Merge pull request #6 from gofflab/claude/add-genome-reference-tool-t3h7n

Fix migration not registering species after pip upgrade

Author: Loyale

VERSIONINFO.md

@@ -1,3 +1,6 @@
+## v0.3.2 - 04.08.2026
+  + Fixed migration bug where species config entries were not restored after `pip install -U`
+  + Index directories found in the old package location are now auto-registered as species
 ## v0.3.1 - 04.08.2026
   + Moved config and indices to persistent user data directory (`~/.hcrprobedesign/`)
   + Reference genomes and configuration now survive `pip install -U` upgrades

setup.cfg

@@ -1,7 +1,7 @@
 [metadata]
 # replace with your username:
 name = hcrprobedesign
-version = 0.3.1
+version = 0.3.2
 author = Loyal A. Goff
 author_email = loyalgoff@jhmi.edu
 description = Probe Design tool for Hybridization Chain Reaction

src/HCRProbeDesign/_datadir.py

@@ -110,6 +110,7 @@ def _migrate_old_data():
     new_species = new_config.setdefault("species", {})
 
     migrated_any = False
+    migrated_index_dirs = []
 
     # Migrate index directories
     if os.path.isdir(old_indices_dir):
@@ -127,16 +128,17 @@ def _migrate_old_data():
                     f"  Skipping index '{entry}' (already exists in {new_indices_dir})",
                     file=sys.stderr,
                 )
-                continue
+            else:
+                print(
+                    f"  Moving index '{entry}' -> {new_entry_path}",
+                    file=sys.stderr,
+                )
+                shutil.copytree(old_entry_path, new_entry_path)
+                migrated_any = True
 
-            print(
-                f"  Moving index '{entry}' -> {new_entry_path}",
-                file=sys.stderr,
-            )
-            shutil.copytree(old_entry_path, new_entry_path)
-            migrated_any = True
+            migrated_index_dirs.append(entry)
 
-    # Migrate species config entries
+    # Migrate species config entries from old config
     for name, entry in old_species.items():
         if name in new_species:
             continue
@@ -157,6 +159,26 @@ def _migrate_old_data():
             file=sys.stderr,
         )
 
+    # Auto-register species for migrated index dirs that have no config entry.
+    # This handles the case where pip replaced the old HCRconfig.yaml with a
+    # fresh copy (species: {}) during upgrade, so the config entries were lost
+    # even though the index files survived.
+    for dirname in migrated_index_dirs:
+        # Check if any existing species entry already points to this index dir
+        index_prefix = os.path.join(new_indices_dir, dirname, dirname)
+        already_registered = any(
+            sp.get("bowtie2_index", "") == index_prefix
+            or os.path.basename(sp.get("bowtie2_index", "")).startswith(dirname)
+            for sp in new_species.values()
+        )
+        if dirname not in new_species and not already_registered:
+            new_species[dirname] = {"bowtie2_index": index_prefix}
+            migrated_any = True
+            print(
+                f"  Auto-registered species '{dirname}' -> {index_prefix}",
+                file=sys.stderr,
+            )
+
     # Preserve default_params from old config if not present in new
     if "default_params" not in new_config and "default_params" in old_config:
         new_config["default_params"] = old_config["default_params"]

tests/test_datadir.py

@@ -111,6 +111,47 @@ def test_migrate_old_data(tmp_path, monkeypatch):
     assert len(bt2_files) == 6
 
 
+def test_migrate_auto_registers_species_when_config_empty(tmp_path, monkeypatch):
+    """Index files survive pip upgrade but old config was replaced with species: {}.
+    Migration should auto-register species based on index directory names."""
+    data_dir = tmp_path / "data"
+    monkeypatch.setenv("HCRPROBEDESIGN_DATA_DIR", str(data_dir))
+
+    # Set up old package dir with index files but EMPTY species config
+    # (simulates pip replacing HCRconfig.yaml with the clean default)
+    old_pkg = tmp_path / "old_pkg"
+    old_indices = old_pkg / "indices" / "mm10"
+    old_indices.mkdir(parents=True)
+    for suffix in ["1.bt2", "2.bt2", "3.bt2", "4.bt2", "rev.1.bt2", "rev.2.bt2"]:
+        (old_indices / f"mm10.{suffix}").write_text("fake")
+
+    old_config_path = old_pkg / "HCRconfig.yaml"
+    old_config = {
+        "species": {},
+        "default_params": {"tileSize": 52},
+    }
+    with open(old_config_path, "w") as fh:
+        yaml.safe_dump(old_config, fh)
+
+    monkeypatch.setattr(_datadir, "_PACKAGE_DIRECTORY", str(old_pkg))
+    monkeypatch.setattr(_datadir, "_SEED_CONFIG", str(old_config_path))
+
+    _datadir.ensure_data_dir()
+
+    # Index files should be migrated
+    new_index_dir = data_dir / "indices" / "mm10"
+    assert new_index_dir.exists()
+    bt2_files = glob.glob(str(new_index_dir / "*.bt2*"))
+    assert len(bt2_files) == 6
+
+    # Species should be auto-registered from the directory name
+    new_config = yaml.safe_load((data_dir / "HCRconfig.yaml").read_text())
+    assert "mm10" in new_config["species"]
+    assert new_config["species"]["mm10"]["bowtie2_index"] == os.path.join(
+        str(data_dir), "indices", "mm10", "mm10"
+    )
+
+
 def test_has_old_data_empty(tmp_path, monkeypatch):
     monkeypatch.setattr(_datadir, "_PACKAGE_DIRECTORY", str(tmp_path))
     # Write an empty config