filip/Stalwart
1
0
Fork 0
Code Issues Pull Requests Actions 1 Packages Projects Releases Wiki Activity

Spam filter performance and accuracy improvements (part 3)

Browse Source
This commit is contained in:
mdecimus
2025-12-02 18:11:01 +01:00
parent 72a175a434
commit e127ac0067
26 changed files with 1137 additions and 1395 deletions
Download Patch File Download Diff File Expand all files Collapse all files

48
crates/store/src/write/blob.rs
View File

@@ -143,8 +143,8 @@ impl Store {
let mut delete_keys = Vec::new();
let now = now();
self.iterate(
IterateParams::new(from_key, to_key).ascending().no_values(),
|key, _| {
IterateParams::new(from_key, to_key).ascending(),
|key, value| {
let hash = BlobHash::try_from_hash_slice(
key.get(0..BLOB_HASH_LEN)
.ok_or_else(|| trc::Error::corrupted_key(key, None, trc::location!()))?,
@@ -181,21 +181,35 @@ impl Store {
to: BlobLink::Temporary { until },
},
));
if account_id != u32::MAX {
delete_keys.push((
Some(account_id),
BlobOp::Quota {
hash: last_hash.clone(),
until,
},
));
delete_keys.push((
Some(account_id),
BlobOp::Undelete {
hash: last_hash.clone(),
until,
},
));
match value.first().copied() {
Some(BlobLink::QUOTA_LINK) => {
delete_keys.push((
Some(account_id),
BlobOp::Quota {
hash: last_hash.clone(),
until,
},
));
}
Some(BlobLink::UNDELETE_LINK) => {
delete_keys.push((
Some(account_id),
BlobOp::Undelete {
hash: last_hash.clone(),
until,
},
));
}
Some(BlobLink::SPAM_SAMPLE_LINK) => {
delete_keys.push((
Some(account_id),
BlobOp::SpamSample {
hash: last_hash.clone(),
until,
},
));
}
_ => {}
}
} else {
last_hash_is_linked = true;

31
crates/store/src/write/key.rs
View File

@@ -289,16 +289,6 @@ impl ValueClass {
.write(if *is_insert { 7u8 } else { 8u8 })
.write(document_id)
.write(index.to_u8()),
TaskQueueClass::SpamTrain {
due,
blob_hash,
learn_spam,
} => serializer
.write(due.inner())
.write(account_id)
.write(if *learn_spam { 1u8 } else { 2u8 })
.write(blob_hash.as_slice())
.write(document_id),
TaskQueueClass::SendAlarm {
due,
event_id,
@@ -348,15 +338,20 @@ impl ValueClass {
.write(*until),
},
BlobOp::Quota { hash, until } => serializer
.write(BlobLink::QUOTA_LINK)
.write(account_id)
.write(0u8)
.write::<&[u8]>(hash.as_ref())
.write(*until),
BlobOp::Undelete { hash, until } => serializer
.write(BlobLink::UNDELETE_LINK)
.write(account_id)
.write(1u8)
.write::<&[u8]>(hash.as_ref())
.write(*until),
BlobOp::SpamSample { hash, until } => serializer
.write(BlobLink::SPAM_SAMPLE_LINK)
.write(*until)
.write(account_id)
.write::<&[u8]>(hash.as_ref()),
},
ValueClass::Config(key) => serializer.write(key.as_slice()),
ValueClass::InMemory(lookup) => match lookup {
@@ -511,6 +506,12 @@ impl ValueClass {
}
}
impl BlobLink {
pub const QUOTA_LINK: u8 = 0;
pub const UNDELETE_LINK: u8 = 1;
pub const SPAM_SAMPLE_LINK: u8 = 2;
}
impl<T: AsRef<[u8]> + Sync + Send + Clone> Key for IndexKey<T> {
fn subspace(&self) -> u8 {
SUBSPACE_INDEXES
@@ -582,10 +583,10 @@ impl ValueClass {
BlobOp::Quota { .. } | BlobOp::Undelete { .. } => {
BLOB_HASH_LEN + U32_LEN + U64_LEN + 1
}
BlobOp::SpamSample { .. } => BLOB_HASH_LEN + U32_LEN + 2,
},
ValueClass::TaskQueue(e) => match e {
TaskQueueClass::UpdateIndex { .. } => (U64_LEN * 2) + 2,
TaskQueueClass::SpamTrain { .. } => BLOB_HASH_LEN + (U64_LEN * 2) + 1,
TaskQueueClass::SendAlarm { .. } | TaskQueueClass::MergeThreads { .. } => {
U64_LEN + (U32_LEN * 3) + 1
}
@@ -642,7 +643,9 @@ impl ValueClass {
ValueClass::TaskQueue { .. } => SUBSPACE_TASK_QUEUE,
ValueClass::Blob(op) => match op {
BlobOp::Commit { .. } | BlobOp::Link { .. } => SUBSPACE_BLOB_LINK,
BlobOp::Quota { .. } | BlobOp::Undelete { .. } => SUBSPACE_BLOB_EXTRA,
BlobOp::Quota { .. } | BlobOp::Undelete { .. } | BlobOp::SpamSample { .. } => {
SUBSPACE_BLOB_EXTRA
}
},
ValueClass::Config(_) => SUBSPACE_SETTINGS,
ValueClass::InMemory(lookup) => match lookup {

6
crates/store/src/write/mod.rs
View File

@@ -226,11 +226,6 @@ pub enum TaskQueueClass {
index: SearchIndex,
is_insert: bool,
},
SpamTrain {
due: TaskEpoch,
blob_hash: BlobHash,
learn_spam: bool,
},
SendAlarm {
due: TaskEpoch,
event_id: u16,
@@ -380,6 +375,7 @@ pub enum BlobOp {
Link { hash: BlobHash, to: BlobLink },
Quota { hash: BlobHash, until: u64 },
Undelete { hash: BlobHash, until: u64 },
SpamSample { hash: BlobHash, until: u64 },
}
#[derive(Debug, PartialEq, Clone, Eq, Hash)]