ENH: Add calamite engine to read_excel

ci/deps/actions-310.yaml

@@ -55,3 +55,6 @@ dependencies:
   - xlrd
   - xlsxwriter
   - zstandard
+
+  - pip:
+    - python-calamine

ci/deps/actions-38-downstream_compat.yaml

@@ -69,3 +69,6 @@ dependencies:
   - pandas-gbq
   - pyyaml
   - py
+
+  - pip:
+    - python-calamine

ci/deps/actions-38-minimum_versions.yaml

@@ -61,3 +61,4 @@ dependencies:
 
   - pip:
     - pyqt5==5.15.1
+    - python-calamine==0.0.5

ci/deps/actions-38.yaml

@@ -44,6 +44,7 @@ dependencies:
   - pymysql
   - pyreadstat
   - pytables
+  - python-calamine
   - python-snappy
   - pyxlsb
   - s3fs>=2021.08.0
@@ -54,3 +55,6 @@ dependencies:
   - xlrd
   - xlsxwriter
   - zstandard
+
+  - pip:
+    - python-calamine

ci/deps/actions-39.yaml

@@ -55,3 +55,6 @@ dependencies:
   - xlrd
   - xlsxwriter
   - zstandard
+
+  - pip:
+    - python-calamine

ci/deps/circle-38-arm64.yaml

@@ -55,3 +55,6 @@ dependencies:
   - xlrd
   - xlsxwriter
   - zstandard
+
+  - pip:
+    - python-calamine

doc/source/whatsnew/v2.0.0.rst

@@ -106,7 +106,7 @@ 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``)
 
 .. ---------------------------------------------------------------------------
 .. _whatsnew_200.notable_bug_fixes:

environment.yml

@@ -118,3 +118,4 @@ dependencies:
 
   - pip:
       - sphinx-toggleprompt
+      - python-calamine

pandas/compat/_optional.py

@@ -48,6 +48,7 @@
     "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

pandas/io/excel/_base.py

@@ -1435,6 +1435,7 @@ class ExcelFile:
             This is not supported, switch to using ``openpyxl`` instead.
     """
 
+    from pandas.io.excel._calamitereader import CalamineExcelReader
     from pandas.io.excel._odfreader import ODFReader
     from pandas.io.excel._openpyxl import OpenpyxlReader
     from pandas.io.excel._pyxlsb import PyxlsbReader
@@ -1445,6 +1446,7 @@ class ExcelFile:
         "openpyxl": OpenpyxlReader,
         "odf": ODFReader,
         "pyxlsb": PyxlsbReader,
+        "calamite": CalamineExcelReader,
     }
 
     def __init__(

pandas/io/excel/_calamitereader.py

@@ -0,0 +1,69 @@
+from __future__ import annotations
+
+from io import (
+    BufferedReader,
+    BytesIO,
+)
+from pathlib import PurePath
+from tempfile import NamedTemporaryFile
+from typing import Any
+
+from python_calamine import (
+    get_sheet_data,
+    get_sheet_names,
+)
+
+from pandas.io.excel._base import (
+    BaseExcelReader,
+    inspect_excel_format,
+)
+
+
+class __calamine__:
+    pass
+
+
+class CalamineExcelReader(BaseExcelReader):
+    book: str
+    _sheet_names: list[str] | None = None
+
+    @property
+    def _workbook_class(self) -> type[__calamine__]:
+        return __calamine__
+
+    def load_workbook(
+        self, filepath_or_buffer: str | PurePath | BufferedReader | BytesIO
+    ) -> str:
+        if isinstance(filepath_or_buffer, BufferedReader):
+            filepath_or_buffer = filepath_or_buffer.name
+
+        elif isinstance(filepath_or_buffer, BytesIO):
+            ext = inspect_excel_format(filepath_or_buffer)
+            with NamedTemporaryFile(suffix=f".{ext}", delete=False) as tmp_file:
+                tmp_file.write(filepath_or_buffer.getvalue())
+                filepath_or_buffer = tmp_file.name
+
+        elif isinstance(filepath_or_buffer, PurePath):
+            filepath_or_buffer = filepath_or_buffer.as_posix()
+
+        assert isinstance(filepath_or_buffer, str)
+
+        self._sheet_names = get_sheet_names(filepath_or_buffer)
+        return filepath_or_buffer
+
+    @property
+    def sheet_names(self) -> list[str]:
+        if self._sheet_names is None:
+            self._sheet_names = get_sheet_names(self.book)
+        return self._sheet_names
+
+    def get_sheet_by_name(self, name: str) -> int:
+        self.raise_if_bad_sheet_by_name(name)
+        return self.sheet_names.index(name)
+
+    def get_sheet_by_index(self, index: int) -> int:
+        self.raise_if_bad_sheet_by_index(index)
+        return index
+
+    def get_sheet_data(self, sheet: int, convert_float: bool) -> list[list[Any]]:
+        return get_sheet_data(self.book, sheet)

pandas/tests/io/excel/test_readers.py

@@ -52,6 +52,7 @@
     ),
     pytest.param("pyxlsb", marks=td.skip_if_no("pyxlsb")),
     pytest.param("odf", marks=td.skip_if_no("odf")),
+    pytest.param("calamite"),
 ]
 
 

pyproject.toml

@@ -104,6 +104,7 @@ all = ['beautifulsoup4>=4.9.3',
        'pytest>=6.0',
        'pytest-xdist>=1.31',
        'pytest-asyncio>=0.17.0',
+       'python-calamine>=0.0.5',
        'python-snappy>=0.6.0',
        'pyxlsb>=1.0.8',
        'qtpy>=2.2.0',

requirements-dev.txt

@@ -87,4 +87,5 @@ feedparser
 pyyaml
 requests
 sphinx-toggleprompt
+python-calamine
 setuptools>=61.0.0