conda env pytest contig table prep fix

Author: dsmutin

.github/workflows/build.yml

@@ -71,6 +71,7 @@ jobs:
         flake8 . --count --exit-zero --max-complexity=10 --max-line-length=127 --statistics
     - name: Test package installation and imports
       run: |
+        # Full test suite (including BLAST/primer3 integration) runs in conda-build only.
         # Verify package can be imported
         python -c "import PROBESt; print('PROBESt imported successfully')"
         python -c "from PROBESt.args import arguments_parse; print('Args module imported successfully')"

app/templates/index.html

@@ -14,7 +14,7 @@
                 <div class="logo-title">
                     <img src="{{ url_for('static', filename='ctlab_probest_white.png') }}" alt="PROBESt" style="width: 30vw;">
                     <div>
-                        <h1>PROBESt v0.2.0</h1>
+                        <h1>PROBESt v0.2.2</h1>
                         <p class="subtitle">ITMO probe generation and optimization tool</p>
                     </div>
                 </div>

extraction/environment.yml

@@ -129,7 +129,7 @@ dependencies:
   - typing_extensions=4.15.0=pyhcf101f3_0
   - typing_utils=0.1.0=pyhd8ed1ab_1
   - uri-template=1.3.0=pyhd8ed1ab_1
-  - wcwidth=0.2.14=pyhd8ed1ab_0
+  - wcwidth=0.2.24=pyhd8ed1ab_0
   - webcolors=24.11.1=pyhd8ed1ab_0
   - webencodings=0.5.1=pyhd8ed1ab_3
   - websocket-client=1.8.0=pyhd8ed1ab_1
@@ -220,7 +220,7 @@ dependencies:
       - opencv-python-headless==4.11.0.86
       - openpyxl==3.1.5
       - outlines==1.2.5
-      - outlines-core==0.2.11
+      - outlines-core==0.2.21
       - pandas==2.3.3
       - pdfminer-six==20250327
       - pdfplumber==0.11.6

src/PROBESt/prepare_blast.py

@@ -42,6 +42,11 @@
 FASTA_EXTENSIONS = {'.fa', '.fasta', '.fna', '.fa.gz', '.fasta.gz', '.fna.gz'}
 
 
+def _logical_fasta_key(basename: str) -> str:
+    """Plain and gzip of the same file share a key (e.g. a.fna and a.fna.gz -> a.fna)."""
+    return basename[:-3] if basename.endswith(".gz") else basename
+
+
 def deduplicate_contig_table(contig_table_path: str, contig_id_col: int = 1) -> None:
     """
     Deduplicate a contig table in-place by contig ID (second column by default).
@@ -112,17 +117,35 @@ def get_fasta_files(directory: str) -> List[str]:
     """
     Get all FASTA files from a directory.
     
+    If both ``assembly.fna`` and ``assembly.fna.gz`` exist, keep only the ``.gz``
+    path so directories may ship compressed assets (CI) while developers keep
+    uncompressed copies locally.
+    
     Args:
         directory: Path to the directory containing FASTA files.
         
     Returns:
         List of paths to FASTA files.
     """
-    fasta_files = []
+    seen: List[str] = []
     for ext in FASTA_EXTENSIONS:
         pattern = os.path.join(directory, f"*{ext}")
-        fasta_files.extend(glob.glob(pattern))
-    return sorted(fasta_files)
+        seen.extend(glob.glob(pattern))
+
+    by_key: Dict[str, str] = {}
+    for path in sorted(set(seen)):
+        key = _logical_fasta_key(os.path.basename(path))
+        prev = by_key.get(key)
+        if prev is None:
+            by_key[key] = path
+        elif path.endswith(".gz") and not prev.endswith(".gz"):
+            by_key[key] = path
+        elif prev.endswith(".gz") and not path.endswith(".gz"):
+            pass
+        else:
+            by_key[key] = min(path, prev)
+
+    return sorted(by_key.values())
 
 
 def prepare_blast_database(

tests/PROBESt/test_prepare_blast.py

@@ -31,6 +31,7 @@
 import sys
 import tempfile
 import shutil
+import gzip
 from pathlib import Path
 
 # Add the project root directory to the Python path
@@ -182,6 +183,20 @@ def test_includes_all_supported_extensions(self, tmp_path):
         fasta_files = get_fasta_files(str(ext_dir))
         assert len(fasta_files) == 3
 
+    def test_prefers_gzip_when_plain_and_gzip_same_stem(self, tmp_path):
+        """Same assembly as .fna and .fna.gz should count once; prefer .gz."""
+        d = tmp_path / "dup"
+        d.mkdir()
+        plain = d / "GCA_asm.fna"
+        plain.write_text(">ctg1\nATGC\n")
+        gzpath = d / "GCA_asm.fna.gz"
+        with gzip.open(gzpath, "wb") as gz:
+            gz.write(b">ctg1\nATGC\n")
+
+        paths = get_fasta_files(str(d))
+        assert len(paths) == 1
+        assert paths[0].endswith(".fna.gz")
+
 
 class TestPrepareBlastDatabase:
     """Tests for prepare_blast_database function."""

tests/test_pipeline_general_integration.py

@@ -43,7 +43,11 @@ def _require_blast_tools():
 
 @pytest.mark.integration
 def test_integration_fish(tmp_path):
-    """Same CLI as: …/test.fna, fasta_base true/false dirs, -N 3, --visualize True --AI True."""
+    """Same CLI as: …/test.fna, fasta_base true/false dirs, -N 3, --visualize True --AI True.
+
+    True-base genomes must be present in git as ``*.fna.gz`` under ``fasta_base/true_base``
+    (plain ``*.fna`` is gitignored); false bases already use gz in the repo.
+    """
     _require_blast_tools()
 
     output = tmp_path / "output"