Merge pull request #2 from kqito/fix-compiled-too-big

Improve grep system

Cargo.toml

@@ -21,6 +21,7 @@ thiserror = "1.0.61"
 argh = "0.1.10"
 atty = "0.2.14"
 walkdir = "2"
+rayon = "1.10.0"
 
 [dev-dependencies]
 pretty_assertions = "1.4.0"

README.md

@@ -54,7 +54,6 @@ Positional Arguments:
 
 Options:
   -d, --debug       help
-  -u, --unique      unique Omit duplicate paths
   --help            display usage information
 ```
 

src/args.rs

@@ -14,11 +14,6 @@ pub struct Args {
     /// help
     pub debug: Option<bool>,
 
-    #[argh(switch, short = 'u')]
-    /// unique
-    /// Omit duplicate paths
-    pub unique: Option<bool>,
-
     #[argh(option, short = 'c')]
     /// current_dir
     /// Set the current directory
@@ -30,6 +25,12 @@ pub struct Args {
     /// ignore
     /// Ignore pattern
     pub ignore: Vec<String>,
+
+    #[argh(option, short = 't')]
+    /// filetype
+    /// Filter by type: file (f), directory (d/dir), symlink (l)
+    /// Default: file (f)
+    pub filetype: Vec<String>,
 }
 
 pub fn get_input() -> Result<Option<String>> {

src/grep/finder.rs

@@ -2,30 +2,26 @@ use regex::Regex;
 use std::fmt::Debug;
 use walkdir::{DirEntry, WalkDir};
 
+use super::params::Filetype;
+
 #[derive(Debug, Clone)]
 pub struct IOFinder {
     pub current_dir: String,
     pub ignore_pattern: Vec<String>,
 }
 
-#[derive(Debug, Clone)]
-pub enum ResourceType {
-    Directory,
-    File,
-}
-
 #[derive(Debug, Clone)]
 pub struct Resource {
     pub path: String,
-    pub resource_type: ResourceType,
+    pub filetype: Filetype,
 }
 
 #[derive(Debug, Clone)]
 pub struct Stats {
     pub resources: Vec<Resource>,
 }
 
-pub trait Finder: FinderClone + Debug {
+pub trait Finder: FinderClone + Debug + Sync + Send {
     fn current_dir(&mut self, current_dir: &str);
     fn ignore(&mut self, ignore: Vec<String>);
     fn find(&mut self) -> Stats;
@@ -87,12 +83,20 @@ impl Finder for IOFinder {
 
         for entry in walker.filter_entry(|e| !self.is_ignored(e)) {
             if let Ok(entry) = entry {
+                let path_str = entry.path().to_str().unwrap();
+                // omit "./" prefix
+                let path = if path_str.starts_with("./") {
+                    path_str[2..].to_string()
+                } else {
+                    path_str.to_string()
+                };
+
                 resources.push(Resource {
-                    path: entry.path().to_str().unwrap().to_string(),
-                    resource_type: if entry.file_type().is_dir() {
-                        ResourceType::Directory
+                    path,
+                    filetype: if entry.file_type().is_dir() {
+                        Filetype::Directory
                     } else {
-                        ResourceType::File
+                        Filetype::File
                     },
                 });
             }
@@ -102,23 +106,18 @@ impl Finder for IOFinder {
     }
 }
 
-impl Stats {
+impl Resource {
     pub fn as_regex(&self) -> Regex {
-        let mut matches: Vec<String> = self
-            .resources
-            .iter()
-            .map(|r| {
-                let path = r.path.replace("/", r"\/");
-                match r.resource_type {
-                    ResourceType::Directory => format!(r"{}", path),
-                    // Support for line and column numbers
-                    ResourceType::File => format!(r"{}(:\d+:\d+)?", path),
-                }
-            })
-            .collect();
-        matches.sort_by(|a, b| b.len().cmp(&a.len()));
-
-        let pattern = matches.join("|");
-        Regex::new(&pattern).unwrap()
+        match self.filetype {
+            Filetype::Directory => {
+                let path = self.path.replace("/", r"\/");
+                Regex::new(&format!(r"{}", path)).unwrap()
+            }
+            // Support for line and column numbers
+            Filetype::File => {
+                let path = self.path.replace("/", r"\/");
+                Regex::new(&format!(r"{}(:\d+:\d+)?", path)).unwrap()
+            }
+        }
     }
 }

src/grep/mod.rs

@@ -2,9 +2,13 @@ mod finder;
 pub mod params;
 mod tests;
 
-use params::GrepParams;
-
 use crate::output::{pretty_print, Status};
