From 110cf1e40c1943385cd7afe7ef2c95684b05f758 Mon Sep 17 00:00:00 2001
From: AndreaGuarracino <andreaguarracino@outlook.com>
Date: Wed, 27 Mar 2024 13:57:09 -0500
Subject: [PATCH 1/2] support gzipped PAF files

---
 Cargo.toml  |  1 +
 src/main.rs |  2 +-
 src/paf.rs  | 26 ++++++++++++++++++++------
 3 files changed, 22 insertions(+), 7 deletions(-)

diff --git a/Cargo.toml b/Cargo.toml
index 650fef4..266fd01 100644
--- a/Cargo.toml
+++ b/Cargo.toml
@@ -13,3 +13,4 @@ itertools = "0.10.0"
 fnv = "1.0.7"
 rgb = "0.8"
 line_drawing = "0.8.0"
+flate2 = "1.0.28"
\ No newline at end of file
diff --git a/src/main.rs b/src/main.rs
index 7b9f150..fbbf568 100644
--- a/src/main.rs
+++ b/src/main.rs
@@ -8,7 +8,7 @@ use crate::paf::{PafFile, paf_query_end, paf_target, paf_target_length, paf_targ
 
 fn main() {
     let matches = App::new("paf2chain")
-        .version("0.1.0")
+        .version("0.1.1")
         .author("Andrea Guarracino")
         .about("Generate a CHAIN format file from a PAF format file")
         .arg(
diff --git a/src/paf.rs b/src/paf.rs
index 965a95e..4abdcef 100644
--- a/src/paf.rs
+++ b/src/paf.rs
@@ -1,7 +1,9 @@
 use boomphf::Mphf;
 use itertools::Itertools;
 use std::fs::File;
-use std::io::{BufRead, BufReader};
+use std::io::{BufRead, BufReader, Read};
+use std::path::Path;
+use flate2::read::GzDecoder;
 
 #[derive(Debug, Clone)]
 struct AlignedSeq {
@@ -81,7 +83,19 @@ pub fn paf_target_end(line: &str) -> usize {
 
 fn _for_each_line_in_file(paf_filename: &str, mut callback: impl FnMut(&str)) {
     let file = File::open(paf_filename).unwrap();
-    let reader = BufReader::new(file);
+    // Determine if the file is gzipped based on its extension
+    let gzipped = Path::new(paf_filename)
+        .extension()
+        .map_or(false, |ext| ext == "gz");
+
+    // Create a dynamic reader based on the file type
+    let box_reader: Box<dyn Read> = if gzipped {
+        Box::new(GzDecoder::new(file))
+    } else {
+        Box::new(file)
+    };
+
+    let reader = BufReader::new(box_reader);
     for line in reader.lines() {
         callback(&line.unwrap());
     }
@@ -109,7 +123,7 @@ impl PafFile {
             let query_id = query_mphf.hash(&query_name) as usize;
             if !seen_queries[query_id] {
                 seen_queries[query_id] = true;
-                let mut query = &mut queries[query_id];
+                let query = &mut queries[query_id];
                 query.name = query_name;
                 query.length = paf_query_length(l);
             }
@@ -117,7 +131,7 @@ impl PafFile {
             let target_id = target_mphf.hash(&target_name) as usize;
             if !seen_targets[target_id] {
                 seen_targets[target_id] = true;
-                let mut target = &mut targets[target_id];
+                let target = &mut targets[target_id];
                 target.name = target_name;
                 target.length = paf_target_length(l);
             }
@@ -129,7 +143,7 @@ impl PafFile {
         let mut target_offset: usize = 0;
         targets_sort.iter().for_each(|t| {
             let target_id = target_mphf.hash(&t.name) as usize;
-            let mut target = &mut targets[target_id];
+            let target = &mut targets[target_id];
             target.rank = target_idx;
             target_idx += 1;
             target.offset = target_offset;
@@ -141,7 +155,7 @@ impl PafFile {
         let mut query_offset: usize = 0;
         queries_sort.iter().for_each(|q| {
             let query_id = query_mphf.hash(&q.name) as usize;
-            let mut query = &mut queries[query_id];
+            let query = &mut queries[query_id];
             query.rank = query_idx;
             query_idx += 1;
             query.offset = query_offset;

From d2e77fb7b05fe20061f3b3c3f789b4d761647d09 Mon Sep 17 00:00:00 2001
From: AndreaGuarracino <andreaguarracino@outlook.com>
Date: Wed, 27 Mar 2024 13:59:21 -0500
Subject: [PATCH 2/2] README

---
 Cargo.toml | 2 +-
 README.md  | 2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/Cargo.toml b/Cargo.toml
index 266fd01..4995736 100644
--- a/Cargo.toml
+++ b/Cargo.toml
@@ -13,4 +13,4 @@ itertools = "0.10.0"
 fnv = "1.0.7"
 rgb = "0.8"
 line_drawing = "0.8.0"
-flate2 = "1.0.28"
\ No newline at end of file
+flate2 = "1.0.28"
diff --git a/README.md b/README.md
index 10be568..287dd8b 100644
--- a/README.md
+++ b/README.md
@@ -18,7 +18,7 @@ cargo install --force --path .
 
 Generate alignments with the cigar string attached to the `cg:Z:` tag.
 These can be made by several aligners, including `minimap2 -c`, `wfmash`, or `lastz --format=paf:wfmash`.
-With alignments in `aln.paf`, we would convert it into a CHAIN format file using this call:
+With alignments in `aln.paf` (or `aln.paf.gz` if gzip-compressed), we would convert it into a CHAIN format file using this call:
 
 ```
 paf2chain -i aln.paf > aln.chain