Perform referenceless contig stitching in sample.py

cfe-lab · Jan 2, 2025 · 72fcbfc · 72fcbfc
1 parent 1db40d9
commit 72fcbfc
Show file tree

Hide file tree

Showing 2 changed files with 12 additions and 5 deletions.
diff --git a/micall/drivers/sample.py b/micall/drivers/sample.py
@@ -21,6 +21,7 @@
 from micall.utils.driver_utils import makedirs
 from micall.utils.fasta_to_csv import fasta_to_csv
 from micall.utils.referencefull_contig_stitcher import referencefull_contig_stitcher
+from micall.utils.referenceless_contig_stitcher import referenceless_contig_stitcher
 from contextlib import contextmanager
 
 logger = logging.getLogger(__name__)
@@ -426,10 +427,14 @@ def run_denovo(self, excluded_seeds):
                    merged_contigs_csv,
                    )
 
+        with open(self.unstitched_contigs_fasta, 'r') as unstitched_contigs_fasta, \
+             open(self.stitched_contigs_fasta, 'w') as stitched_contigs_fasta:
+            referenceless_contig_stitcher(unstitched_contigs_fasta, stitched_contigs_fasta)
+
         with open(self.unstitched_contigs_csv, 'w') as unstitched_contigs_csv, \
              open(self.merged_contigs_csv, 'r') as merged_contigs_csv, \
              open(self.blast_csv, 'w') as blast_csv:
-            fasta_to_csv(Path(self.unstitched_contigs_fasta),
+            fasta_to_csv(Path(self.stitched_contigs_fasta),
                          unstitched_contigs_csv,
                          merged_contigs_csv,
                          blast_csv=blast_csv,

diff --git a/micall/utils/referenceless_contig_stitcher.py b/micall/utils/referenceless_contig_stitcher.py
@@ -3,6 +3,7 @@
 from fractions import Fraction
 from Bio import SeqIO, Seq
 from Bio.SeqRecord import SeqRecord
+from micall.utils.contig_stitcher_context import StitcherContext
 
 from micall.utils.contig_stitcher_contigs import Contig
 from micall.utils.find_maximum_overlap import find_maximum_overlap
@@ -144,10 +145,11 @@ def stitch_consensus(contigs: Iterable[Contig]) -> Iterable[Contig]:
 
 
 def write_contigs(output_fasta: TextIO, contigs: Iterable[Contig]):
-    records = (SeqRecord(Seq.Seq(contig.seq),
-                         name=contig.unique_name)
-               for contig in contigs)
-    SeqIO.write(records, output_fasta, "fasta")
+    with StitcherContext.fresh():
+        records = (SeqRecord(Seq.Seq(contig.seq),
+                             name=contig.unique_name)
+                   for contig in contigs)
+        SeqIO.write(records, output_fasta, "fasta")
 
 
 def read_contigs(input_fasta: TextIO) -> Iterable[Contig]: