refactor: use Python 3 stdlib instead of six

misc/docker/seqrepo.df

@@ -37,7 +37,6 @@ RUN pip3 install --upgrade \
     pygments \
     pysam \
     simplegeneric \
-    six \
     tqdm \
     traitlets \
     wcwidth \

pyproject.toml

@@ -22,7 +22,6 @@ dependencies = [
     "ipython ~= 8.4",
     "pysam ~= 0.22",
     "requests ~= 2.31",
-    "six ~= 1.16",
     "tqdm ~= 4.66",
     "yoyo-migrations ~= 8.2",
 ]

src/biocommons/seqrepo/fastadir/bases.py

@@ -1,11 +1,8 @@
 import abc
 from typing import Optional
 
-import six
 
-
-@six.add_metaclass(abc.ABCMeta)
-class BaseReader:
+class BaseReader(metaclass=abc.ABCMeta):
     @abc.abstractmethod
     def fetch(self, seq_id: str, start: Optional[int] = None, end: Optional[int] = None) -> str:
         raise NotImplementedError
@@ -14,8 +11,7 @@ def __getitem__(self, ac: str) -> str:
         return self.fetch(ac)
 
 
-@six.add_metaclass(abc.ABCMeta)
-class BaseWriter:
+class BaseWriter(metaclass=abc.ABCMeta):
     @abc.abstractmethod
     def store(self, seq_id: str, seq: str) -> str:
         pass  # pragma: no cover

src/biocommons/seqrepo/fastadir/fabgz.py

@@ -16,7 +16,6 @@
 from types import TracebackType
 from typing import Optional, Type
 
-import six
 from pysam import FastaFile
 from typing_extensions import Self
 
@@ -45,15 +44,14 @@ def _get_bgzip_version(exe: str) -> str:
 
 def _find_bgzip() -> str:
     """return path to bgzip if found and meets version requirements, else exception"""
-    missing_file_exception = OSError if six.PY2 else FileNotFoundError
     min_bgzip_version = ".".join(map(str, min_bgzip_version_info))
     exe = os.environ.get("SEQREPO_BGZIP_PATH", shutil.which("bgzip") or "/usr/bin/bgzip")
 
     try:
         bgzip_version = _get_bgzip_version(exe)
     except AttributeError:
         raise RuntimeError("Didn't find version string in bgzip executable ({exe})".format(exe=exe))
-    except missing_file_exception:
+    except FileNotFoundError:
         raise RuntimeError(
             "{exe} doesn't exist; you need to install htslib and tabix "
             "(See https://github.com/biocommons/biocommons.seqrepo#requirements)".format(exe=exe)

tests/test_fabgz.py

@@ -3,7 +3,6 @@
 import tempfile
 
 import pytest
-import six
 
 from biocommons.seqrepo.fastadir.fabgz import FabgzReader, FabgzWriter
 
@@ -19,18 +18,18 @@ def test_write_reread():
 
     # write sequences
     faw = FabgzWriter(fabgz_fn)
-    for seq_id, seq in six.iteritems(sequences):
+    for seq_id, seq in sequences.items():
         faw.store(seq_id, seq)
     # add twice to demonstrate non-redundancy
-    for seq_id, seq in six.iteritems(sequences):
+    for seq_id, seq in sequences.items():
         faw.store(seq_id, seq)
     faw.close()
 
     # now read them back
     far = FabgzReader(fabgz_fn)
-    assert far.filename.startswith(tmpdir.encode())
+    assert far.filename.startswith(tmpdir.encode())  # type: ignore
     assert set(far.keys()) == set(sequences.keys())
-    assert 5 == len(far), "expected 5 sequences"
+    assert 5 == len(far), "expected 5 sequences"  # type: ignore
     assert "l10" in far.keys()
     assert far["l10"] == seed * 10
     for seq_id in far.keys():

tests/test_fastaiter.py

@@ -1,6 +1,6 @@
+from io import StringIO
+
 import pytest
-import six
-from six.moves import StringIO
 
 from biocommons.seqrepo.fastaiter import FastaIter
 
@@ -12,7 +12,7 @@ def test_empty():
 
     # should return an empty generator
     with pytest.raises(StopIteration):
-        six.next(iterator)
+        next(iterator)
 
 
 def test_noheader():
@@ -22,54 +22,54 @@ def test_noheader():
 
     # should return an empty generator
     with pytest.raises(StopIteration):
-        six.next(iterator)
+        next(iterator)
 
 
 def test_single():
     data = StringIO(">seq1\nACGT\n")
 
     iterator = FastaIter(data)
 
-    header, seq = six.next(iterator)
+    header, seq = next(iterator)
     assert header == "seq1"
     assert seq == "ACGT"
 
     # should be empty now
     with pytest.raises(StopIteration):
-        six.next(iterator)
+        next(iterator)
 
 
 def test_multiple():
     data = StringIO(">seq1\nACGT\n>seq2\nTGCA\n\n>seq3\nTTTT")
 
     iterator = FastaIter(data)
 
-    header, seq = six.next(iterator)
+    header, seq = next(iterator)
     assert header == "seq1"
     assert seq == "ACGT"
 
-    header, seq = six.next(iterator)
+    header, seq = next(iterator)
     assert header == "seq2"
     assert seq == "TGCA"
 
-    header, seq = six.next(iterator)
+    header, seq = next(iterator)
     assert header == "seq3"
     assert seq == "TTTT"
 
     # should be empty now
     with pytest.raises(StopIteration):
-        six.next(iterator)
+        next(iterator)
 
 
 def test_multiline():
     data = StringIO(">seq1\nACGT\nTGCA")
 
     iterator = FastaIter(data)
 
-    header, seq = six.next(iterator)
+    header, seq = next(iterator)
     assert header == "seq1"
     assert seq == "ACGTTGCA"
 
     # should be empty now
     with pytest.raises(StopIteration):
-        six.next(iterator)
+        next(iterator)