support .zstd, .xz and .bz2 compressed inputs

Cargo.lock

@@ -117,6 +117,16 @@ dependencies = [
  "libc",
 ]
 
+[[package]]
+name = "bzip2-rs"
+version = "0.1.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "beeb59e7e4c811ab37cc73680c798c7a5da77fc9989c62b09138e31ee740f735"
+dependencies = [
+ "crc32fast",
+ "tinyvec",
+]
+
 [[package]]
 name = "bzip2-sys"
 version = "0.1.11+1.0.8"
@@ -648,6 +658,7 @@ version = "0.1.5"
 dependencies = [
  "ahash",
  "base64 0.22.1",
+ "bzip2-rs",
  "clap",
  "derive_more",
  "flate2",
@@ -659,6 +670,7 @@ dependencies = [
  "ratatui",
  "rayon",
  "regex",
+ "ruzstd",
  "serde",
  "serde_json",
  "tar",
@@ -666,6 +678,7 @@ dependencies = [
  "tikv-jemallocator",
  "time",
  "tinystr",
+ "xz2",
  "zip",
 ]
 
@@ -696,6 +709,17 @@ dependencies = [
  "crc",
 ]
 
+[[package]]
+name = "lzma-sys"
+version = "0.1.20"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "5fda04ab3764e6cde78b9974eec4f779acaba7c4e84b36eca3cf77c581b85d27"
+dependencies = [
+ "cc",
+ "libc",
+ "pkg-config",
+]
+
 [[package]]
 name = "main_error"
 version = "0.1.2"
@@ -1003,6 +1027,15 @@ version = "1.0.18"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "0e819f2bc632f285be6d7cd36e25940d45b2391dd6d9b939e79de557f7014248"
 
+[[package]]
+name = "ruzstd"
+version = "0.7.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "99c3938e133aac070997ddc684d4b393777d293ba170f2988c8fd5ea2ad4ce21"
+dependencies = [
+ "twox-hash",
+]
+
 [[package]]
 name = "ryu"
 version = "1.0.18"
@@ -1269,6 +1302,22 @@ dependencies = [
  "serde",
 ]
 
+[[package]]
+name = "tinyvec"
+version = "1.8.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "445e881f4f6d382d5f27c034e25eb92edd7c784ceab92a0937db7f2e9471b938"
+
+[[package]]
+name = "twox-hash"
+version = "1.6.3"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "97fee6b57c6a41524a810daee9286c02d7752c4253064d0b05472833a438f675"
+dependencies = [
+ "cfg-if",
+ "static_assertions",
+]
+
 [[package]]
 name = "typenum"
 version = "1.17.0"
@@ -1452,6 +1501,15 @@ dependencies = [
  "rustix",
 ]
 
+[[package]]
+name = "xz2"
+version = "0.1.7"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "388c44dc09d76f1536602ead6d325eb532f5c122f17782bd57fb47baeeb767e2"
+dependencies = [
+ "lzma-sys",
+]
+
 [[package]]
 name = "zerocopy"
 version = "0.7.35"

Cargo.toml

@@ -14,8 +14,6 @@ regex = "1.10.5"
 log = "0.4.22"
 clap = { version = "=4.1.3", features = ["derive"] }
 logsmash-data = { version = "0.1.0", path = "./data" }
-zip = "2.1.5"
-flate2 = "1.0.31"
 itertools = "0.13.0"
 ratatui = "0.29.0"
 tinystr = { version = "0.7.6", features = ["serde"] }
@@ -26,6 +24,11 @@ base64 = "0.22.1"
 derive_more = { version = "1.0.0-beta.6", features = ["from"] }
 rayon = "1.10.0"
 tar = "0.4.42"
+zip = "2.1.5"
+flate2 = "1.0.31"
+xz2 = "0.1.7"
+bzip2-rs = "0.1.2"
+ruzstd = "0.7.2"
 
 [target.'cfg(not(target_os = "windows"))'.dependencies]
 tikv-jemallocator = "0.6.0"

src/error.rs

@@ -1,3 +1,4 @@
+use ruzstd::frame_decoder::FrameDecoderError;
 use std::string::FromUtf8Error;
 use thiserror::Error;
 use zip::result::ZipError;
@@ -22,6 +23,8 @@ pub enum ReadError {
     Io(#[from] std::io::Error),
     #[error(transparent)]
     Zip(#[from] ZipError),
+    #[error(transparent)]
+    Zstd(#[from] FrameDecoderError),
     #[error("archive contains multiple files")]
     MultipleFiles,
     #[error("archive contains no files")]

src/logfile/mod.rs

@@ -2,9 +2,12 @@ mod archive;
 
 use crate::error::ReadError;
 use crate::logfile::archive::{Archive, ArchiveEntry, TarArchive, ZipArchive};
+use bzip2_rs::DecoderReader;
 use flate2::read::GzDecoder;
+use ruzstd::StreamingDecoder;
 use std::fs::File;
-use std::io::Read;
+use std::io::{BufReader, Read};
+use xz2::read::XzDecoder;
 
 pub struct LogFile {
     content: String,
@@ -13,6 +16,7 @@ pub struct LogFile {
 impl LogFile {
     pub fn open(path: &str) -> Result<LogFile, ReadError> {
         let file = File::open(path)?;
+        let file = BufReader::new(file);
         if path.ends_with(".zip") {
             let mut zip = ZipArchive::new(file)?;
             let content = select_file(&mut zip)?;
@@ -23,6 +27,15 @@ impl LogFile {
         if let Some(path) = path.strip_suffix(".gz") {
             let decoder = GzDecoder::new(file);
             return Self::open_no_seek(path, decoder);
+        } else if let Some(path) = path.strip_suffix(".xz") {
+            let decoder = XzDecoder::new(file);
+            return Self::open_no_seek(path, decoder);
+        } else if let Some(path) = path.strip_suffix(".bz2") {
+            let decoder = DecoderReader::new(file);
+            return Self::open_no_seek(path, decoder);
+        } else if let Some(path) = path.strip_suffix(".zst") {
+            let decoder = StreamingDecoder::new(file)?;
+            return Self::open_no_seek(path, decoder);
         }
 
         Self::open_no_seek(path, Box::new(file))