uproot 5.3.2__tar.gz → 5.3.3__tar.gz

{uproot-5.3.2 → uproot-5.3.3}/.github/workflows/build-test.yml

@@ -92,3 +92,27 @@ jobs:
       - name: Run pytest
         run: |
           python -m pytest -vv tests --reruns 10 --reruns-delay 30 --only-rerun "(?i)http|ssl|timeout|expired|connection|socket"
+
+  numpy2-build:
+    strategy:
+      fail-fast: false
+      matrix:
+        platform: [windows-latest, ubuntu-latest, macos-latest]
+        python-version: ['3.11']
+
+    runs-on: ${{ matrix.platform }}
+    timeout-minutes: 30
+
+    steps:
+      - uses: actions/checkout@v4
+
+      - uses: actions/setup-python@v5
+        with:
+          python-version: ${{ matrix.python-version }}
+
+      - name: Pip install the package
+        run: python -m pip install 'numpy>=2.0.0b1' .[test]
+
+      - name: Run pytest
+        run: |
+          python -m pytest -vv tests --reruns 10 --reruns-delay 30 --only-rerun "(?i)http|ssl|timeout|expired|connection|socket"

{uproot-5.3.2 → uproot-5.3.3}/.pre-commit-config.yaml

@@ -4,7 +4,7 @@ ci:
 
 repos:
   - repo: https://github.com/pre-commit/pre-commit-hooks
-    rev: v4.5.0
+    rev: v4.6.0
     hooks:
       - id: check-added-large-files
       - id: check-case-conflict
@@ -23,7 +23,7 @@ repos:
       - id: black
 
   - repo: https://github.com/astral-sh/ruff-pre-commit
-    rev: v0.3.4
+    rev: v0.3.5
     hooks:
       - id: ruff
         args: [--fix, --show-fixes]
@@ -36,7 +36,7 @@ repos:
         args: [--py38-plus]
 
   - repo: https://github.com/macisamuele/language-formatters-pre-commit-hooks
-    rev: v2.12.0
+    rev: v2.13.0
     hooks:
       - id: pretty-format-toml
         args: [--autofix]

{uproot-5.3.2 → uproot-5.3.3}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: uproot
-Version: 5.3.2
+Version: 5.3.3
 Summary: ROOT I/O in pure Python and NumPy.
 Project-URL: Download, https://github.com/scikit-hep/uproot5/releases
 Project-URL: Homepage, https://github.com/scikit-hep/uproot5

{uproot-5.3.2 → uproot-5.3.3}/src/uproot/_dask.py

@@ -606,6 +606,7 @@ def _get_dask_array(
                 filter_typename=filter_typename,
                 filter_branch=real_filter_branch,
                 full_paths=full_paths,
+                ignore_duplicates=True,
             )
 
             if common_keys is None:
@@ -747,6 +748,7 @@ def _get_dask_array_delay_open(
         filter_typename=filter_typename,
         filter_branch=filter_branch,
         full_paths=full_paths,
+        ignore_duplicates=True,
     )
 
     dask_dict = {}
@@ -1441,6 +1443,7 @@ def _get_dak_array(
                 filter_typename=filter_typename,
                 filter_branch=real_filter_branch,
                 full_paths=full_paths,
+                ignore_duplicates=True,
             )
 
             if common_keys is None:
@@ -1586,7 +1589,7 @@ def _get_dak_array_delay_open(
     ffile_path, fobject_path = files[0][0:2]
 
     if known_base_form is not None:
-        common_keys = list(known_base_form.fields)
+        common_keys = list(dict.fromkeys(known_base_form.fields))
         base_form = known_base_form
     else:
         obj = uproot._util.regularize_object_path(
@@ -1598,6 +1601,7 @@ def _get_dak_array_delay_open(
             filter_typename=filter_typename,
             filter_branch=filter_branch,
             full_paths=full_paths,
+            ignore_duplicates=True,
         )
         base_form = _get_ttree_form(
             awkward, obj, common_keys, interp_options.get("ak_add_doc")

