From 788f168d5fa24b1896315132f03f7a9cc3b76ef2 Mon Sep 17 00:00:00 2001
From: JeanMainguy <jean.mainguy@outlook.fr>
Date: Wed, 18 Dec 2024 19:09:21 +0100
Subject: [PATCH] pyfastx index stored in temporary file dir

---
 binette/contig_manager.py |  4 ++--
 binette/main.py           | 14 ++++++++++++--
 2 files changed, 14 insertions(+), 4 deletions(-)

diff --git a/binette/contig_manager.py b/binette/contig_manager.py
index fc783f5..91cdf6e 100644
--- a/binette/contig_manager.py
+++ b/binette/contig_manager.py
@@ -2,7 +2,7 @@
 from typing import Dict, Iterable, Tuple, Set, Any, Union
 
 
-def parse_fasta_file(fasta_file: str) -> pyfastx.Fasta:
+def parse_fasta_file(fasta_file: str, index_file: str) -> pyfastx.Fasta:
     """
     Parse a FASTA file and return a pyfastx.Fasta object.
 
@@ -10,7 +10,7 @@ def parse_fasta_file(fasta_file: str) -> pyfastx.Fasta:
 
     :return: A pyfastx.Fasta object representing the parsed FASTA file.
     """
-    fa = pyfastx.Fasta(fasta_file, build_index=True)
+    fa = pyfastx.Fasta(fasta_file, build_index=True, index_file=index_file)
     return fa
 
 
diff --git a/binette/main.py b/binette/main.py
index 5f95cb4..8533f4e 100755
--- a/binette/main.py
+++ b/binette/main.py
@@ -213,6 +213,7 @@ def parse_input_files(
     bin_dirs: List[Path],
     contig2bin_tables: List[Path],
     contigs_fasta: Path,
+    temporary_dir: Path,
     fasta_extensions: Set[str] = {".fasta", ".fna", ".fa"},
 ) -> Tuple[
     Dict[str, Set[bin_manager.Bin]], Set[bin_manager.Bin], Set[str], Dict[str, int]
@@ -254,7 +255,10 @@ def parse_input_files(
     original_bins = bin_manager.dereplicate_bin_sets(bin_set_name_to_bins.values())
 
     logging.info(f"Parsing contig fasta file: {contigs_fasta}")
-    contigs_object = contig_manager.parse_fasta_file(contigs_fasta.as_posix())
+    index_file = temporary_dir / f"{contigs_fasta.name}.fxi"
+    contigs_object = contig_manager.parse_fasta_file(
+        contigs_fasta.as_posix(), index_file=index_file.as_posix()
+    )
 
     unexpected_contigs = {
         contig for contig in contigs_in_bins if contig not in contigs_object
@@ -276,6 +280,7 @@ def parse_input_files(
 def manage_protein_alignement(
     faa_file: Path,
     contigs_fasta: Path,
+    temporary_dir: Path,
     contig_to_length: Dict[str, int],
     contigs_in_bins: Set[str],
     diamond_result_file: Path,
@@ -314,9 +319,12 @@ def manage_protein_alignement(
         )
 
     else:
+        index_file = temporary_dir / f"{contigs_fasta.name}.fxi"
         contigs_iterator = (
             seq
-            for seq in contig_manager.parse_fasta_file(contigs_fasta.as_posix())
+            for seq in contig_manager.parse_fasta_file(
+                contigs_fasta.as_posix(), index_file=index_file.as_posix()
+            )
             if seq.name in contigs_in_bins
         )
         contig_to_genes = cds.predict(contigs_iterator, faa_file.as_posix(), threads)
@@ -502,6 +510,7 @@ def main():
         args.contig2bin_tables,
         args.contigs,
         fasta_extensions=set(args.fasta_extensions),
+        temporary_dir=out_tmp_dir,
     )
 
     if args.proteins and not args.resume:
@@ -517,6 +526,7 @@ def main():
     contig_to_kegg_counter, contig_to_genes = manage_protein_alignement(
         faa_file=faa_file,
         contigs_fasta=args.contigs,
+        temporary_dir=out_tmp_dir,
         contig_to_length=contig_to_length,
         contigs_in_bins=contigs_in_bins,
         diamond_result_file=diamond_result_file,