[feat] 0.23.7

izihawa · May 11, 2024 · 8b06507 · 8b06507
1 parent f1d4dc4
commit 8b06507
Show file tree

Hide file tree

Showing 6 changed files with 280 additions and 6 deletions.
diff --git a/aiosumma/aiosumma/client.py b/aiosumma/aiosumma/client.py
@@ -13,7 +13,6 @@
 from .proto import index_service_pb2 as index_service_pb
 from .proto import query_pb2 as query_pb
 from .proto import reflection_service_pb2 as reflection_service_pb
-from .proto import search_service_pb2 as search_service_pb
 from .proto.consumer_service_pb2_grpc import ConsumerApiStub
 from .proto.index_service_pb2_grpc import IndexApiStub
 from .proto.reflection_service_pb2_grpc import ReflectionApiStub

diff --git a/poetry.lock b/poetry.lock
diff --git a/pyproject.toml b/pyproject.toml
@@ -7,6 +7,7 @@ readme = "README.md"
 
 [tool.poetry.dependencies]
 python = "^3.9"
+aiogrpcclient = "^1.3.5"
 
 
 [build-system]

diff --git a/summa-core/src/components/query_parser/morphology/mod.rs b/summa-core/src/components/query_parser/morphology/mod.rs
@@ -1,3 +1,4 @@
+use std::collections::HashSet;
 use summa_proto::proto;
 use tantivy::query::{DisjunctionMaxQuery, Query, TermQuery};
 use tantivy::schema::{Field, FieldType, IndexRecordOption};
@@ -20,12 +21,14 @@ pub trait Morphology: MorphologyClone + Send + Sync {
             let term = cast_field_to_term(field, full_path, field_type, text, false);
             return Box::new(TermQuery::new(term, IndexRecordOption::WithFreqs)) as Box<dyn Query>;
         };
-        let mut terms = vec![text.to_string()];
+        let mut terms = HashSet::new();
+        terms.insert(text.to_string());
         if let Some(other_tense_text) = self.derive_tenses(text) {
-            terms.push(other_tense_text);
+            terms.insert(other_tense_text);
         }
-        let corrected_spellings: Vec<_> = terms.iter().filter_map(|t| self.derive_spelling(t)).collect();
-        terms.extend(corrected_spellings);
+        terms.extend(terms.iter().filter_map(|t| self.derive_spelling(t)).collect::<Vec<_>>());
+        terms.extend(terms.iter().filter_map(|t| self.derive_tenses(t)).collect::<Vec<_>>());
+        let terms = Vec::from_iter(terms.into_iter());
         if terms.len() == 1 {
             Box::new(TermQuery::new(
                 cast_field_to_term(field, full_path, field_type, &terms[0], false),

diff --git a/summa-core/src/components/query_parser/summa_ql.rs b/summa-core/src/components/query_parser/summa_ql.rs
@@ -1196,6 +1196,11 @@ mod tests {
         assert_eq!(format!("{:?}", query), "Ok(BooleanQuery { subqueries: [(Should, TermQuery(Term(field=0, type=Str, \"red1\"))), (Should, DisjunctionMaxQuery { disjuncts: [TermQuery(Term(field=0, type=Str, \"search\")), TermQuery(Term(field=0, type=Str, \"searches\"))], tie_breaker: 0.3 }), (Should, DisjunctionMaxQuery { disjuncts: [TermQuery(Term(field=0, type=Str, \"engine\")), TermQuery(Term(field=0, type=Str, \"engines\"))], tie_breaker: 0.3 }), (Should, TermQuery(Term(field=0, type=Str, \"going\")))] })");
         let query = query_parser.parse_query("iso 34-1:2022");
         assert_eq!(format!("{:?}", query), "Ok(BooleanQuery { subqueries: [(Should, DisjunctionMaxQuery { disjuncts: [TermQuery(Term(field=0, type=Str, \"iso\")), TermQuery(Term(field=0, type=Str, \"isos\"))], tie_breaker: 0.3 }), (Should, TermQuery(Term(field=0, type=Str, \"34\"))), (Should, TermQuery(Term(field=0, type=Str, \"1\")))] })");
+        let query = query_parser.parse_query("http://www.aaa.com-inward-record-b-c\naaaa");
+        assert_eq!(
+            format!("{:?}", query),
+            ""
+        );
     }
 
     #[test]

diff --git a/summa-server/Cargo.toml b/summa-server/Cargo.toml
@@ -1,7 +1,7 @@
 [package]
 edition = "2021"
 name = "summa-server"
-version = "0.23.6"
+version = "0.23.7"
 license-file = "LICENSE"
 description = "Fast full-text search server"
 homepage = "https://github.com/izihawa/summa"