polystore 0.2.0__tar.gz → 0.2.2__tar.gz

{polystore-0.2.0/src/polystore.egg-info → polystore-0.2.2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: polystore
-Version: 0.2.0
+Version: 0.2.2
 Summary: Framework-agnostic multi-backend storage abstraction for ML and scientific computing
 Author-email: Tristan Simas <tristan.simas@mail.mcgill.ca>
 License-Expression: MIT
@@ -58,6 +58,7 @@ Requires-Dist: pytest-cov>=4.1.0; extra == "dev"
 Requires-Dist: coverage>=7.3.2; extra == "dev"
 Requires-Dist: genbadge[coverage]; extra == "dev"
 Requires-Dist: pandas>=2.3.1; extra == "dev"
+Requires-Dist: jpype1>=1.4.0; extra == "dev"
 Requires-Dist: black>=23.0.0; extra == "dev"
 Requires-Dist: ruff>=0.1.0; extra == "dev"
 Provides-Extra: docs

{polystore-0.2.0 → polystore-0.2.2}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "polystore"
-version = "0.2.0"
+version = "0.2.2"
 description = "Framework-agnostic multi-backend storage abstraction for ML and scientific computing"
 readme = "README.md"
 requires-python = ">=3.11"
@@ -94,6 +94,7 @@ dev = [
     "coverage>=7.3.2",
     "genbadge[coverage]",
     "pandas>=2.3.1",
+    "jpype1>=1.4.0",
     "black>=23.0.0",
     "ruff>=0.1.0",
 ]

{polystore-0.2.0 → polystore-0.2.2}/src/polystore/__init__.py

@@ -40,6 +40,7 @@ from .filemanager import FileManager
 from .formats import DEFAULT_IMAGE_EXTENSIONS, FileFormat
 from .imagej_runtime import (
     FIJI_IMAGEJ_RUNTIME,
+    ImageJJvmTeardown,
     ImageJRuntimePolicy,
     ImageJRuntimeUnavailableError,
 )
@@ -96,6 +97,7 @@ __all__ = [
     "Backend",
     "MemoryType",
     "FileFormat",
+    "ImageJJvmTeardown",
     "ImageJRuntimePolicy",
     "ImageJRuntimeUnavailableError",
     "FIJI_IMAGEJ_RUNTIME",

{polystore-0.2.0 → polystore-0.2.2}/src/polystore/disk.py

@@ -24,6 +24,7 @@ from .exceptions import StorageResolutionError
 from .formats import FileFormat
 
 logger = logging.getLogger(__name__)
+DISK_TEXT_ENCODING = "utf-8"
 
 
 def optional_import(module_name):
@@ -293,34 +294,36 @@ class DiskStorageBackend(StorageBackend):
         return numeric_arrays[0]
 
     def _text_writer(self, path, data, **kwargs):
-        """Write text data to file. Accepts and ignores extra kwargs for compatibility."""
-        path.write_text(str(data))
+        """Write UTF-8 text bytes without platform newline translation."""
+
+        del kwargs
+        path.write_bytes(str(data).encode(DISK_TEXT_ENCODING))
 
     def _text_reader(self, path):
-        return path.read_text()
+        return path.read_text(encoding=DISK_TEXT_ENCODING)
 
     def _json_writer(self, path, data, **kwargs):
         import json
 
         # Ensure parent directory exists
         path.parent.mkdir(parents=True, exist_ok=True)
-        path.write_text(json.dumps(data, indent=2))
+        self._text_writer(path, json.dumps(data, indent=2))
 
     def _json_reader(self, path):
         import json
 
-        return json.loads(path.read_text())
+        return json.loads(self._text_reader(path))
 
     def _csv_writer(self, path, data, **kwargs):
         import csv
 
         # Handle pre-formatted CSV strings (from pandas to_csv)
         if isinstance(data, str):
-            path.write_text(data)
+            self._text_writer(path, data)
             return
 
         # Assume data is a list of rows or a dict
-        with path.open("w", newline="") as f:
+        with path.open("w", newline="", encoding=DISK_TEXT_ENCODING) as f:
             if isinstance(data, dict):
                 # Write dict as CSV with headers
                 writer = csv.DictWriter(f, fieldnames=data.keys())
@@ -355,7 +358,7 @@ class DiskStorageBackend(StorageBackend):
     def _csv_reader(self, path):
         import csv
 
-        with path.open("r", newline="") as f:
+        with path.open("r", newline="", encoding=DISK_TEXT_ENCODING) as f:
             reader = csv.DictReader(f)
             return list(reader)
 

{polystore-0.2.0 → polystore-0.2.2}/src/polystore/imagej_runtime.py

@@ -5,6 +5,7 @@ from __future__ import annotations
 import logging
 import time
 from dataclasses import dataclass
+from enum import Enum
 from typing import Any
 
 logger = logging.getLogger(__name__)
@@ -14,6 +15,20 @@ class ImageJRuntimeUnavailableError(RuntimeError):
     """Raised when an ImageJ distribution cannot use its declared Java runtime."""
 
 
+class ImageJJvmTeardown(Enum):
+    """JPype teardown behavior owned by an ImageJ process runtime."""
+
+    PROCESS_EXIT = False
+    DESTROY_JVM = True
+
+    def configure(self) -> None:
+        """Apply this teardown behavior before the process starts its JVM."""
+
+        import jpype.config
+
+        jpype.config.destroy_jvm = self.value
+
+
 @dataclass(frozen=True, slots=True)
 class ImageJRuntimePolicy:
     """Own the endpoint and managed Java requirement for an ImageJ runtime."""
@@ -22,6 +37,7 @@ class ImageJRuntimePolicy:
     java_fetch: str
     java_vendor: str
     java_version: str
+    jvm_teardown: ImageJJvmTeardown
     initialization_retry_delays_seconds: tuple[float, ...] = ()
 
     @property
@@ -66,6 +82,7 @@ class ImageJRuntimePolicy:
     def configure_java(self, scyjava_module: Any) -> None:
         """Select managed Java before startup or validate the active JVM."""
 
+        self.jvm_teardown.configure()
         if scyjava_module.jvm_started():
             self.require_compatible_active_java(scyjava_module)
             return
@@ -105,5 +122,6 @@ FIJI_IMAGEJ_RUNTIME = ImageJRuntimePolicy(
     java_fetch="always",
     java_vendor="zulu-jre",
     java_version="21",
+    jvm_teardown=ImageJJvmTeardown.PROCESS_EXIT,
     initialization_retry_delays_seconds=(1.0, 2.0),
 )

{polystore-0.2.0 → polystore-0.2.2}/src/polystore/omero_local.py

@@ -28,7 +28,7 @@ import numpy as np
 
 from .array_payload import storage_numpy_array
 from .base import PicklableBackend, VirtualBackend
-from .omero_tables import OMERO_TABLE_SERVICE
+from .omero_tables import OMERO_TABLE_SERVICE, OMEROTableColumnType
 from .omero_text import OMEROTextFormat
 
 logger = logging.getLogger(__name__)
@@ -616,7 +616,6 @@ class OMEROLocalBackend(VirtualBackend, PicklableBackend):
         from io import StringIO
 
         import pandas as pd
-        from omero.grid import BoolColumn, DoubleColumn, LongColumn, StringColumn
         from omero.model import FileAnnotationI
         from omero.rtypes import rstring
 
@@ -650,30 +649,9 @@ class OMEROLocalBackend(VirtualBackend, PicklableBackend):
         # OMERO table names cannot contain dots except for the .h5 extension
         table_name = output_path.name.split(".")[0]
 
