From 1a5403c0a825ff1241c1950eea4c346ab14186d6 Mon Sep 17 00:00:00 2001
From: bluenote-1577 <jimshawster@gmail.com>
Date: Thu, 30 May 2024 17:53:45 +0000
Subject: [PATCH] v0.5.5 with output format similar to strobealign

---
 CHANGELOG.md   |  4 ++++
 Cargo.lock     |  2 +-
 Cargo.toml     |  2 +-
 src/cmdline.rs |  3 +++
 src/contain.rs | 13 +++++++++----
 5 files changed, 18 insertions(+), 6 deletions(-)

diff --git a/CHANGELOG.md b/CHANGELOG.md
index 3a57936..db77e77 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -1,3 +1,7 @@
+## fairy v0.5.5
+
+* Added the --aemb_format option which is the same as strobealign's format (for single samples).
+
 ## fairy v0.5.4
 
 Updating to try and fix bioconda bug. 
diff --git a/Cargo.lock b/Cargo.lock
index 4d3ecaf..4db6b7a 100644
--- a/Cargo.lock
+++ b/Cargo.lock
@@ -330,7 +330,7 @@ dependencies = [
 
 [[package]]
 name = "fairy"
-version = "0.5.4"
+version = "0.5.5"
 dependencies = [
  "assert_cmd",
  "bincode",
diff --git a/Cargo.toml b/Cargo.toml
index 58c1040..42b9196 100644
--- a/Cargo.toml
+++ b/Cargo.toml
@@ -1,6 +1,6 @@
 [package]
 name = "fairy"
-version = "0.5.4"
+version = "0.5.5"
 edition = "2021"
 license = "MIT OR Apache-2.0"
 
diff --git a/src/cmdline.rs b/src/cmdline.rs
index 034e3d8..40969c8 100644
--- a/src/cmdline.rs
+++ b/src/cmdline.rs
@@ -117,4 +117,7 @@ pub struct ContainArgs {
     pub out_file_name: Option<String>,
     #[clap(long="maxbin-format", help = "Remove contig length, average depth, and variance columns. (default: MetaBAT2 format with variances)", help_heading="OUTPUT")]
     pub concoct_format: bool,
+    #[clap(long="aemb-format", help = "Strobealign --aemb format (default: MetaBAT2 format with variances)", help_heading="OUTPUT")]
+    pub aemb_format: bool,
+
 }
diff --git a/src/contain.rs b/src/contain.rs
index 1098ee5..7ea471b 100644
--- a/src/contain.rs
+++ b/src/contain.rs
@@ -35,23 +35,28 @@ fn print_cov_matrix(ani_results: Vec<AniResult>, read_files: &FxHashSet<String>,
 
     //sort(&mut contig_list_sorted);
     sort(&mut read_list_sorted);
+    let no_var_and_mean = args.concoct_format | args.aemb_format;
 
     if args.concoct_format{
         write!(writer, "contigName").unwrap();
         for read_name in read_list_sorted.iter(){
             write!(writer, "\t{}", read_name).unwrap();
         }
+        write!(writer, "\n").unwrap();
+    }
+    else if args.aemb_format{
     }
     else{
         write!(writer, "contigName\tcontigLen\ttotalAvgDepth").unwrap();
         for read_name in read_list_sorted.iter(){
             write!(writer, "\t{}\t{}-var", read_name, read_name).unwrap();
         }
+        write!(writer, "\n").unwrap();
     }
-    write!(writer, "\n").unwrap();
+
     for contig in contig_list_sorted{
         write!(writer, "{}", contig.split(' ').collect::<Vec<&str>>()[0]).unwrap();
-        if !args.concoct_format{
+        if !no_var_and_mean{
             write!(writer, "\t{}", contig_to_size[contig]).unwrap();
         }
         let mut avg_cov = 0.;
@@ -63,11 +68,11 @@ fn print_cov_matrix(ani_results: Vec<AniResult>, read_files: &FxHashSet<String>,
         else{
         }
         avg_cov /= read_list_sorted.len() as f64;
-        if !args.concoct_format{
+        if !no_var_and_mean{
             write!(writer, "\t{}", avg_cov).unwrap();
         }
         for read in read_list_sorted.iter(){
-            if !args.concoct_format{
+            if !no_var_and_mean{
                 if matrix.contains_key(contig) && matrix[contig].contains_key(read){
                     let (cov, var) = matrix[&contig][read];
                     write!(writer, "\t{}\t{}", cov, var).unwrap();