{uproot-5.3.2 → uproot-5.3.3}/src/uproot/_util.py

@@ -80,7 +80,9 @@ def ensure_numpy(array, types=(numpy.bool_, numpy.integer, numpy.floating)):
 
     awkward = uproot.extras.awkward()
     with warnings.catch_warnings():
-        warnings.simplefilter("error", numpy.VisibleDeprecationWarning)
+        warnings.simplefilter(
+            "error", getattr(numpy, "exceptions", numpy).VisibleDeprecationWarning
+        )
         if isinstance(array, awkward.contents.Content):
             out = awkward.to_numpy(array)
         else:

{uproot-5.3.2 → uproot-5.3.3}/src/uproot/behaviors/TBranch.py

@@ -1126,6 +1126,7 @@ class HasBranches(Mapping):
         filter_branch=no_filter,
         recursive=True,
         full_paths=True,
+        ignore_duplicates=False,
     ):
         """
         Args:
@@ -1143,6 +1144,7 @@ class HasBranches(Mapping):
             full_paths (bool): If True, include the full path to each subbranch
                 with slashes (``/``); otherwise, use the descendant's name as
                 the output name.
+            ignore_duplicates (bool): If True, return a set of the keys; otherwise, return the full list of keys.
 
         Returns the names of the subbranches as a list of strings.
         """
@@ -1153,6 +1155,7 @@ class HasBranches(Mapping):
                 filter_branch=filter_branch,
                 recursive=recursive,
                 full_paths=full_paths,
+                ignore_duplicates=ignore_duplicates,
             )
         )
 
@@ -1279,6 +1282,7 @@ class HasBranches(Mapping):
         filter_branch=no_filter,
         recursive=True,
         full_paths=True,
+        ignore_duplicates=False,
     ):
         """
         Args:
@@ -1296,6 +1300,8 @@ class HasBranches(Mapping):
             full_paths (bool): If True, include the full path to each subbranch
                 with slashes (``/``); otherwise, use the descendant's name as
                 the output name.
+            ignore_duplicates (bool): If True, return a set of the keys; otherwise, return the full list of keys.
+
 
         Returns the names of the subbranches as an iterator over strings.
         """
@@ -1305,6 +1311,7 @@ class HasBranches(Mapping):
             filter_branch=filter_branch,
             recursive=recursive,
             full_paths=full_paths,
+            ignore_duplicates=ignore_duplicates,
         ):
             yield k
 
@@ -1353,6 +1360,7 @@ class HasBranches(Mapping):
         filter_branch=no_filter,
         recursive=True,
         full_paths=True,
+        ignore_duplicates=False,
     ):
         """
         Args:
@@ -1370,6 +1378,8 @@ class HasBranches(Mapping):
             full_paths (bool): If True, include the full path to each subbranch
                 with slashes (``/``) in the name; otherwise, use the descendant's
                 name as the name without modification.
+            ignore_duplicates (bool): If True, return a set of the keys; otherwise, return the full list of keys.
+
 
         Returns (name, branch) pairs of the subbranches as an iterator over
         2-tuples of (str, :doc:`uproot.behaviors.TBranch.TBranch`).
@@ -1385,6 +1395,8 @@ class HasBranches(Mapping):
                 f"filter_branch must be None or a function: TBranch -> bool, not {filter_branch!r}"
             )
 