-        # Build column objects based on DataFrame dtypes
-        columns = []
-        for col_name in df.columns:
-            col_data = df[col_name]
-
-            # Determine column type from pandas dtype
-            if pd.api.types.is_integer_dtype(col_data):
-                col = LongColumn(col_name, "", [])
-                col.values = col_data.astype(int).tolist()
-            elif pd.api.types.is_float_dtype(col_data):
-                col = DoubleColumn(col_name, "", [])
-                col.values = col_data.astype(float).tolist()
-            elif pd.api.types.is_bool_dtype(col_data):
-                col = BoolColumn(col_name, "", [])
-                col.values = col_data.astype(bool).tolist()
-            else:
-                # Default to string column
-                # Calculate max string length (OMERO requires size > 0)
-                str_values = col_data.astype(str).tolist()
-                max_len = max(len(s) for s in str_values) if str_values else 1
-                col = StringColumn(col_name, "", max_len, [])
-                col.values = str_values
-
-            columns.append(col)
+        columns = [
+            OMEROTableColumnType.column_for(col_name, df[col_name]) for col_name in df.columns
+        ]
 
         table = OMERO_TABLE_SERVICE.create_table(conn, f"{table_name}.h5")
 

{polystore-0.2.0 → polystore-0.2.2}/src/polystore/omero_tables.py

@@ -4,11 +4,107 @@ from __future__ import annotations
 
 import logging
 import time
+from collections.abc import Callable
 from dataclasses import dataclass
+from enum import Enum
 from typing import Any
 
 logger = logging.getLogger(__name__)
 
+OMEROColumnMatcher = Callable[[Any], bool]
+OMEROColumnBuilder = Callable[[str, Any], Any]
+
+
+def _integer_column_matches(series: Any) -> bool:
+    import pandas as pd
+
+    return bool(pd.api.types.is_integer_dtype(series))
+
+
+def _float_column_matches(series: Any) -> bool:
+    import pandas as pd
+
+    return bool(pd.api.types.is_float_dtype(series))
+
+
+def _bool_column_matches(series: Any) -> bool:
+    import pandas as pd
+
+    return bool(pd.api.types.is_bool_dtype(series))
+
+
+def _fallback_column_matches(series: Any) -> bool:
+    del series
+    return True
+
+
+def _integer_column(name: str, series: Any) -> Any:
+    from omero.grid import LongColumn
+
+    column = LongColumn(name, "", [])
+    column.values = series.astype(int).tolist()
+    return column
+
+
+def _float_column(name: str, series: Any) -> Any:
+    from omero.grid import DoubleColumn
+
+    column = DoubleColumn(name, "", [])
+    column.values = series.astype(float).tolist()
+    return column
+
+
+def _bool_column(name: str, series: Any) -> Any:
+    from omero.grid import BoolColumn
+
+    column = BoolColumn(name, "", [])
+    column.values = series.astype(bool).tolist()
+    return column
+
+
+def _string_column(name: str, series: Any) -> Any:
+    from omero.grid import StringColumn
+
+    values = [str(value) for value in series.tolist()]
+    column = StringColumn(
+        name,
+        "",
+        max((len(value) for value in values), default=1),
+        [],
+    )
+    column.values = values
+    return column
+
+
+class OMEROTableColumnType(Enum):
+    """One pandas dtype family's complete OMERO column projection."""
+
+    def __new__(
+        cls,
+        identity: str,
+        matches: OMEROColumnMatcher,
+        build: OMEROColumnBuilder,
+    ):
+        member = object.__new__(cls)
+        member._value_ = identity
+        member._matches = matches
+        member._build = build
+        return member
+
+    INTEGER = ("integer", _integer_column_matches, _integer_column)
+    FLOAT = ("float", _float_column_matches, _float_column)
+    BOOLEAN = ("boolean", _bool_column_matches, _bool_column)
+    STRING = ("string", _fallback_column_matches, _string_column)
+
+    @classmethod
+    def column_for(cls, name: str, series: Any) -> Any:
+        """Build the first declared OMERO column matching one pandas series."""
+
+        return next(member for member in cls if member._matches(series))._build(
+            name,
+            series,
+        )
+
 
 class OMEROTableServiceUnavailableError(RuntimeError):
     """Raised when OMERO.tables cannot satisfy a table operation."""

