Merge pull request #1 from dimastbk/issue-50395-dima

calamite -> calamine, updated some tests for calamine

ci/deps/actions-38-minimum_versions.yaml

@@ -61,4 +61,4 @@ dependencies:
 
   - pip:
     - pyqt5==5.15.1
-    - python-calamine==0.0.5
+    - python-calamine==0.0.6

doc/source/whatsnew/v2.0.0.rst

@@ -165,7 +165,7 @@ Other enhancements
 - Improved error message in :func:`to_datetime` for non-ISO8601 formats, informing users about the position of the first error (:issue:`50361`)
 - Improved error message when trying to align :class:`DataFrame` objects (for example, in :func:`DataFrame.compare`) to clarify that "identically labelled" refers to both index and columns (:issue:`50083`)
 - Performance improvement in :func:`to_datetime` when format is given or can be inferred (:issue:`50465`)
-- Added ``calamite`` as an engine to ``read_excel`` (:issue: ``50395``)
+- Added ``calamine`` as an engine to ``read_excel`` (:issue: ``50395``)
 - Added :meth:`DatetimeIndex.as_unit` and :meth:`TimedeltaIndex.as_unit` to convert to different resolutions; supported resolutions are "s", "ms", "us", and "ns" (:issue:`50616`)
 
 .. ---------------------------------------------------------------------------

pandas/compat/_optional.py

@@ -34,6 +34,7 @@
     "pyarrow": "6.0.0",
     "pyreadstat": "1.1.2",
     "pytest": "7.0.0",
+    "python-calamine": "0.0.6",
     "pyxlsb": "1.0.8",
     "s3fs": "2021.08.0",
     "scipy": "1.7.1",
@@ -48,7 +49,6 @@
     "tzdata": "2022.1",
     "qtpy": "2.2.0",
     "pyqt5": "5.15.1",
-    "python-calamine": "0.0.5",
 }
 
 # A mapping from import name to package name (on PyPI) for packages where
@@ -62,6 +62,7 @@
     "lxml.etree": "lxml",
     "odf": "odfpy",
     "pandas_gbq": "pandas-gbq",
+    "python_calamine": "python-calamine",
     "snappy": "python-snappy",
     "sqlalchemy": "SQLAlchemy",
     "tables": "pytables",

pandas/io/excel/_calaminereader.py

@@ -8,7 +8,7 @@
 from tempfile import NamedTemporaryFile
 from typing import Any
 
-from pandas.compat._optional import import_optional_dependency
+from pandas._typing import StorageOptions
 
 from pandas.io.excel._base import (
     BaseExcelReader,
@@ -23,7 +23,12 @@ class __calamine__:
 class CalamineExcelReader(BaseExcelReader):
     book: str
     _sheet_names: list[str] | None = None
-    import_optional_dependency("python_calamine")
+
+    def __init__(
+        self, filepath_or_buffer, storage_options: StorageOptions = None
+    ) -> None:
+        import_optional_dependency("python_calamine")
+        super().__init__(filepath_or_buffer, storage_options=storage_options)
 
     @property
     def _workbook_class(self) -> type[__calamine__]:

pandas/tests/io/excel/test_readers.py

@@ -52,7 +52,7 @@
     ),
     pytest.param("pyxlsb", marks=td.skip_if_no("pyxlsb")),
     pytest.param("odf", marks=td.skip_if_no("odf")),
-    pytest.param("calamine", marks=td.skip_if_no("calamine")),
+    pytest.param("calamine", marks=td.skip_if_no("python_calamine")),
 ]
 
 
@@ -66,11 +66,11 @@ def _is_valid_engine_ext_pair(engine, read_ext: str) -> bool:
         return False
     if engine == "odf" and read_ext != ".ods":
         return False
-    if read_ext == ".ods" and engine != "odf":
+    if read_ext == ".ods" and engine not in {"odf", "calamine"}:
         return False
     if engine == "pyxlsb" and read_ext != ".xlsb":
         return False
-    if read_ext == ".xlsb" and engine != "pyxlsb":
+    if read_ext == ".xlsb" and engine not in {"pyxlsb", "calamine"}:
         return False
     if engine == "xlrd" and read_ext != ".xls":
         return False
@@ -834,6 +834,11 @@ def test_corrupt_bytes_raises(self, engine):
                 "Unsupported format, or corrupt file: Expected BOF "
                 "record; found b'foo'"
             )
+        elif engine == "calamine":
+            import python_calamine
+
+            error = python_calamine._python_calamine.CalamineError
+            msg = "Cannot detect file format"
         else:
             error = BadZipFile
             msg = "File is not a zip file"
@@ -1374,7 +1379,7 @@ def test_trailing_blanks(self, read_ext):
 
     def test_ignore_chartsheets_by_str(self, request, engine, read_ext):
         # GH 41448
-        if engine == "odf":
+        if read_ext == ".ods":
             pytest.skip("chartsheets do not exist in the ODF format")
         if engine == "pyxlsb":
             request.node.add_marker(
@@ -1387,7 +1392,7 @@ def test_ignore_chartsheets_by_str(self, request, engine, read_ext):
 
     def test_ignore_chartsheets_by_int(self, request, engine, read_ext):
         # GH 41448
-        if engine == "odf":
+        if read_ext == ".ods":
             pytest.skip("chartsheets do not exist in the ODF format")
         if engine == "pyxlsb":
             request.node.add_marker(
@@ -1656,7 +1661,7 @@ def test_engine_invalid_option(self, read_ext):
 
     def test_ignore_chartsheets(self, request, engine, read_ext):
         # GH 41448
-        if engine == "odf":
+        if read_ext == ".ods":
             pytest.skip("chartsheets do not exist in the ODF format")
         if engine == "pyxlsb":
             request.node.add_marker(
@@ -1676,6 +1681,10 @@ def test_corrupt_files_closed(self, engine, read_ext):
             import xlrd
 
             errors = (BadZipFile, xlrd.biffh.XLRDError)
+        elif engine == "calamine":
+            import python_calamine
+
+            errors = (python_calamine._python_calamine.CalamineError,)
 
         with tm.ensure_clean(f"corrupt{read_ext}") as file:
             Path(file).write_text("corrupt")

pyproject.toml

@@ -104,7 +104,7 @@ all = ['beautifulsoup4>=4.9.3',
        'pytest>=7.0.0',
        'pytest-xdist>=2.2.0',
        'pytest-asyncio>=0.17.0',
-       'python-calamine>=0.0.5',
+       'python-calamine>=0.0.6',
        'python-snappy>=0.6.0',
        'pyxlsb>=1.0.8',
        'qtpy>=2.2.0',