gtfparse 2.8.0__tar.gz → 2.8.1__tar.gz

{gtfparse-2.8.0 → gtfparse-2.8.1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: gtfparse
-Version: 2.8.0
+Version: 2.8.1
 Summary: Parsing library for extracting data frames of genomic features from GTF files
 Author-email: Alex Rubinsteyn <alex.rubinsteyn@unc.edu>
 Project-URL: Homepage, https://github.com/openvax/gtfparse

{gtfparse-2.8.0 → gtfparse-2.8.1}/gtfparse/__init__.py

@@ -24,7 +24,7 @@ from .read_gtf import (
 )
 from .write_gtf import write_gtf
 
-__version__ = "2.8.0"
+__version__ = "2.8.1"
 
 __all__ = [
     "GENCODE_BIOTYPE_ALIASES",

{gtfparse-2.8.0 → gtfparse-2.8.1}/gtfparse/read_gtf.py

@@ -110,9 +110,10 @@ DEFAULT_COLUMN_DTYPES = {
 def parse_with_polars_lazy(
     filepath_or_buffer, split_attributes=True, features=None, fix_quotes_columns=["attribute"]
 ):
-    # use a global string cache so that all strings get intern'd into
-    # a single numbering system
-    polars.enable_string_cache()
+    # Categories shares categorical mappings on modern Polars. Older releases
+    # need the global string cache for categoricals from separate reads to match.
+    if not hasattr(polars, "Categories"):
+        polars.enable_string_cache()
     kwargs = {
         "has_header": False,
         "separator": "\t",

{gtfparse-2.8.0 → gtfparse-2.8.1}/gtfparse.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: gtfparse
-Version: 2.8.0
+Version: 2.8.1
 Summary: Parsing library for extracting data frames of genomic features from GTF files
 Author-email: Alex Rubinsteyn <alex.rubinsteyn@unc.edu>
 Project-URL: Homepage, https://github.com/openvax/gtfparse

{gtfparse-2.8.0 → gtfparse-2.8.1}/gtfparse.egg-info/SOURCES.txt

@@ -21,6 +21,7 @@ tests/test_expand_attributes.py
 tests/test_gencode_gtf.py
 tests/test_multiple_values_for_tag_attribute.py
 tests/test_parse_gtf_lines.py
+tests/test_polars_compatibility.py
 tests/test_read_stringtie_gtf.py
 tests/test_refseq_gtf.py
 tests/test_write_gtf.py

gtfparse-2.8.1/tests/test_polars_compatibility.py

@@ -0,0 +1,51 @@
+from io import StringIO
+
+import polars
+import pytest
+
+from gtfparse import read_gtf
+from gtfparse.read_gtf import parse_with_polars_lazy
+
+GTF_TEXT = (
+    '1\tensembl\tgene\t10\t20\t.\t+\t.\tgene_id "g1";\n'
+    '2\thavana\texon\t30\t40\t.\t-\t.\tgene_id "g2";\n'
+)
+
+pytestmark = pytest.mark.filterwarnings("error::DeprecationWarning")
+
+
+@pytest.mark.parametrize("result_type", ["pandas", "polars"])
+def test_repeated_reads_preserve_categoricals(result_type):
+    for _ in range(2):
+        df = read_gtf(StringIO(GTF_TEXT), result_type=result_type, features={"exon"})
+        assert list(df["seqname"]) == ["2"]
+        assert list(df["feature"]) == ["exon"]
+        assert list(df["gene_id"]) == ["g2"]
+        if result_type == "polars":
+            assert df["seqname"].dtype == polars.Categorical
+        else:
+            assert df["seqname"].dtype.name == "category"
+
+
+def test_categoricals_from_separate_lazy_reads_can_join():
+    left = parse_with_polars_lazy(StringIO(GTF_TEXT))
+    reversed_text = "\n".join(reversed(GTF_TEXT.splitlines())) + "\n"
+    right = parse_with_polars_lazy(StringIO(reversed_text))
+    keys = ["seqname", "source", "feature", "strand"]
+    joined = left.join(right, on=keys).sort("start").collect()
+
+    assert joined.height == 2
+    assert joined["start"].to_list() == [10, 30]
+    assert joined["start_right"].to_list() == [10, 30]
+    for key in keys:
+        assert joined[key].dtype == polars.Categorical
+
+
+@pytest.mark.skipif(not hasattr(polars, "Categories"), reason="Requires modern Polars")
+def test_modern_reads_do_not_toggle_global_string_cache(monkeypatch):
+    def fail_on_cache_toggle():
+        pytest.fail("Reading a GTF must not toggle the global string cache on modern Polars")
+
+    monkeypatch.setattr(polars, "enable_string_cache", fail_on_cache_toggle)
+    monkeypatch.setattr(polars, "disable_string_cache", fail_on_cache_toggle)
+    assert read_gtf(StringIO(GTF_TEXT)).height == 2