{polystore-0.2.0 → polystore-0.2.2/src/polystore.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: polystore
-Version: 0.2.0
+Version: 0.2.2
 Summary: Framework-agnostic multi-backend storage abstraction for ML and scientific computing
 Author-email: Tristan Simas <tristan.simas@mail.mcgill.ca>
 License-Expression: MIT
@@ -58,6 +58,7 @@ Requires-Dist: pytest-cov>=4.1.0; extra == "dev"
 Requires-Dist: coverage>=7.3.2; extra == "dev"
 Requires-Dist: genbadge[coverage]; extra == "dev"
 Requires-Dist: pandas>=2.3.1; extra == "dev"
+Requires-Dist: jpype1>=1.4.0; extra == "dev"
 Requires-Dist: black>=23.0.0; extra == "dev"
 Requires-Dist: ruff>=0.1.0; extra == "dev"
 Provides-Extra: docs

{polystore-0.2.0 → polystore-0.2.2}/src/polystore.egg-info/requires.txt

@@ -25,6 +25,7 @@ pytest-cov>=4.1.0
 coverage>=7.3.2
 genbadge[coverage]
 pandas>=2.3.1
+jpype1>=1.4.0
 black>=23.0.0
 ruff>=0.1.0
 

{polystore-0.2.0 → polystore-0.2.2}/tests/test_disk_backend.py

@@ -3,6 +3,7 @@
 Keep tests small and focused — these hit CSV/JSON/TEXT handlers, listing,
 ensure_directory idempotence, and symlink creation.
 """
+
 from pathlib import Path
 
 import numpy as np
@@ -38,6 +39,23 @@ def test_text_json_csv_save_load(tmp_path: Path):
     assert loaded[0]["a"] == "1"
 
 
+def test_preformatted_text_is_written_as_exact_utf8_bytes(
+    tmp_path: Path,
+    monkeypatch,
+) -> None:
+    content = "well,count\r\nA01,2\r\n"
+    output_path = tmp_path / "summary.csv"
+
+    def reject_platform_text_translation(*args, **kwargs):
+        del args, kwargs
+        raise AssertionError("text persistence must bypass platform newline translation")
+
+    monkeypatch.setattr(Path, "write_text", reject_platform_text_translation)
+    DiskBackend().save(content, output_path)
+
+    assert output_path.read_bytes() == content.encode(disk_module.DISK_TEXT_ENCODING)
+
+
 def test_tiff_save_normalizes_external_array_payload(tmp_path, monkeypatch) -> None:
     class ExternalArray:
         pass

{polystore-0.2.0 → polystore-0.2.2}/tests/test_imagej_runtime.py

@@ -6,6 +6,7 @@ import pytest
 
 from polystore.imagej_runtime import (
     FIJI_IMAGEJ_RUNTIME,
+    ImageJJvmTeardown,
     ImageJRuntimePolicy,
     ImageJRuntimeUnavailableError,
 )
@@ -49,7 +50,10 @@ class _ImageJ:
         return self.gateway
 
 
-def test_fiji_runtime_selects_managed_java_before_initialization() -> None:
+def test_fiji_runtime_selects_managed_java_before_initialization(monkeypatch) -> None:
+    import jpype.config
+
+    monkeypatch.setattr(jpype.config, "destroy_jvm", True)
     scyjava = _ScyJava(started=False, java_version="21.0.8")
     imagej = _ImageJ(scyjava)
 
