wathiede/letterbox
1
0
Fork 0
Code Issues 1 Pull Requests 1 Actions Releases Wiki Activity

server: fix pagination counts for tantivy results

Browse Source
This commit is contained in:
Bill Thiede 2024-10-27 10:55:49 -07:00
parent 88f8a9d537
commit 93073c9602
2 changed files with 30 additions and 28 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
server/src/lib.rs
View File

@ -605,7 +605,7 @@ fn compute_offset_limit(
} }
} }
#[derive(Debug)] #[derive(Debug, Default)]
pub struct Query { pub struct Query {
pub unread_only: bool, pub unread_only: bool,
pub tags: Vec<String>, pub tags: Vec<String>,

56
server/src/tantivy.rs
View File

@ -57,7 +57,7 @@ impl TantivyConnection {
); );
let start_time = std::time::Instant::now(); let start_time = std::time::Instant::now();
let (searcher, _query) = self.searcher_and_query("")?; let (searcher, _query) = self.searcher_and_query(&Query::default())?;
let docs = searcher.search(&AllQuery, &DocSetCollector)?; let docs = searcher.search(&AllQuery, &DocSetCollector)?;
let uid = self.get_index()?.schema().get_field("uid")?; let uid = self.get_index()?.schema().get_field("uid")?;
let t_uids: Vec<_> = docs let t_uids: Vec<_> = docs
@ -187,7 +187,7 @@ impl TantivyConnection {
} }
fn searcher_and_query( fn searcher_and_query(
&self, &self,
term: &str, query: &Query,
) -> Result<(Searcher, Box<dyn query::Query>), ServerError> { ) -> Result<(Searcher, Box<dyn query::Query>), ServerError> {
let index = self.get_index()?; let index = self.get_index()?;
let reader = index.reader()?; let reader = index.reader()?;
@ -197,20 +197,41 @@ impl TantivyConnection {
let summary = schema.get_field("summary")?; let summary = schema.get_field("summary")?;
let query_parser = QueryParser::for_index(&index, vec![title, summary]); let query_parser = QueryParser::for_index(&index, vec![title, summary]);
// Tantivy uses '*' to match all docs, not empty string // Tantivy uses '*' to match all docs, not empty string
let term = &query.remainder.join(" ");
let term = if term.is_empty() { "*" } else { term }; let term = if term.is_empty() { "*" } else { term };
info!("query_parser('{term}')"); info!("query_parser('{term}')");
let query = query_parser.parse_query(&term)?;
Ok((searcher, query)) let tantivy_query = query_parser.parse_query(&term)?;
let tag = self.get_index()?.schema().get_field("tag")?;
let is_read = self.get_index()?.schema().get_field("is_read")?;
let mut terms = vec![(Occur::Must, tantivy_query)];
for t in &query.tags {
let facet = Facet::from(&format!("/{t}"));
let facet_term = Term::from_facet(tag, &facet);
let facet_term_query = Box::new(TermQuery::new(facet_term, IndexRecordOption::Basic));
terms.push((Occur::Must, facet_term_query));
}
if query.unread_only {
info!("searching for unread only");
let term = Term::from_field_bool(is_read, false);
terms.push((
Occur::Must,
Box::new(TermQuery::new(term, IndexRecordOption::Basic)),
));
}
let search_query = BooleanQuery::new(terms);
Ok((searcher, Box::new(search_query)))
} }
pub async fn count(&self, query: &Query) -> Result<usize, ServerError> { pub async fn count(&self, query: &Query) -> Result<usize, ServerError> {
if !is_tantivy_query(query) { if !is_tantivy_query(query) {
return Ok(0); return Ok(0);
} }
info!("tantivy::count {query:?}");
use tantivy::collector::Count; use tantivy::collector::Count;
let term = query.remainder.join(" "); let term = query.remainder.join(" ");
let (searcher, query) = self.searcher_and_query(&term)?; let (searcher, query) = self.searcher_and_query(&query)?;
Ok(searcher.search(&query, &Count)?) Ok(searcher.search(&query, &Count)?)
} }
pub async fn search( pub async fn search(
@ -233,27 +254,8 @@ impl TantivyConnection {
limit = limit + 1; limit = limit + 1;
} }
let term = query.remainder.join(" "); let (searcher, search_query) = self.searcher_and_query(&query)?;
let (searcher, tantivy_query) = self.searcher_and_query(&term)?; info!("Tantivy::search(query '{query:?}', off {offset}, lim {limit}, search_query {search_query:?})");
let tag = self.get_index()?.schema().get_field("tag")?;
let is_read = self.get_index()?.schema().get_field("is_read")?;
let mut terms = vec![(Occur::Must, tantivy_query)];
for t in &query.tags {
let facet = Facet::from(&format!("/{t}"));
let facet_term = Term::from_facet(tag, &facet);
let facet_term_query = Box::new(TermQuery::new(facet_term, IndexRecordOption::Basic));
terms.push((Occur::Must, facet_term_query));
}
if query.unread_only {
info!("searching for unread only");
let term = Term::from_field_bool(is_read, false);
terms.push((
Occur::Must,
Box::new(TermQuery::new(term, IndexRecordOption::Basic)),
));
}
let search_query = BooleanQuery::new(terms);
info!("Tantivy::search(term '{term}', off {offset}, lim {limit}, search_query {search_query:?})");
let top_docs = searcher.search( let top_docs = searcher.search(
&search_query, &search_query,
&TopDocs::with_limit(limit as usize) &TopDocs::with_limit(limit as usize)