inbuxa/inbuxa-server
1
0
Fork 0
Code Issues Pull Requests 2 Actions Packages Projects Releases 3 Wiki Activity

Search: find addresses by local part, domain or name on PostgreSQL and MySQL
ci / fork-checks (pull_request) Successful in 43s
Details
ci / build (pull_request) Successful in 4m20s
Details

Browse Source 
A 3-node PostgreSQL rehearsal found IMAP SEARCH FROM "noreply" matched
0-2 messages where RocksDB matched 23 of 930. The message indexer hands
each address and display name of From/To/Cc/Bcc to the search store as
keyword text (Language::None). The built-in index splits keyword text
into lowercase runs of alphanumerics, so an address is found by its full
form, its local part, its domain or a display-name word. The SQL
backends didn't:

- PostgreSQL's text parser keeps "[email protected]" as one email
  token (host names and URLs likewise), so neither "noreply" nor
  "amazon.com" ever matched it. Keyword text is now split the same way
  as the built-in index (SpaceTokenizer) before to_tsvector on insert
  and before plainto_tsquery/phraseto_tsquery on search, still under
  the 'simple' configuration, so the GIN index keeps serving the query.
  The sort columns keep the raw text.
- MySQL's FULLTEXT parser already splits on punctuation, but InnoDB
  never indexes its stopwords ("com", "de", "www", ...) or words under
  innodb_ft_min_token_size (3), and a required +word it hasn't indexed
  matches no row. So "amazon.com", "[email protected]" or "jane doe" found
  nothing. Those words are now matched with a word-boundary REGEXP on
  the rows the indexed words select. In language text (bodies,
  subjects) they are dropped when other words remain, and only checked
  when nothing else is left, so "the invoice" no longer finds nothing
  either.

Existing PostgreSQL search indexes hold the old single-token vectors and
need a reindex (the reindexAccounts task) before address searches find
old messages. MySQL needs none: only the query changed.