@@ -58,6 +62,7 @@ def test_fiji_runtime_selects_managed_java_before_initialization() -> None:
     assert gateway is imagej.gateway
     assert scyjava.constraints == [{"fetch": "always", "vendor": "zulu-jre", "version": "21"}]
     assert imagej.calls == [("sc.fiji:fiji", "headless")]
+    assert jpype.config.destroy_jvm is False
 
 
 def test_runtime_accepts_compatible_active_java_without_reconfiguration() -> None:
@@ -90,6 +95,7 @@ def test_runtime_reports_an_unparseable_active_java_version() -> None:
         java_fetch="always",
         java_vendor="example-jre",
         java_version="21",
+        jvm_teardown=ImageJJvmTeardown.PROCESS_EXIT,
     )
     scyjava = _ScyJava(started=True, java_version="unknown")
 
@@ -132,6 +138,7 @@ def test_runtime_retries_initialization_only_before_jvm_start(monkeypatch) -> No
         java_fetch="always",
         java_vendor="example-jre",
         java_version="21",
+        jvm_teardown=ImageJJvmTeardown.PROCESS_EXIT,
         initialization_retry_delays_seconds=(0.1, 0.2),
     )
     scyjava = _ScyJava(started=False, java_version="21")
@@ -162,6 +169,7 @@ def test_runtime_preserves_final_failure_after_retry_schedule() -> None:
         java_fetch="always",
         java_vendor="example-jre",
         java_version="21",
+        jvm_teardown=ImageJJvmTeardown.PROCESS_EXIT,
         initialization_retry_delays_seconds=(0.0, 0.0),
     )
     scyjava = _ScyJava(started=False, java_version="21")

{polystore-0.2.0 → polystore-0.2.2}/tests/test_omero_tables.py

@@ -1,10 +1,13 @@
 """Tests for the declaration-owned OMERO.tables capability boundary."""
 
-from types import SimpleNamespace
+import sys
+from types import ModuleType, SimpleNamespace
 
+import pandas as pd
 import pytest
 
 from polystore.omero_tables import (
+    OMEROTableColumnType,
     OMEROTableService,
     OMEROTableServiceUnavailableError,
 )
@@ -12,6 +15,64 @@ from polystore.omero_tables import (
 _DEFAULT_TABLE = object()
 
 
+class _Column:
+    def __init__(self, *declaration) -> None:
+        self.declaration = declaration
+        self.values = []
+
+
+class _LongColumn(_Column):
+    pass
+
+
+class _DoubleColumn(_Column):
+    pass
+
+
+class _BoolColumn(_Column):
+    pass
+
+
+class _StringColumn(_Column):
+    pass
+
+
+@pytest.fixture
+def omero_grid_columns(monkeypatch):
+    grid = ModuleType("omero.grid")
+    grid.LongColumn = _LongColumn
+    grid.DoubleColumn = _DoubleColumn
+    grid.BoolColumn = _BoolColumn
+    grid.StringColumn = _StringColumn
+    omero = ModuleType("omero")
+    omero.__path__ = []
+    monkeypatch.setitem(sys.modules, "omero", omero)
+    monkeypatch.setitem(sys.modules, "omero.grid", grid)
+
+
+@pytest.mark.parametrize(
+    ("series", "column_type", "values"),
+    (
+        (pd.Series([1, 2]), _LongColumn, [1, 2]),
+        (pd.Series([1.5, 2.5]), _DoubleColumn, [1.5, 2.5]),
+        (pd.Series([True, False]), _BoolColumn, [True, False]),
+        (pd.Series(["A01", float("nan")], dtype=object), _StringColumn, ["A01", "nan"]),
+    ),
+)
+def test_table_column_type_owns_dtype_projection(
+    omero_grid_columns,
+    series,
+    column_type,
+    values,
+) -> None:
+    column = OMEROTableColumnType.column_for("value", series)
+
+    assert isinstance(column, column_type)
+    assert column.values == values
+    if isinstance(column, _StringColumn):
+        assert column.declaration == ("value", "", 3, [])
+
+
 class _Resources:
     def __init__(
         self,