snowykami/meilisearch
1
0
Fork 0
mirror of https://github.com/meilisearch/meilisearch.git synced 2025-03-06 13:51:03 +08:00
Code Issues Actions 7 Packages Projects Releases Wiki Activity

Fix the algorithm

Browse Source
This commit is contained in:
Kerollmops 2025-02-26 16:21:32 +01:00
parent 3dcf681d46
commit 21db4b6b29
No known key found for this signature in database
GPG Key ID: F250A4C4E3AE5F5F
1 changed files with 11 additions and 8 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

19
crates/milli/src/update/new/indexer/post_processing/facet_bulk.rs
View File

@ -76,8 +76,9 @@ fn compute_level(
field_id: FieldId, field_id: FieldId,
base_level: u8, base_level: u8,
) -> Result<Vec<grenad::Reader<BufReader<File>>>, crate::Error> { ) -> Result<Vec<grenad::Reader<BufReader<File>>>, crate::Error> {
let thread_count = rayon::current_num_threads();
let rtxns = iter::repeat_with(|| index.env.nested_read_txn(wtxn)) let rtxns = iter::repeat_with(|| index.env.nested_read_txn(wtxn))
.take(rayon::current_num_threads()) .take(thread_count)
.collect::<heed::Result<Vec<_>>>()?; .collect::<heed::Result<Vec<_>>>()?;
let range = { let range = {
@ -106,13 +107,9 @@ fn compute_level(
let start_of_group = i % FACET_GROUP_SIZE as usize == 0; let start_of_group = i % FACET_GROUP_SIZE as usize == 0;
let group_index = i / FACET_GROUP_SIZE as usize; let group_index = i / FACET_GROUP_SIZE as usize;
let group_for_thread = group_index % FACET_GROUP_SIZE as usize == thread_id; let group_for_thread = group_index % thread_count == thread_id;
if group_for_thread { if group_for_thread {
// Lazily decode the bitmaps we are interested in.
let value = lazy_value.decode().map_err(heed::Error::Decoding)?;
group_docids.push(value.bitmap);
if start_of_group { if start_of_group {
if let Some(left_bound) = left_bound.take() { if let Some(left_bound) = left_bound.take() {
// We store the bitmaps in a Vec this way we can use // We store the bitmaps in a Vec this way we can use
@ -122,7 +119,8 @@ fn compute_level(
// We also don't forget to store the group size corresponding // We also don't forget to store the group size corresponding
// to the number of entries merged in this group. // to the number of entries merged in this group.
ser_buffer.clear(); ser_buffer.clear();
ser_buffer.push(group_docids.len().try_into().unwrap()); let group_len: u8 = group_docids.len().try_into().unwrap();
ser_buffer.push(group_len);
let group_docids = mem::take(&mut group_docids); let group_docids = mem::take(&mut group_docids);
let docids = group_docids.into_iter().union(); let docids = group_docids.into_iter().union();
CboRoaringBitmapCodec::serialize_into_vec(&docids, &mut ser_buffer); CboRoaringBitmapCodec::serialize_into_vec(&docids, &mut ser_buffer);
@ -130,6 +128,10 @@ fn compute_level(
} }
left_bound = Some(key.left_bound); left_bound = Some(key.left_bound);
} }
// Lazily decode the bitmaps we are interested in.
let value = lazy_value.decode().map_err(heed::Error::Decoding)?;
group_docids.push(value.bitmap);
} }
} }
@ -137,7 +139,8 @@ fn compute_level(
ser_buffer.clear(); ser_buffer.clear();
// We don't forget to store the group size corresponding // We don't forget to store the group size corresponding
// to the number of entries merged in this group. // to the number of entries merged in this group.
ser_buffer.push(group_docids.len().try_into().unwrap()); let group_len: u8 = group_docids.len().try_into().unwrap();
ser_buffer.push(group_len);
let group_docids = group_docids.into_iter().union(); let group_docids = group_docids.into_iter().union();
CboRoaringBitmapCodec::serialize_into_vec(&group_docids, &mut ser_buffer); CboRoaringBitmapCodec::serialize_into_vec(&group_docids, &mut ser_buffer);
writer.insert(left_bound, &ser_buffer)?; writer.insert(left_bound, &ser_buffer)?;