+        keys_set = set()
+
         for branch in self.branches:
             if (
                 (
@@ -1394,7 +1406,11 @@ class HasBranches(Mapping):
                 and (filter_typename is no_filter or filter_typename(branch.typename))
                 and (filter_branch is no_filter or filter_branch(branch))
             ):
-                yield branch.name, branch
+                if ignore_duplicates and branch.name in keys_set:
+                    pass
+                else:
+                    keys_set.add(branch.name)
+                    yield branch.name, branch
 
             if recursive:
                 for k1, v in branch.iteritems(
@@ -1408,7 +1424,11 @@ class HasBranches(Mapping):
                     if filter_name is no_filter or _filter_name_deep(
                         filter_name, self, v
                     ):
-                        yield k2, v
+                        if ignore_duplicates and branch.name in keys_set:
+                            pass
+                        else:
+                            keys_set.add(k2)
+                            yield k2, v
 
     def itertypenames(
         self,

{uproot-5.3.2 → uproot-5.3.3}/src/uproot/containers.py

@@ -864,7 +864,9 @@ class AsVectorLike(AsContainer):
 
     def __init__(self, header, values):
         self.header = header
-        if isinstance(values, AsContainer):
+        if isinstance(self, uproot.containers.AsBitSet):
+            self._items = numpy.dtype(numpy.bool_)
+        elif isinstance(values, AsContainer):
             self._items = values
         elif isinstance(values, type) and issubclass(
             values, (uproot.model.Model, uproot.model.DispatchByVersion)
@@ -1117,7 +1119,6 @@ class AsList(AsVectorLike):
         header (bool): Sets the :ref:`uproot.containers.AsContainer.header`.
         values (:doc:`uproot.model.Model` or :doc:`uproot.containers.Container`): Data
             type for data nested in the container.
-
     A :doc:`uproot.containers.AsContainer` for ``std::list``.
     """
 
@@ -1139,6 +1140,36 @@ class AsList(AsVectorLike):
         return STLList
 
 
+class AsBitSet(AsVectorLike):
+    """
+    Args:
+        header (bool): Sets the :ref:`uproot.containers.AsContainer.header`.
+        keys (:doc:`uproot.model.Model` or :doc:`uproot.containers.Container`): Data
+            type for data nested in the container.
+
+    A :doc:`uproot.containers.AsContainer` for ``std::bitset``.
+    """
+
+    _specialpathitem_name = "bitset"
+
+    @property
+    def typename(self):
+        return f"std::bitset<{_content_typename(self.keys)}>"
+
+    @property
+    def keys(self):
+        """
+        Data type for data nested in the container.
+        """
+        return self._items
+
+    _form_parameters = {"__array__": "bitset"}
+
+    @property
+    def _container_type(self):
+        return STLBitSet
+
+
 class AsSet(AsVectorLike):
     """
     Args:
@@ -1633,6 +1664,49 @@ class STLVector(Container, Sequence):
         ]
 
 
+class STLBitSet(Container, Sequence):
+    """
+    Args:
+        values (``numpy.ndarray`` or iterable): Contents of the ``std::vector``.
+
+    Representation of a C++ ``std::bitset`` as a Python ``Sequence``.
+    """
+
+    def __init__(self, numbytes):
+        self._numbytes = numpy.asarray(numbytes)
+
+    def __str__(self, limit=85):
+        def tostring(i):
+            return _tostring(self._values[i])
+
+        return _str_with_ellipsis(tostring, len(self), "[", "]", limit)
+
+    def __repr__(self, limit=85):
+        return f"<STLBitSet {self.__str__(limit=limit - 30)} at 0x{id(self):012x}>"
+
+    def __getitem__(self, where):
+        return self._numbytes[where]
+
+    def __len__(self):
+        return self._numbytes
+
+    def __iter__(self):
+        return iter(self._numbytes)
+
+    def __eq__(self, other):
+        if isinstance(other, STLBitSet):
+            return self._numbytes == other._numbytes
+        elif isinstance(other, Sequence):
+            return self._numbytes == other
+        else:
+            return False
+
+    def tolist(self):
+        return [
+            x.tolist() if isinstance(x, (Container, numpy.ndarray)) else x for x in self
+        ]
+
+
 class STLSet(Container, Set):
     """
     Args:

{uproot-5.3.2 → uproot-5.3.3}/src/uproot/interpretation/identify.py

@@ -964,7 +964,7 @@ def _parse_node(tokens, i, typename, file, quote, header, inner_header):
         return (
             i + 4,
             _parse_maybe_quote(
-                f'uproot.containers.AsFIXME("std::bitset<{num_bits}>")',
+                f"uproot.containers.AsBitSet({header}, {num_bits})",
                 quote,
             ),
         )

{uproot-5.3.2 → uproot-5.3.3}/src/uproot/interpretation/numerical.py

@@ -106,7 +106,7 @@ class Numerical(uproot.interpretation.Interpretation):
 
                 start = stop
 
-        output = output.newbyteorder("=")
+        output = output.view(output.dtype.newbyteorder("="))
         self.hook_before_library_finalize(
             basket_arrays=basket_arrays,
             entry_start=entry_start,

{uproot-5.3.2 → uproot-5.3.3}/src/uproot/source/cursor.py

@@ -386,7 +386,7 @@ class Cursor:
             length_data = chunk.get(start, stop, self, context)
             length = numpy.frombuffer(length_data, dtype=self._u1).view(self._i4)[0]
         start = stop
-        stop = start + length
+        stop = start + int(length)
         if move:
             self._index = stop
         return uproot._util.tobytes(chunk.get(start, stop, self, context))

{uproot-5.3.2 → uproot-5.3.3}/src/uproot/streamers.py

@@ -1128,7 +1128,7 @@ class Model_TStreamerBasicType(Model_TStreamerElement):
             uproot.const.kULong,
             uproot.const.kLong,
         ):
-            self._bases[0]._members["fSize"] = numpy.dtype(numpy.compat.long).itemsize
+            self._bases[0]._members["fSize"] = numpy.dtype(numpy.int64).itemsize
 
         elif self._bases[0]._members["fType"] in (
             uproot.const.kULong64,

{uproot-5.3.2 → uproot-5.3.3}/src/uproot/version.py

@@ -12,5 +12,5 @@ __version__: str
 __version_tuple__: VERSION_TUPLE
 version_tuple: VERSION_TUPLE
 
-__version__ = version = '5.3.2'
-__version_tuple__ = version_tuple = (5, 3, 2)
+__version__ = version = '5.3.3'
+__version_tuple__ = version_tuple = (5, 3, 3)

{uproot-5.3.2 → uproot-5.3.3}/tests/test_0692_fsspec_reading.py

@@ -1,12 +1,5 @@
 # BSD 3-Clause License; see https://github.com/scikit-hep/uproot4/blob/main/LICENSE
 
-import pytest
-import uproot
-import uproot.source.fsspec
-import uproot.source.file
-import uproot.source.xrootd
-import uproot.source.s3
-
 from typing import BinaryIO
 import skhep_testdata
 import queue
@@ -15,6 +8,14 @@ import requests
 import os
 import sys
 
+import pytest
+
+import uproot
+import uproot.source.fsspec
+import uproot.source.file
+import uproot.source.xrootd
+import uproot.source.s3
+
 is_windows = sys.platform.startswith("win")
 
 
@@ -243,7 +244,7 @@ def test_fsspec_chunks(http_server):
             chunk = notifications.get()
             expected.pop((chunk.start, chunk.stop))
 
-        chunk_data_sum = {sum(chunk.raw_data) for chunk in chunks}
+        chunk_data_sum = {sum(map(int, chunk.raw_data)) for chunk in chunks}
         assert chunk_data_sum == {3967, 413, 10985}, "Chunk data does not match"
 
 

