parthjadhav / rust_search

Blazingly fast file search library built in Rust

Blazingly fast file search crate built in Rust 🔥

📦 Usage

Please report any problems you encounter when using rust search here: Issues

Add rust_search = "2.0.0" in Cargo.toml.

[dependencies]
rust_search = "2.0.0"

Examples

• General use

use rust_search::SearchBuilder;

fn main(){
    let search: Vec<String> = SearchBuilder::default()
        .location("~/path/to/directory")
        .search_input("what to search")
        .more_locations(vec!["/anotherPath/to/search", "/keepAddingIfYouWant/"])
        .limit(1000) // results to return
        .ext("extension")
        .strict()
        .depth(1)
        .ignore_case()
        .hidden()
        .build()
        .collect();

    for path in search {
        println!("{}", path);
    }
}

• Sort the output by similarity with the input

 use rust_search::{SearchBuilder, similarity_sort};
 fn main() {
     let search_input = "fly";
     let mut search: Vec<String> = SearchBuilder::default()
         .location("~/Desktop/")
         .search_input(search_input)
         .depth(1)
         .ignore_case()
         .build()
        .collect();

     similarity_sort(&mut search, &search_input);
     for path in search {
        println!("{:?}", path);
     }
 }
 

search without similarity sort ["afly.txt", "bfly.txt", "flyer.txt", "fly.txt"]

search with similarity sort ["fly.txt", "flyer.txt", "afly.txt", "bfly.txt",]

• To get all the files with a specific extension in a directory, use:

use rust_search::SearchBuilder;

let files: Vec<String> = SearchBuilder::default()
    .location("/path/to/directory")
    .ext("file_extension")
    .build()
    .collect();

• To get all the files in a directory, use:

use rust_search::SearchBuilder;

let files: Vec<String> = SearchBuilder::default()
    .location("/path/to/directory")
    .depth(1)
    .build()
    .collect();

To filter files by date_created, date_modified, file_size and/or custom_filter, use:

use rust_search::{FileSize, FilterExt, SearchBuilder};
use std::time::{Duration, SystemTime};

let search: Vec<String> = SearchBuilder::default()
		.location("~/path/to/directory")
		.file_size_greater(FileSize::Kilobyte(200.0))
		.file_size_smaller(FileSize::Megabyte(10.0))
		.created_after(SystemTime::now() - Duration::from_secs(3600 * 24 * 10))
		.created_before(SystemTime::now())
		.modified_after(SystemTime::now() - Duration::from_secs(3600 * 24 * 5))
		.custom_filter(|dir| dir.metadata().unwrap().is_file())
		.custom_filter(|dir| !dir.metadata().unwrap().permissions().readonly())
		.build()
		.collect();

👉 For more examples, please refer to the Documentation

⚙️ Benchmarks

The difference in sample size is due to the fact that fd and glob are different tools and have different use cases. fd is a command line tool that searches for files and directories. glob is a library that can be used to search for files and directories. The benchmark is done on a MacBook Air M2, 16 GB Unified memory.

Benchmarks are done using hyperfine, Benchmarks files are available in the benchmarks drive folder.

- Rust Search vs Glob

The benchmark was done on a directories containing 300K files.

Command / Library	Mean [s]	Min [s]	Max [s]	Relative
rust_search	1.317 ± 0.002	1.314	1.320	1.00
glob	22.728 ± 0.023	22.690	22.746	17.25 ± 0.03

---

- Rust Search vs FD

The benchmark was done on a directories containing 45K files.

Command / Library	Mean [ms]	Min [ms]	Max [ms]	Relative
rust_search	680.5 ± 2.1	678.3	683.6	1.00
fd -e .js	738.7 ± 10.2	720.8	746.7	1.09 ± 0.02

---

Results:-

+ rust_search is 17.25 times faster than Glob.

+ rust_search** is 1.09 times faster than FD.

👨‍💻 Contributors

Any contributions would be greatly valued as this library is still in its early stages.

• Documentation
• Benchmarks
• Implementation guidelines
• Code Improvement

If you want to contribute to this project, please follow the steps below:

1. Fork the project
2. Clone the forked repository
3. Create a feature branch
4. Make changes to the code
5. Commit the changes
6. Push the changes to the forked repository
7. Create a pull request
8. Wait for the pull request to be reviewed and merged (if approved)

License

This project is licensed under the terms of the MIT license.

Discord server & Linkedin

Click the button below to join the discord server or Linkedin

rust_search's Issues

Low speed on x86_64-unknown-linux-musl

Why so different speed for some targets?
For x86_64-unknown-linux-gnu

./target/release/walk    
Time elapsed: 1.022172369s
Found 11 files

and x7 slow for x86_64-unknown-linux-musl

./target/x86_64-unknown-linux-musl/release/walk
Time elapsed: 7.560947213s
Found 11 files

Why?