store::search_tests gains test_address_search: five messages, 28
FROM/TO/CC/BCC searches by full address, local part, domain, domain
labels, display name and hyphenated local part, plus a TEXT-style OR,
with the same expected ids on every backend. It passes on RocksDB,
SQLite, PostgreSQL and MySQL; on main it fails on PostgreSQL (From
"noreply") and MySQL (From "[email protected]").
This commit is contained in:
jcoffey-dev
2026-09-24 11:25:25 -07:00
parent 9311c1a38b
commit 639a415a4f
3 changed files with 312 additions and 23 deletions
Download Patch File Download Diff File Expand all files Collapse all files
tests/src/store/query.rs
+157
View File
@@ -128,6 +128,11 @@ pub async fn test(test: &TestServer) {
println!("Running trace document tests...");
test_trace_documents(store.clone()).await;
// inbuxa: address fields match by full address, local part, domain and
// display name on every backend
println!("Running address search tests...");
test_address_search(store.clone()).await;
// Large document insert test
println!("Running large document insert tests...");
let mut large_text = String::with_capacity(20 * 1024 * 1024);
@@ -972,3 +977,155 @@ async fn test_trace_documents(store: SearchStore) {
.unwrap();
}
}
// inbuxa: the message indexer passes each display name and each address of
// From/To/Cc/Bcc as keyword text (Language::None). The built-in index splits
// that text into words, so an address is found by its full form, its local
// part, its domain or a display-name word; PostgreSQL kept the whole address
// as one token and MySQL dropped stopwords such as "com" and words under three
// characters. The expected results below are the built-in (RocksDB/SQLite)
// results and must be the same on every backend.
async fn test_address_search(store: SearchStore) {
const ACCOUNT_ID: u32 = 7;
let messages: [[&[(&str, &str)]; 4]; 5] = [
// From, To, Cc, Bcc
[
&[("Amazon.com", "[email protected]")],
&[("Jane Doe", "[email protected]")],
&[],
&[],
],
[
&[("", "[email protected]")],
&[("", "[email protected]")],
&[("Jane Doe", "[email protected]")],
&[],
],
[
&[("GitHub", "[email protected]")],
&[("Jo Li", "[email protected]")],
&[],
&[("Audit", "[email protected]")],
],
[
&[("Jane Doe", "[email protected]")],
&[("Amazon Web Services", "[email protected]")],
&[("Bob", "[email protected]")],
&[("", "[email protected]")],
],
[
&[("Newsletter", "[email protected]")],
&[("", "[email protected]")],
&[],
&[],
],
];
let fields = [
EmailSearchField::From,
EmailSearchField::To,
EmailSearchField::Cc,
EmailSearchField::Bcc,
];
let mut documents = Vec::new();
let mut mask = RoaringBitmap::new();
for (document_id, message) in messages.iter().enumerate() {
let mut document = IndexDocument::new(SearchIndex::Email)
.with_account_id(ACCOUNT_ID)
.with_document_id(document_id as u32);
for (field, addresses) in fields.iter().zip(message.iter()) {
for (name, address) in addresses.iter() {
if !name.is_empty() {
document.index_text(field.clone(), name, Language::None);
}
document.index_text(field.clone(), address, Language::None);
}
}
document.index_unsigned(EmailSearchField::ReceivedAt, document_id as u64);
documents.push(document);
mask.insert(document_id as u32);
}
store.index(documents).await.unwrap();
if let SearchStore::ElasticSearch(store) = &store {
store.refresh_index(SearchIndex::Email).await.unwrap();
}
for (field, text, expected) in [
// full address
(EmailSearchField::From, "[email protected]", vec![0u32]),
(EmailSearchField::To, "[email protected]", vec![0]),
(EmailSearchField::Cc, "[email protected]", vec![1]),
(EmailSearchField::Bcc, "[email protected]", vec![3]),
(EmailSearchField::To, "[email protected]", vec![1, 2]),
// local part
(EmailSearchField::From, "noreply", vec![0, 2]),
(EmailSearchField::To, "jo", vec![1, 2]),
(EmailSearchField::Cc, "bob", vec![3]),
(EmailSearchField::Bcc, "audit", vec![2]),
// domain
(EmailSearchField::From, "amazon.com", vec![0, 1]),
(EmailSearchField::From, "amazon", vec![0, 1]),
(EmailSearchField::To, "example.org", vec![0, 4]),
(EmailSearchField::To, "io.de", vec![1, 2]),
(EmailSearchField::Cc, "example.net", vec![3]),
(EmailSearchField::Bcc, "example.org", vec![2]),
(EmailSearchField::From, "www.example.com", vec![4]),
(EmailSearchField::From, "com", vec![0, 1, 2, 4]),
// display name
(EmailSearchField::From, "Jane", vec![3]),
(EmailSearchField::From, "jane doe", vec![3]),
(EmailSearchField::To, "Web Services", vec![3]),
(EmailSearchField::To, "Li", vec![2]),
(EmailSearchField::Cc, "Doe", vec![1]),
(EmailSearchField::Bcc, "Audit", vec![2]),
// hyphenated local part
(EmailSearchField::From, "shipment-tracking", vec![1]),
(EmailSearchField::From, "tracking", vec![1]),
// no match
(EmailSearchField::From, "amazon.org", vec![]),
(EmailSearchField::To, "noreply", vec![]),
(EmailSearchField::Bcc, "jane", vec![]),
] {
let ids = store
.query_account(
SearchQuery::new(SearchIndex::Email)
.with_filters(vec![
SearchFilter::eq(SearchField::AccountId, ACCOUNT_ID),
SearchFilter::has_keyword(field.clone(), text),
])
.with_comparator(SearchComparator::ascending(EmailSearchField::ReceivedAt))
.with_mask(mask.clone()),
)
.await
.unwrap();
assert_eq!(ids, expected, "{field:?} {text:?}");
}
// TEXT-style search across all address fields
let ids = store
.query_account(
SearchQuery::new(SearchIndex::Email)
.with_filters(vec![
SearchFilter::eq(SearchField::AccountId, ACCOUNT_ID),
SearchFilter::Or,
SearchFilter::has_keyword(EmailSearchField::From, "example.org"),
SearchFilter::has_keyword(EmailSearchField::To, "example.org"),
SearchFilter::has_keyword(EmailSearchField::Cc, "example.org"),
SearchFilter::has_keyword(EmailSearchField::Bcc, "example.org"),
SearchFilter::End,
])
.with_comparator(SearchComparator::ascending(EmailSearchField::ReceivedAt))
.with_mask(mask.clone()),
)
.await
.unwrap();
assert_eq!(ids, vec![0, 1, 2, 3, 4]);
store
.unindex(
SearchQuery::new(SearchIndex::Email)
.with_filter(SearchFilter::eq(SearchField::AccountId, ACCOUNT_ID)),
)
.await
.unwrap();
}