{uproot-5.3.2 → uproot-5.3.3}/tests/test_0692_fsspec_writing.py

@@ -51,12 +51,19 @@ def test_fsspec_writing_local(tmp_path, scheme):
     "slash_prefix",
     [""] if is_windows else ["", "/"],
 )
-def test_fsspec_writing_local_uri(tmp_path, scheme, slash_prefix, filename):
-    uri = scheme + slash_prefix + os.path.join(tmp_path, "some", "path", filename)
-    with uproot.create(uri) as f:
-        f["tree"] = {"x": np.array([1, 2, 3])}
-    with uproot.open(uri) as f:
-        assert f["tree"]["x"].array().tolist() == [1, 2, 3]
+def test_fsspec_writing_local_uri(tmp_path, scheme, slash_prefix, filename, request):
+    os.chdir(tmp_path)
+
+    try:
+        uri = scheme + slash_prefix + os.path.join(tmp_path, "some", "path", filename)
+
+        with uproot.create(uri) as f:
+            f["tree"] = {"x": np.array([1, 2, 3])}
+        with uproot.open(uri) as f:
+            assert f["tree"]["x"].array().tolist() == [1, 2, 3]
+
+    finally:
+        os.chdir(request.config.invocation_params.dir)
 
 
 @pytest.mark.parametrize(

{uproot-5.3.2 → uproot-5.3.3}/tests/test_1070_pandas_dataframe_building_performance_fix.py

@@ -1,9 +1,12 @@
 # BSD 3-Clause License; see https://github.com/scikit-hep/uproot5/blob/main/LICENSE
 
 import pytest
-import uproot
 import skhep_testdata
 
+import uproot
+
+pytest.importorskip("pandas")
+
 
 def test_pandas_performance_many_branches(tmp_path):
     for array in uproot.iterate(

{uproot-5.3.2 → uproot-5.3.3}/tests/test_1120_check_decompression_executor_pass_for_dask.py

@@ -1,9 +1,12 @@
 # BSD 3-Clause License; see https://github.com/scikit-hep/uproot5/blob/main/LICENSE
 
 import pytest
-import uproot
 import skhep_testdata
 
+import uproot
+
+pytest.importorskip("pandas")
+
 
 def test_decompression_executor_for_dask():
 

uproot-5.3.3/tests/test_1182_add_support_for_bitset.py

@@ -0,0 +1,40 @@
+# BSD 3-Clause License; see https://github.com/scikit-hep/uproot5/blob/main/LICENSE
+
+import uproot
+import skhep_testdata
+
+
+def test_add_support_for_bitset():
+    with uproot.open(skhep_testdata.data_path("uproot-issue-40.root")) as f:
+        assert (
+            repr(f["tree"]["bitset8"].interpretation)
+            == "AsObjects(AsBitSet(True, dtype('bool')))"
+        )
+        assert (
+            repr(f["tree"]["bitset16"].interpretation)
+            == "AsObjects(AsBitSet(True, dtype('bool')))"
+        )
+
+        assert f["tree"]["bitset8"].array().tolist() == [
+            [True, False, True, False, True, False, False, False]
+        ]
+        assert f["tree"]["bitset16"].array().tolist() == [
+            [
+                False,
+                True,
+                False,
+                True,
+                False,
+                True,
+                True,
+                True,
+                True,
+                False,
+                True,
+                False,
+                True,
+                False,
+                False,
+                False,
+            ]
+        ]

uproot-5.3.3/tests/test_1189_dask_failing_on_duplicate_keys.py

@@ -0,0 +1,24 @@
+# BSD 3-Clause License; see https://github.com/scikit-hep/uproot5/blob/main/LICENSE
+
+import pytest
+import skhep_testdata
+
+import uproot
+
+pytest.importorskip("pandas")
+
+
+def test_dask_duplicated_keys():
+
+    lazy = uproot.dask(
+        skhep_testdata.data_path("uproot-metadata-performance.root") + ":Events"
+    )
+    materialized = lazy.FatJet_btagDDBvLV2.compute()
+
+    lazy = uproot.dask(skhep_testdata.data_path("uproot-issue513.root") + ":Delphes")
+    materialized = lazy.Particle.compute()
+
+    lazy = uproot.dask(
+        skhep_testdata.data_path("uproot-issue443.root") + ":muonDataTree"
+    )
+    materialized = lazy.hitEnd.compute()