+use params::{Filetype, GrepParams};
+use rayon::prelude::*;
+use std::{
+    fmt::Debug,
+    sync::{Arc, RwLock},
+};
 
 #[derive(Debug, PartialEq)]
 pub enum GrepItemType {
@@ -18,68 +22,77 @@ pub struct GrepItem {
     pub line: Option<usize>,
     pub column: Option<usize>,
     pub item_type: GrepItemType,
+    pub filetype: Filetype,
 }
 
 /// Extract path in string message with regex
 pub fn grep(params: &GrepParams) -> Vec<GrepItem> {
-    let mut items: Vec<GrepItem> = Vec::new();
-    let mut finder = params.finder.clone();
-    finder.current_dir(&params.current_dir);
-    finder.ignore(params.ignore_pattern.clone());
-    let find_list = finder.find();
+    let finder = Arc::new(RwLock::new(params.finder.clone()));
+    let mut finder_lock = finder.write().unwrap();
+    finder_lock.current_dir(&params.current_dir);
+    finder_lock.ignore(params.ignore_pattern.clone());
+    let find_list = finder_lock.find();
+    drop(finder_lock);
 
     if params.debug {
+        pretty_print(&format!("Content: {:#?}", &params.content), Status::Info);
         pretty_print(&format!("Finder: {:#?}", &find_list), Status::Info);
-        pretty_print(
-            &format!("Finder Regex: {:#?}", &find_list.as_regex()),
-            Status::Info,
-        );
     }
 
-    // Iterate over all matches in the content
-    for cap in find_list.as_regex().captures_iter(&params.content) {
-        let matched = cap[0].to_string();
-        let parts: Vec<&str> = matched.split(':').collect();
+    let items: Vec<_> = find_list
+        .resources
+        .par_iter()
+        .filter_map(|r| {
+            // Improve performance by checking if it matches without using regular expressions
+            if !params.content.contains(&r.path) {
+                return None;
+            }
+
+            if params.debug {
+                pretty_print(&format!("Matched: {}", &r.path), Status::Info);
+            }
 
-        let path = parts.get(0).unwrap().to_string();
-        let line: Option<usize> = match parts.get(1) {
-            Some(line) => match line.parse::<usize>() {
+            let numbers = match r.as_regex().find(&params.content) {
+                Some(numbers) => numbers,
+                None => return None,
+            };
+            let line = match numbers.as_str().parse::<usize>() {
                 Ok(line) => Some(line),
                 Err(_) => None,
-            },
-            None => None,
-        };
-
-        let column: Option<usize> = match parts.get(2) {
-            Some(column) => match column.parse::<usize>() {
+            };
+            let column = match numbers.as_str().parse::<usize>() {
                 Ok(column) => Some(column),
                 Err(_) => None,
-            },
-            None => None,
-        };
-
-        let item_type = match path.starts_with('/') {
-            true => GrepItemType::AbsolutePath,
-            false => GrepItemType::RelativePath,
-        };
+            };
+            let item_type = match &r.path.starts_with('/') {
+                true => GrepItemType::AbsolutePath,
+                false => GrepItemType::RelativePath,
+            };
 
-        items.push(GrepItem {
-            path,
-            line,
-            column,
-            item_type,
-        });
-    }
+            Some(GrepItem {
+                path: r.path.clone(),
+                line,
+                column,
+                item_type,
+                filetype: r.filetype.clone(),
+            })
+        })
+        .collect();
 
+    let mut unique_items: Vec<GrepItem> = items.into_iter().collect();
     // dedup by item.path
-    items.sort_by(|a, b| {
+    unique_items.sort_by(|a, b| {
         a.path
             .partial_cmp(&b.path)
             .unwrap_or(std::cmp::Ordering::Equal)
     });
-    if params.unique {
-        items.dedup_by(|a, b| a.path == b.path);
-    }
+    unique_items.dedup_by(|a, b| a.path == b.path);
+
+    // filetypeでのフィルタリング
+    unique_items.retain(|item| match item.filetype {
+        Filetype::File => params.filetype.contains(&Filetype::File),
+        Filetype::Directory => params.filetype.contains(&Filetype::Directory),
+    });
 
-    items
+    unique_items
 }