Refactor reviews_for_fsrs function for improved performance

Replaced the previous implementation with a more efficient approach using a single loop and pre-allocated vectors. This change reduces the complexity of creating FSRSItems and enhances overall performance, especially for larger datasets.
This commit is contained in:
Jarrett Ye 2025-09-17 12:28:24 +08:00
parent b97fb45e06
commit ecd1cf45e9
No known key found for this signature in database
GPG key ID: EBFC55E0C1A352BB

View file

@ -481,24 +481,22 @@ pub(crate) fn reviews_for_fsrs(
let skip = if training { 1 } else { 0 };
// Convert the remaining entries into separate FSRSItems, where each item
// contains all reviews done until then.
let items: Vec<(RevlogId, FSRSItem)> = entries
.iter()
.enumerate()
.skip(skip)
.map(|(outer_idx, entry)| {
let reviews = entries
.iter()
.take(outer_idx + 1)
.enumerate()
.map(|(inner_idx, r)| FSRSReview {
rating: r.button_chosen as u32,
delta_t: delta_ts[inner_idx],
})
.collect();
(entry.id, FSRSItem { reviews })
})
.filter(|(_, item)| !training || item.reviews.last().unwrap().delta_t > 0)
.collect_vec();
let mut items = Vec::with_capacity(entries.len());
let mut current_reviews = Vec::with_capacity(entries.len());
for (idx, (entry, &delta_t)) in entries.iter().zip(delta_ts.iter()).enumerate() {
current_reviews.push(FSRSReview {
rating: entry.button_chosen as u32,
delta_t,
});
if idx >= skip {
if !training || current_reviews.last().unwrap().delta_t > 0 {
let item = FSRSItem {
reviews: current_reviews.clone(),
};
items.push((entry.id, item));
}
}
}
if items.is_empty() {
None
} else {