Add it to Rust Awesome

Of course because it's awesome. https://github.com/rust-unofficial/awesome-rust

[request] Add example of returning results

I'm a total newb to Rust. Trying to figure out how to use this library in the context of https://github.com/tauri-apps/tauri but am having trouble getting the return value right. Trying to have it return the file paths in a simple array (to be used in JS).

Based on the example code:

use rust_search::SearchBuilder;

fn main(){
    let search: Vec<String> = SearchBuilder::default()
        .location("~/path/to/directory")
        .search_input("what to search")
        .more_locations(vec!["/anotherPath/to/search", "/keepAddingIfYouWant/"])
        .limit(1000) // results to return
        .ext("extension")
        .strict()
        .depth(1)
        .ignore_case()
        .hidden()
        .build()
        .collect();

    for path in search {
        println!("{}", path);
    }

   // How to return `search` here as a simple array?
}

I've tried simply adding:

return search;

But that returns "expected (), found struct Vec".

I know that for Rust developers this is likely embarrassingly simple. Please forgive my ignorance! I've done quite a bit of searching and tested different things but nothing has worked.

Possibility to set a limit for upcoming files

^ Title

I will work on it

[Bug]: Search input includes `location` to search for

let search: Vec<String> = SearchBuilder::default()
        .location("/System/")
        .search_input("Syste")
        .depth(1)
        .ignore_case()
        .build()
        .collect();

This will return everything contained in /System/ directory because the Regex will be matched with the path input i.e /System/ and return everything.

This could be solved by adding: .split("/").last().unwrap() to path

reg_exp.is_match(&path.split("/").last().unwrap())

@TheAwiteb , Is there a better way to do this? Or are there any implications using this approach ?

Replace `~` with home directory in unix-like system

^ Title

I can claim it, will work in it after merge #9

[Feature]: Sort the paths by file names

Currently the results are scattered & random.

Eg:

Current output:
Search input: "a"
Search output: ["man", "cat", "ant", "ban", "tan"]

Expected result with .sort(true) :
Search input: "a"
Search output: ["ant", "ban", "cat", "man", "tan"]

Is it possible to exclude directories?

This would help performance when using a high depth value and knowing that desired results are not in directories {...}. For example, node_modules (which often have deeply nested structures with lots of files).

[Documentation]: Add benchmarks

Benchmarks requests:

1. Glob Vs Rust_Search
2. FD vs Rust_Search
   ...
   More suggestions are welcome 😄

How to ?

Using hyperfine you can easily create benchmarks

Receiving the file extension without a dot

This will make it easier to solve #4 , and also to understand

Return &Path or OsString object from iter

I am very impressed with the simplicity and speed of your solution!

Are you planning to abandon the use of String as a return value in the future? On Linux, this can create some problems, as far as I can tell.

[Feature Request]: Add search filters (Date Created, Date Modified, File Size)

Hi ,Support for increasing file size, establishment time and modification time.

[Feature Request]: Implement Caching

Question:

Should we? I'm not sure about this... Suggestions are open

Is there any example using this library in c++?

This is a cool library. I want to using this library in my c++ project, but I cannot find any example. I know I can use rust cxx to embed rust library to c++, but that's not easy to return string vector from rust to c++.
So hope any cxx/c++ exmaple using this library. Have you ever tired cxx? It will be awsome to have an example of that.

Ability to use outside variables in custom_filter() for better filtering

I want to use variables outside the scope of the closure inside custom_function(|d: DirEntry| { //code })
I am new to rust so please forgive me if I make a mistake.

Currently, this yeilds a type error because the closure is capturing outside variables.

Even using move or inner functions does not solve this problem.
This functionality would be particularly useful when the value of outside_var can change or is not determined from the start.

[Feature Request]: Make `Search` struct easier

Breaking changes

I think add search instance function to Search struct will make it easier. This will make it easier to do something like this

use rust_search::Search;

fn main() {
    let mains = Search::default().input("main").search();
    let readmes = Search::default().input("readme")
    .ignore_case() // will ignore ascii case
    .extension("md")
    .search();
}

[Bug]: Search work without file extension, but with it's not

^ Title

For example i have file it's name is some_random_word_yes_it_is.md in /home/User/foo/bar/

use rust_search::Search;

fn main(){
    let search = Search::new(
        "/home/User/foo/bar/",
        Some("yes"),
        None,
        None);

    for path in search {
        println!("{}", path);
    }
}

This will work, but when i add the file extinction it's not, for example

use rust_search::Search;

fn main(){
    let search = Search::new(
        "/home/User/foo/bar/",
        Some("yes"),
        Some(".md"), // also "md"
        None);

    for path in search {
        println!("{}", path);
    }
}

rust_search::Filesize cannot be imported

error[E0432]: unresolved import rust_search::FileType
--> src/main.rs:3:5
|
3 | use rust_search::FileType;
| ^^^^^^^^^^^^^^^^^^^^^ no FileType in the root
|
help: consider importing this struct instead
|
3 | use std::fs::FileType;
| ~~~~~~~~~~~~~~~~~~

error[E0433]: failed to resolve: use of undeclared type FileSize