fix: universal compaction condition (#97)

Signed-off-by: Alex Chi <iskyzh@gmail.com>

mini-lsm/src/compact/tiered.rs

@@ -16,6 +16,7 @@ pub struct TieredCompactionOptions {
     pub max_size_amplification_percent: usize,
     pub size_ratio: usize,
     pub min_merge_width: usize,
+    pub max_merge_width: Option<usize>,
 }
 
 pub struct TieredCompactionController {
@@ -61,25 +62,29 @@ impl TieredCompactionController {
         for id in 0..(snapshot.levels.len() - 1) {
             size += snapshot.levels[id].1.len();
             let next_level_size = snapshot.levels[id + 1].1.len();
-            let current_size_ratio = size as f64 / next_level_size as f64;
-            if current_size_ratio >= size_ratio_trigger && id + 2 >= self.options.min_merge_width {
+            let current_size_ratio = next_level_size as f64 / size as f64;
+            if current_size_ratio > size_ratio_trigger && id + 1 >= self.options.min_merge_width {
                 println!(
-                    "compaction triggered by size ratio: {}",
-                    current_size_ratio * 100.0
+                    "compaction triggered by size ratio: {} > {}",
+                    current_size_ratio * 100.0,
+                    size_ratio_trigger * 100.0
                 );
                 return Some(TieredCompactionTask {
                     tiers: snapshot
                         .levels
                         .iter()
-                        .take(id + 2)
+                        .take(id + 1)
                         .cloned()
                         .collect::<Vec<_>>(),
-                    bottom_tier_included: id + 2 >= snapshot.levels.len(),
+                    bottom_tier_included: id + 1 >= snapshot.levels.len(),
                 });
             }
         }
         // trying to reduce sorted runs without respecting size ratio
-        let num_tiers_to_take = snapshot.levels.len() - self.options.num_tiers + 2;
+        let num_tiers_to_take = snapshot
+            .levels
+            .len()
+            .min(self.options.max_merge_width.unwrap_or(usize::MAX));
         println!("compaction triggered by reducing sorted runs");
         return Some(TieredCompactionTask {
             tiers: snapshot

mini-lsm/src/table/bloom.rs

@@ -87,7 +87,7 @@ impl Bloom {
         filter.resize(nbytes, 0);
         for h in keys {
             let mut h = *h;
-            let delta = (h >> 17) | (h << 15);
+            let delta = h.rotate_left(15);
             for _ in 0..k {
                 let bit_pos = (h as usize) % nbits;
                 filter.set_bit(bit_pos, true);
@@ -107,7 +107,7 @@ impl Bloom {
             true
         } else {
             let nbits = self.filter.bit_len();
-            let delta = (h >> 17) | (h << 15);
+            let delta = h.rotate_left(15);
             for _ in 0..self.k {
                 let bit_pos = h % (nbits as u32);
                 if !self.filter.get_bit(bit_pos as usize) {

mini-lsm/src/tests/harness.rs

@@ -368,6 +368,7 @@ pub fn check_compaction_ratio(storage: Arc<MiniLsm>) {
             max_size_amplification_percent,
             size_ratio,
             min_merge_width,
+            ..
         }) => {
             let size_ratio_trigger = (100.0 + size_ratio as f64) / 100.0;
             assert_eq!(l0_sst_num, 0);

mini-lsm/src/tests/week2_day3.rs

@@ -18,6 +18,7 @@ fn test_integration() {
                 max_size_amplification_percent: 200,
                 size_ratio: 1,
                 min_merge_width: 2,
+                max_merge_width: None,
             },
         )),
     )

mini-lsm/src/tests/week2_day5.rs

@@ -29,6 +29,7 @@ fn test_integration_tiered() {
         max_size_amplification_percent: 200,
         size_ratio: 1,
         min_merge_width: 3,
+        max_merge_width: None,
     }))
 }
 

mini-lsm/src/tests/week2_day6.rs

@@ -26,6 +26,7 @@ fn test_integration_tiered() {
         max_size_amplification_percent: 200,
         size_ratio: 1,
         min_merge_width: 3,
+        max_merge_width: None,
     }))
 }