From 7872ac35045f063caffb95b7f4a84184fc4ab2cd Mon Sep 17 00:00:00 2001
From: wdecoster <decosterwouter@gmail.com>
Date: Mon, 4 Sep 2023 14:01:02 +0200
Subject: [PATCH] fix memory problem from faidx

temporarily, see https://github.com/rust-bio/rust-htslib/issues/401
---
 Cargo.toml     |  3 ++-
 src/main.rs    |  2 +-
 src/repeats.rs | 21 ++++++++++++---------
 3 files changed, 15 insertions(+), 11 deletions(-)

diff --git a/Cargo.toml b/Cargo.toml
index 6e42cf7..027ad97 100644
--- a/Cargo.toml
+++ b/Cargo.toml
@@ -1,6 +1,6 @@
 [package]
 name = "STRdust"
-version = "0.2.0"
+version = "0.2.1"
 edition = "2021"
 
 
@@ -26,6 +26,7 @@ distance = "0.4.0"
 levenshtein = "1.0.5"
 rand = "0.8.5"
 libz-sys = "1.1.12"
+libc = "0.2.147"
 
 [dev-dependencies]
 ctor = "*"
diff --git a/src/main.rs b/src/main.rs
index 0bdc891..d973406 100644
--- a/src/main.rs
+++ b/src/main.rs
@@ -64,7 +64,7 @@ pub struct Cli {
 
 fn is_file(pathname: &str) -> Result<(), String> {
     let path = PathBuf::from(pathname);
-    if path.is_file() {
+    if path.is_file() || pathname.starts_with("http") {
         Ok(())
     } else {
         Err(format!("Input file {} is invalid", path.display()))
diff --git a/src/repeats.rs b/src/repeats.rs
index 6a5ce8e..ed52cc8 100644
--- a/src/repeats.rs
+++ b/src/repeats.rs
@@ -1,6 +1,6 @@
 use log::error;
 use rust_htslib::faidx;
-use std::{fmt, io::Write};
+use std::fmt;
 
 pub struct RepeatInterval {
     pub chrom: String,
@@ -45,16 +45,18 @@ impl RepeatInterval {
                 (self.end + flanking - 2) as usize,
             )
             .expect("Failed to extract fas_right sequence from fasta for {chrom}:{start}-{end}");
-        // write the new reference sequence to a file
-        let mut newref_file =
-            std::fs::File::create("newref.fa").expect("Unable to create newref.fa");
-        newref_file
-            .write_all(&[fas_left, fas_right].concat())
-            .expect("Unable to write to newref.fa");
-        [fas_left, fas_right].concat()
+        // // write the new reference sequence to a file
+        // let mut newref_file =
+        //     std::fs::File::create("newref.fa").expect("Unable to create newref.fa");
+        // newref_file
+        //     .write_all(&[fas_left, fas_right].concat())
+        //     .expect("Unable to write to newref.fa");
+        let newref = [fas_left, fas_right].concat();
+        unsafe { libc::free(fas_left.as_ptr() as *mut std::ffi::c_void) }; // Free up memory
+        unsafe { libc::free(fas_right.as_ptr() as *mut std::ffi::c_void) }; // Free up memory
+        newref
     }
 
-    /// If the repeat sequence is out of bounds, None is returned
     pub fn reference_repeat_sequence(&self, fasta: &String) -> Option<String> {
         let fas = faidx::Reader::from_path(fasta).expect("Failed to read fasta");
         let repeat_ref_sequence = std::str::from_utf8(
@@ -63,6 +65,7 @@ impl RepeatInterval {
         )
         .expect("Failed to convert repeat sequence to string for {chrom}:{start}-{end}")
         .to_string();
+        // If the repeat sequence is out of bounds, None is returned
         if repeat_ref_sequence == "N" {
             eprintln!(
                 "Cannot genotype repeat at {self} because it is out of bounds for the fasta file",