test: run db-download and plass_class depth tests in temp dirs

Stop the suite mutating committed fixtures:
- test_db: the download test installs into a TemporaryDirectory rather than
  removing/overwriting the committed Plassembler_Test_DB fixtures in place.
- test_plass_class: the get_depth / get_depth_long / combine_tsv tests (Plass
  and Assembly) copy their fixture dir into a TemporaryDirectory and work there,
  so committed intermediates (e.g. assembly_class/depth/combined_short.sam) are
  no longer regenerated over the tracked copies.

Validated with real tools: the 75MB DB download passes and the plass_class
tests pass, both leaving the committed fixtures untouched.

Author: gbouras13

tests/test_db.py

@@ -6,6 +6,7 @@
 """
 
 import sys
+import tempfile
 
 # import
 import unittest
@@ -42,12 +43,13 @@ def test_check_db_installation_good(self):
         check_db_installation(db_path, force=False, install_flag=False)
 
     # for plassembler download
-    # NOTE: force=True + install_flag=True downloads the real (~75MB) database
-    # over the network and overwrites the committed test fixtures in place, so
-    # this is neither offline-safe nor idempotent. Kept out of the fast subset.
+    # NOTE: downloads the real (~75MB) database over the network, so this is slow
+    # and offline-unsafe. It installs into a throwaway tmp dir so the committed
+    # fixture DB is never removed or overwritten.
     @pytest.mark.slow
     def test_check_db_installation_good_d(self):
-        check_db_installation(db_path, force=True, install_flag=True)
+        with tempfile.TemporaryDirectory() as tmp:
+            check_db_installation(Path(tmp) / "db", force=True, install_flag=True)
 
     def test_check_db_installation_bad(self):
         with self.assertRaises(SystemExit):

tests/test_plass_class.py

@@ -6,12 +6,13 @@
 """
 
 # import
+import shutil
+import tempfile
 import unittest
 from pathlib import Path
 
 import pytest
 
-from src.plassembler.utils.cleanup import remove_file
 from src.plassembler.utils.plass_class import Assembly, Plass
 
 # import functions
@@ -65,11 +66,12 @@ def test_check_get_depth(self):
         plass = Plass()
         pacbio_model = "nothing"
         threads = 1
-        # set to the depth dir for  intermediate files
-        plass.outdir = plass_class_depth_dir
-        plass.get_depth(logdir, pacbio_model, threads)
-        remove_file(Path(f"{plass_class_depth_dir}/combined_short.sam"))
-        remove_file(Path(f"{plass_class_depth_dir}/combined_sorted_short.bam"))
+        with tempfile.TemporaryDirectory() as tmp:
+            # work in a copy so the committed fixtures are not mutated
+            workdir = Path(tmp) / "depth"
+            shutil.copytree(plass_class_depth_dir, workdir)
+            plass.outdir = workdir
+            plass.get_depth(logdir, pacbio_model, threads)
         self.assertEqual(expected, True)
 
     @pytest.mark.slow
@@ -78,14 +80,12 @@ def test_check_get_depth_long(self):
         plass = Plass()
         pacbio_model = "nothing"
         threads = 1
-        plasmids_for_sketching = Path(
-            f"{plass_class_depth_dir}/plasmids_for_sketching.fasta"
-        )
-        # set to the depth dir for  intermediate files
-        plass.outdir = plass_class_depth_dir
-        plass.get_depth_long(logdir, pacbio_model, threads, plasmids_for_sketching)
-        remove_file(Path(f"{plass_class_depth_dir}/combined_long.sam"))
-        remove_file(Path(f"{plass_class_depth_dir}/combined_sorted_long.bam"))
+        with tempfile.TemporaryDirectory() as tmp:
+            workdir = Path(tmp) / "depth"
+            shutil.copytree(plass_class_depth_dir, workdir)
+            plasmids_for_sketching = workdir / "plasmids_for_sketching.fasta"
+            plass.outdir = workdir
+            plass.get_depth_long(logdir, pacbio_model, threads, plasmids_for_sketching)
         self.assertEqual(expected, True)
 
