfg(test)]
|
||||
mod tests {
|
||||
use super::*;
|
||||
|
||||
#[test]
|
||||
fn updates_order_removal_and_values_match_tree_map() {
|
||||
let mut flat = NumericFactorMap::new();
|
||||
let mut tree = BTreeMap::new();
|
||||
let mut seed = 71_u64;
|
||||
for index in 0..10000 {
|
||||
seed = seed.wrapping_mul(6364136223846793005).wrapping_add(1);
|
||||
let key: Cow<'static, str> = Cow::Owned(format!("factor_{:04}", (seed >> 32) % 1000));
|
||||
let value = (index as f64 - 5000.0) / 7.0;
|
||||
if index % 11 == 0 {
|
||||
assert_eq!(flat.remove(key.as_ref()), tree.remove(key.as_ref()));
|
||||
} else {
|
||||
assert_eq!(flat.insert(key.clone(), value), tree.insert(key, value));
|
||||
}
|
||||
}
|
||||
assert_eq!(
|
||||
flat.iter().collect::<Vec<_>>(),
|
||||
tree.iter().collect::<Vec<_>>()
|
||||
);
|
||||
flat.retain(|_, value| *value > 100.0);
|
||||
tree.retain(|_, value| *value > 100.0);
|
||||
assert_eq!(
|
||||
flat.iter().collect::<Vec<_>>(),
|
||||
tree.iter().collect::<Vec<_>>()
|
||||
);
|
||||
assert_eq!(
|
||||
std::mem::size_of::<NumericFactorMap>(),
|
||||
std::mem::size_of_val(&tree)
|
||||
);
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn bulk_load_is_sorted_and_keeps_last_value_for_each_field() {
|
||||
let input = vec![
|
||||
(Cow::Borrowed("z"), 2.0),
|
||||
(Cow::Borrowed("a"), 1.0),
|
||||
(Cow::Borrowed("z"), 3.0),
|
||||
(Cow::Borrowed("z"), 4.0),
|
||||
];
|
||||
let flat: NumericFactorMap = input.clone().into_iter().collect();
|
||||
let tree: BTreeMap<_, _> = input.into_iter().collect();
|
||||
assert_eq!(
|
||||
flat.iter().collect::<Vec<_>>(),
|
||||
tree.iter().collect::<Vec<_>>()
|
||||
);
|
||||
assert_eq!(flat["z"], 4.0);
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn serialization_keeps_the_map_contract_and_precise_numbers() {
|
||||
let input = [
|
||||
(Cow::Borrowed("zero"), -0.0),
|
||||
(Cow::Borrowed("tiny"), 1.0000000000000002),
|
||||
(Cow::Borrowed("large"), 9.123456789123456e20),
|
||||
];
|
||||
let flat = NumericFactorMap::from(input.clone());
|
||||
let tree = BTreeMap::from(input);
|
||||
let json = serde_json::to_string(&flat).unwrap();
|
||||
assert_eq!(json, serde_json::to_string(&tree).unwrap());
|
||||
let decoded: NumericFactorMap = serde_json::from_str(&json).unwrap();
|
||||
for (key, value) in &flat {
|
||||
assert_eq!(value.to_bits(), decoded[key.as_ref()].to_bits());
|
||||
}
|
||||
let duplicate: NumericFactorMap = serde_json::from_str(r#"{"x":1,"x":2}"#).unwrap();
|
||||
assert_eq!(duplicate["x"], 2.0);
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn clone_does_not_share_mutable_values() {
|
||||
let original = NumericFactorMap::from([(Cow::Borrowed("signal"), 1.0)]);
|
||||
let mut copy = original.clone();
|
||||
*copy.get_mut("signal").unwrap() = 0.0;
|
||||
copy.insert(Cow::Borrowed("other"), 2.0);
|
||||
assert_eq!(original["signal"], 1.0);
|
||||
assert!(!original.contains_key("other"));
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn wide_batch_merge_matches_tree_and_preserves_old_key_ownership() {
|
||||
let entries = (0..4096)
|
||||
.map(|index| (Cow::Owned(format!("f_{index:05}")), index as f64))
|
||||
.collect::<Vec<_>>();
|
||||
let mut flat: NumericFactorMap = entries.clone().into_iter().collect();
|
||||
let mut tree = BTreeMap::from_iter(entries);
|
||||
flat.insert(Cow::Borrowed("shared"), -0.0);
|
||||
tree.insert(Cow::Borrowed("shared"), -0.0);
|
||||
let incoming = (2048..8192)
|
||||
.rev()
|
||||
.map(|index| (Cow::Owned(format!("f_{index:05}")), -(index as f64)))
|
||||
.chain([(Cow::Owned("shared".to_owned()), 1.0)])
|
||||
.collect::<Vec<_>>();
|
||||
flat.extend(incoming.clone());
|
||||
tree.extend(incoming);
|
||||
assert_eq!(
|
||||
flat.iter().collect::<Vec<_>>(),
|
||||
tree.iter().collect::<Vec<_>>()
|
||||
);
|
||||
assert!(matches!(flat.keys().last(), Some(Cow::Borrowed("shared"))));
|
||||
flat.extend([(Cow::Borrowed("zz"), f64::NAN)]);
|
||||
assert!(flat["zz"].is_nan());
|
||||
flat.extend(std::iter::empty());
|
||||
assert_eq!(flat.len(), tree.len() + 1);
|
||||
}
|
||||
}
|
||||
Reference in New Issue
Block a user