methplotlib 0.23.2__tar.gz → 0.23.3__tar.gz

{methplotlib-0.23.2/methplotlib.egg-info → methplotlib-0.23.3}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: methplotlib
-Version: 0.23.2
+Version: 0.23.3
 Summary: Plot nucleotide modification data from nanopore sequencing
 Home-page: https://github.com/wdecoster/methplotlib
 Author: Wouter De Coster

{methplotlib-0.23.2 → methplotlib-0.23.3}/methplotlib/import_methylation.py

@@ -91,6 +91,7 @@ def parse_nanopolish(filename, file_type, name, window, smoothen=5):
                 raise
             header = gzip.open(filename, "rt").readline().rstrip().split("\t")
             table = pd.read_csv(tabix_stream.stdout, sep="\t", header=None, names=header)
+            check_tabix(tabix_stream, filename, window)
         else:
             logging.info(f"Reading {filename} slowly by splitting the file in chunks.")
             sys.stderr.write(f"\nReading {filename} would be faster with bgzip and tabix.\n")
@@ -235,6 +236,19 @@ def no_records_warning(filename, window):
     sys.stderr.write(f"\nWARNING: {message}\nThis dataset is skipped.\n")
 
 
+def check_tabix(tabix_stream, filename, window):
+    """
+    Exit if tabix failed, as a failing tabix also writes nothing to stdout,
+    which would otherwise be reported as a window without records.
+    Only call this after its stdout has been read.
+    """
+    if tabix_stream.wait() != 0:
+        sys.exit(
+            f"\n\nERROR: tabix failed to read {window.chromosome}:{window.begin}-{window.end} "
+            f"from {filename}:\n{tabix_stream.stderr.read().decode()}"
+        )
+
+
 def read_in_chunks(filename, window, sep, usecols, index_suggestion):
     """
     Read a headerless file in chunks, keeping only the lines of the chromosome of interest.
@@ -303,6 +317,7 @@ def parse_bedgraph(filename, name, window):
                 names=colnames,
                 dtype={0: str},
             )
+            check_tabix(tabix_stream, filename, window)
         else:
             table = read_in_chunks(
                 filename,
@@ -380,13 +395,21 @@ def parse_bedmethyl(filename, name, window, smoothen=5, flavor="modkit", mods_of
                 sys.stderr.write("\n\nERROR when opening a tabix stream.\n")
                 sys.stderr.write("Is tabix installed and on the PATH?.")
                 raise
-            table = pd.read_csv(
-                tabix_stream.stdout,
-                sep=sep,
-                header=None,
-                usecols=usecols,
-                dtype={0: str},
-            ).rename(columns=colnames)
+            try:
+                table = pd.read_csv(
+                    tabix_stream.stdout,
+                    sep=sep,
+                    header=None,
+                    usecols=usecols,
+                    dtype={0: str},
+                ).rename(columns=colnames)
+            except pd.errors.EmptyDataError:
+                table = None
+            check_tabix(tabix_stream, filename, window)
+            # tabix writes nothing if the window has no records or the chromosome is not in the index
+            if table is None:
+                no_records_warning(filename, window)
+                return []
         else:
             table = read_in_chunks(
                 filename,

methplotlib-0.23.3/methplotlib/version.py

@@ -0,0 +1 @@
+__version__ = "0.23.3"

{methplotlib-0.23.2 → methplotlib-0.23.3/methplotlib.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: methplotlib
-Version: 0.23.2
+Version: 0.23.3
 Summary: Plot nucleotide modification data from nanopore sequencing
 Home-page: https://github.com/wdecoster/methplotlib
 Author: Wouter De Coster