     def test_process_mash_tsv(self):
@@ -100,18 +100,16 @@ def test_process_mash_tsv(self):
     def test_combine_tsv(self):
         expected = True
         plass = Plass()
-        plass.outdir = plass_class_depth_dir
         prefix = "plassembler"
         pacbio_model = "nothing"
         threads = 1
-        plass.get_depth(logdir, pacbio_model, threads)
-        plass.process_mash_tsv(plassembler_db_dir)
-        plass.combine_depth_mash_tsvs(prefix, depth_filter=0.1, skip_mash=False)
-        remove_file(Path(f"{plass_class_depth_dir}/combined_long.sam"))
-        remove_file(Path(f"{plass_class_depth_dir}/combined_short.sam"))
-        remove_file(Path(f"{plass_class_depth_dir}/combined_sorted_long.bam"))
-        remove_file(Path(f"{plass_class_depth_dir}/combined_sorted_short.bam"))
-        remove_file(Path(f"{plass_class_depth_dir}/{prefix}_summary.tsv"))
+        with tempfile.TemporaryDirectory() as tmp:
+            workdir = Path(tmp) / "depth"
+            shutil.copytree(plass_class_depth_dir, workdir)
+            plass.outdir = workdir
+            plass.get_depth(logdir, pacbio_model, threads)
+            plass.process_mash_tsv(plassembler_db_dir)
+            plass.combine_depth_mash_tsvs(prefix, depth_filter=0.1, skip_mash=False)
         self.assertEqual(expected, True)
 
 
@@ -133,13 +131,11 @@ def test_check_get_depth(self):
         assembly = Assembly()
         pacbio_model = "nothing"
         threads = 1
-        # set to the depth dir for  intermediate files
-        assembly.outdir = assembly_depth_dir
-        assembly.get_depth(logdir, pacbio_model, threads)
-        remove_file(Path(f"{assembly_depth_dir}/combined_long.sam"))
-        remove_file(Path(f"{assembly_depth_dir}/combined_short.sam"))
-        remove_file(Path(f"{assembly_depth_dir}/combined_sorted_long.bam"))
-        remove_file(Path(f"{assembly_depth_dir}/combined_sorted_short.bam"))
+        with tempfile.TemporaryDirectory() as tmp:
+            workdir = Path(tmp) / "depth"
+            shutil.copytree(assembly_depth_dir, workdir)
+            assembly.outdir = workdir
+            assembly.get_depth(logdir, pacbio_model, threads)
         self.assertEqual(expected, True)
 
     def test_process_mash_tsv(self):
@@ -155,17 +151,15 @@ def test_process_mash_tsv(self):
     def test_combine_tsv(self):
         expected = True
         assembly = Assembly()
-        assembly.outdir = assembly_depth_dir
         prefix = "plassembler"
         pacbio_model = "nothing"
         threads = 1
-        plasmid_fasta = Path(f"{assembly_depth_dir}/plasmids.fasta")
-        assembly.get_depth(logdir, pacbio_model, threads)
-        assembly.process_mash_tsv(plassembler_db_dir, plasmid_fasta)
-        assembly.combine_depth_mash_tsvs(prefix, False)
-        remove_file(Path(f"{assembly_depth_dir}/combined_long.sam"))
-        remove_file(Path(f"{assembly_depth_dir}/combined_short.bam"))
-        remove_file(Path(f"{assembly_depth_dir}/combined_sorted_long.bam"))
-        remove_file(Path(f"{assembly_depth_dir}/combined_sorted_short.bam"))
-        remove_file(Path(f"{assembly_depth_dir}/{prefix}_summary.tsv"))
+        with tempfile.TemporaryDirectory() as tmp:
+            workdir = Path(tmp) / "depth"
+            shutil.copytree(assembly_depth_dir, workdir)
+            assembly.outdir = workdir
+            plasmid_fasta = workdir / "plasmids.fasta"
+            assembly.get_depth(logdir, pacbio_model, threads)
+            assembly.process_mash_tsv(plassembler_db_dir, plasmid_fasta)
+            assembly.combine_depth_mash_tsvs(prefix, False)
         self.assertEqual(expected, True)