format

python/src/lib.rs

@@ -177,14 +177,14 @@ fn json_to_schema(json: &str) -> PyResult<PyArrowType<ArrowSchema>> {
 #[pyfunction]
 pub fn language_model_home() -> PyResult<String> {
     let Some(p) = lance_index::scalar::inverted::language_model_home() else {
-        return Err(pyo3::exceptions::PyValueError::new_err(format!(
-            "Failed to get language model home"
-        )));
+        return Err(pyo3::exceptions::PyValueError::new_err(
+            "Failed to get language model home",
+        ));
     };
     let Some(pstr) = p.to_str() else {
-        return Err(pyo3::exceptions::PyValueError::new_err(format!(
-            "Failed to convert language model home to str"
-        )));
+        return Err(pyo3::exceptions::PyValueError::new_err(
+            "Failed to convert language model home to str",
+        ));
     };
     Ok(String::from(pstr))
 }

rust/lance-index/src/scalar/inverted/tokenizer.rs

@@ -195,7 +195,7 @@ pub fn language_model_home() -> Option<PathBuf> {
 #[cfg(feature = "tokenizer-common")]
 trait TokenizerBuilder: Sized {
     type Config: serde::de::DeserializeOwned + Default;
-    fn load(p: &PathBuf) -> Result<Self> {
+    fn load(p: &std::path::Path) -> Result<Self> {
         if !p.is_dir() {
             return Err(Error::io(
                 format!("{} is not a valid directory", p.display()),
@@ -214,7 +214,7 @@ trait TokenizerBuilder: Sized {
         Self::new(config, p)
     }
 
-    fn new(config: Self::Config, root: &PathBuf) -> Result<Self>;
+    fn new(config: Self::Config, root: &std::path::Path) -> Result<Self>;
 
     fn build(&self) -> Result<tantivy::tokenizer::TextAnalyzerBuilder>;
 }

rust/lance-index/src/scalar/inverted/tokenizer/jieba.rs

@@ -1,28 +1,19 @@
 // SPDX-License-Identifier: Apache-2.0
 // SPDX-FileCopyrightText: Copyright The Lance Authors
 
-use std::path::PathBuf;
+use std::path::{Path, PathBuf};
 
 use super::TokenizerBuilder;
 use lance_core::{Error, Result};
 use serde::{Deserialize, Serialize};
 use snafu::{location, Location};
 
-#[derive(Serialize, Deserialize)]
+#[derive(Serialize, Deserialize, Default)]
 pub struct JiebaConfig {
     main: Option<String>,
     users: Option<Vec<String>>,
 }
 
-impl Default for JiebaConfig {
-    fn default() -> Self {
-        Self {
-            main: Default::default(),
-            users: Default::default(),
-        }
-    }
-}
-
 pub struct JiebaBuilder {
     root: PathBuf,
     config: JiebaConfig,
@@ -47,10 +38,10 @@ impl JiebaBuilder {
 impl TokenizerBuilder for JiebaBuilder {
     type Config = JiebaConfig;
 
-    fn new(config: Self::Config, root: &PathBuf) -> Result<Self> {
-        Ok(JiebaBuilder {
+    fn new(config: Self::Config, root: &Path) -> Result<Self> {
+        Ok(Self {
             config,
-            root: root.clone(),
+            root: root.to_path_buf(),
         })
     }
 

rust/lance-index/src/scalar/inverted/tokenizer/lindera.rs

@@ -1,7 +1,7 @@
 // SPDX-License-Identifier: Apache-2.0
 // SPDX-FileCopyrightText: Copyright The Lance Authors
 
-use std::path::PathBuf;
+use std::path::{Path, PathBuf};
 
 use super::TokenizerBuilder;
 use lance_core::{Error, Result};
@@ -17,23 +17,13 @@ use serde::{Deserialize, Serialize};
 use serde_json::{Map, Value};
 use snafu::{location, Location};
 
-#[derive(Serialize, Deserialize)]
+#[derive(Serialize, Deserialize, Default)]
 pub struct LinderaConfig {
     main: Option<String>,
     user: Option<String>,
     user_kind: Option<String>,
 }
 
-impl Default for LinderaConfig {
-    fn default() -> Self {
-        Self {
-            main: Default::default(),
-            user: Default::default(),
-            user_kind: Default::default(),
-        }
-    }
-}
-
 pub struct LinderaBuilder {
     root: PathBuf,
     config: LinderaConfig,
@@ -73,10 +63,10 @@ impl LinderaBuilder {
 impl TokenizerBuilder for LinderaBuilder {
     type Config = LinderaConfig;
 
-    fn new(config: Self::Config, root: &PathBuf) -> Result<Self> {
-        Ok(LinderaBuilder {
+    fn new(config: Self::Config, root: &Path) -> Result<Self> {
+        Ok(Self {
             config,
-            root: root.clone(),
+            root: root.to_path_buf(),
         })
     }