Merge pull request #393 from Ensembl/mbarba/ruff_check

Add ruff checks

pyproject.toml

@@ -173,6 +173,17 @@ disable = [
 max-attributes = 12
 max-branches = 15
 
+[tool.ruff]
+line-length = 110
+show-fixes = true
+
+[tool.ruff.format]
+docstring-code-format = true
+
+[tool.ruff.lint.per-file-ignores]
+# Ignore `F403` (unable to detect undefined names) in all `__init__.py` files
+"__init__.py" = ["F403"]
+
 [tool.mypy]
 mypy_path = "src/python"
 explicit_package_bases = true

src/python/ensembl/brc4/runnable/seqregion_parser.py

@@ -16,7 +16,6 @@
 
 from pathlib import Path
 import re
-import gzip
 import csv
 from typing import Any, Dict, Tuple
 

src/python/ensembl/io/genomio/assembly/download.py

@@ -194,7 +194,7 @@ def _download_file(
     """
     has_md5 = True
     expected_sum = ""
-    if not ftp_file in md5_sums:
+    if ftp_file not in md5_sums:
         logging.warning(f" File not in the md5 checksums: {ftp_file}")
         has_md5 = False
     else:

src/python/ensembl/io/genomio/genome_metadata/prepare.py

@@ -92,7 +92,7 @@ def add_provider(genome_metadata: Dict, ncbi_data: Dict) -> None:
 
     # Add assembly provider (if missing)
     assembly = genome_metadata["assembly"]
-    if (not "provider_name" in assembly) and (not "provider_url" in assembly):
+    if ("provider_name" not in assembly) and ("provider_url" not in assembly):
         assembly["provider_name"] = provider["assembly"]["provider_name"]
         assembly["provider_url"] = f'{provider["assembly"]["provider_url"]}/{accession}'
 
@@ -111,7 +111,7 @@ def add_assembly_version(genome_data: Dict) -> None:
         genome_data: Genome information of assembly, accession and annotation.
     """
     assembly = genome_data["assembly"]
-    if not "version" in assembly:
+    if "version" not in assembly:
         accession = assembly["accession"]
         version = accession.partition(".")[2]
         if version:
@@ -128,9 +128,9 @@ def add_genebuild_metadata(genome_data: Dict) -> None:
     """
     genebuild = genome_data.setdefault("genebuild", {})
     current_date = datetime.date.today().isoformat()
-    if not "version" in genebuild:
+    if "version" not in genebuild:
         genebuild["version"] = current_date
-    if not "start_date" in genebuild:
+    if "start_date" not in genebuild:
         genebuild["start_date"] = current_date
 
 

src/python/ensembl/io/genomio/genome_stats/dump.py

@@ -104,7 +104,7 @@ def get_feature_stats(self, table: Any) -> Dict[str, int]:
         totals_st = select(func.count(table.stable_id))
         (total,) = session.execute(totals_st).one()
         # pylint: disable-next=singleton-comparison
-        no_desc_st = select(func.count(table.stable_id)).filter(table.description == None)
+        no_desc_st = select(func.count(table.stable_id)).filter(table.description.is_(None))
         (no_desc,) = session.execute(no_desc_st).one()
         xref_desc_st = select(func.count(table.stable_id)).where(table.description.like("%[Source:%"))
         (xref_desc,) = session.execute(xref_desc_st).one()