{methplotlib-0.23.2 → methplotlib-0.23.3}/methplotlib.egg-info/SOURCES.txt

@@ -21,4 +21,5 @@ methplotlib/differential/__init__.py
 methplotlib/differential/differential.py
 scripts/allele_specific_modification
 scripts/differential_modification
-tests/test_differential.py
+tests/test_differential.py
+tests/test_import_methylation.py

methplotlib-0.23.3/tests/test_import_methylation.py

@@ -0,0 +1,78 @@
+import shutil
+
+import pysam
+import pytest
+
+from methplotlib.import_methylation import parse_bedgraph, parse_bedmethyl, parse_nanopolish
+from methplotlib.utils import Region
+
+pytestmark = pytest.mark.skipif(shutil.which("tabix") is None, reason="tabix not on the PATH")
+
+# `modkit pileup --only-tabs` records, as shared in issue #52
+BEDMETHYL = """\
+1\t10468\t10469\th\t1\t+\t10468\t10469\t255,0,0\t1\t0.00\t0\t0\t1\t1\t0\t0\t0
+1\t10468\t10469\tm\t1\t+\t10468\t10469\t255,0,0\t1\t100.00\t1\t0\t0\t1\t0\t0\t0
+1\t10469\t10470\th\t2\t-\t10469\t10470\t255,0,0\t2\t0.00\t0\t1\t1\t0\t0\t0\t0
+1\t10469\t10470\tm\t2\t-\t10469\t10470\t255,0,0\t2\t50.00\t1\t1\t0\t0\t0\t0\t0
+"""
+
+BEDGRAPH = """\
+1\t10468\t10469\t0.5
+1\t10469\t10470\t1.0
+"""
+
+NANOPOLISH_FREQ = """\
+chromosome\tstart\tend\tnum_motifs_in_group\tcalled_sites\tcalled_sites_methylated\tmethylated_frequency\tgroup_sequence
+1\t10468\t10468\t1\t10\t5\t0.5\tACGTA
+1\t10470\t10470\t1\t10\t10\t1.0\tACGTA
+"""
+
+
+def indexed(tmp_path, name, content, **index_args):
+    path = tmp_path / name
+    path.write_text(content)
+    # bgzips the file and creates the .tbi index next to it
+    return pysam.tabix_index(str(path), **index_args)
+
+
+@pytest.fixture
+def indexed_bedmethyl(tmp_path):
+    return indexed(tmp_path, "test.bed", BEDMETHYL, preset="bed")
+
+
+def corrupt_index(filename):
+    with open(filename + ".tbi", "w") as index:
+        index.write("not an index")
+
+
+def test_indexed_bedmethyl_with_records(indexed_bedmethyl):
+    mods = parse_bedmethyl(indexed_bedmethyl, "test", Region("1:10000-11000"))
+    assert sorted(m.name for m in mods) == ["test_h", "test_m"]
+
+
+@pytest.mark.parametrize("window", ["chrX:146,991,437-147,034,645", "1:500000-600000"])
+def test_indexed_bedmethyl_without_records(indexed_bedmethyl, window):
+    assert parse_bedmethyl(indexed_bedmethyl, "test", Region(window)) == []
+
+
+@pytest.mark.parametrize(
+    "name, content, index_args, parse",
+    [
+        ("test.bed", BEDMETHYL, {"preset": "bed"}, parse_bedmethyl),
+        ("test.bedgraph", BEDGRAPH, {"preset": "bed"}, parse_bedgraph),
+        (
+            "freq.tsv",
+            NANOPOLISH_FREQ,
+            {"seq_col": 0, "start_col": 1, "end_col": 2, "line_skip": 1},
+            lambda f, n, w: parse_nanopolish(f, "nanopolish_freq", n, w),
+        ),
+    ],
+    ids=["bedmethyl", "bedgraph", "nanopolish_freq"],
+)
+def test_tabix_failure(tmp_path, name, content, index_args, parse):
+    # a corrupt index makes tabix exit with an error and no output,
+    # which should not be reported as a window without records
+    filename = indexed(tmp_path, name, content, **index_args)
+    corrupt_index(filename)
+    with pytest.raises(SystemExit, match="tabix failed"):
+        parse(filename, "test", Region("1:10000-11000"))

methplotlib-0.23.2/methplotlib/version.py

@@ -1 +0,0 @@
-__version__ = "0.23.2"