Allow multiple netcdf files

Author: mcuntz

.zenodo.json

@@ -38,5 +38,5 @@
         }
     ],
     "access_right": "open",
-    "description": "<p><strong>ncvue</strong> is a minimal GUI for a quick view of netCDF files. It is aiming to be a drop-in replacement for <a href=\"http://meteora.ucsd.edu/~pierce/ncview_home_page.html\">ncview</a>, being slightly more general than <a href=\"http://meteora.ucsd.edu/~pierce/ncview_home_page.html\">ncview</a>, which targets maps. If <strong>ncvue</strong> is used with maps, it supports mostly structured grids, more precisely the grids supported by <a href=\"https://scitools.org.uk/cartopy/docs/latest/\">cartopy</a>.</p>\n\n<p><strong>ncvue</strong> is a Python script that can be called from within Python or as a command line tool. It is not supposed to produce publication-ready plots but rather provide a quick overview of the netCDF file.</p>\n\n<p>The complete documentation for <strong>ncvue</strong> is available from: <a href=\"https://mcuntz.github.io/ncvue/\">https://mcuntz.github.io/ncvue/</a></p>\n\n<p>The current version features three panels, that means three different plotting styles: Scatter/Line plots, Contour Plots, and Maps, with extensive tooltips on buttons, sliders, entry boxes, spinboxes, and menus. It includes the possibility to open and change netCDF files within <strong>ncvue</strong>. The netCDF file will be changed in all panels of the primary window and any secondary window once focus changes on the panel or window. Graphical documentation exists on <a href=\"https://mcuntz.github.io/ncvue/\">Github Pages</a>.</p>\n\n<p>The current version also provides two standalone applications on macOS and on Windows that come with all necessary libraries to run <strong>ncvue</strong>, including Python. Links to the installers are given on <a href=\"https://github.com/mcuntz/ncvue/\">Github</a> and in the <a href=\"https://mcuntz.github.io/ncvue/\">documentation</a>.</p>\n\n<p>Version 3.6 added different designs on macOS, Windows and Linux.</p>\n\n<p>Version 4.0 moved to a new project structure, using pyproject.toml, automatic versioning, src layout, and Github actions for continuous integration.</p>\n\n<p>Version 4.1 ported <strong>ncvue</strong> to conda-forge.</p>\n\n<p>Version 4.2 allowed groups in netcdf files.</p>"
+    "description": "<p><strong>ncvue</strong> is a minimal GUI for a quick view of netCDF files. It is aiming to be a drop-in replacement for <a href=\"http://meteora.ucsd.edu/~pierce/ncview_home_page.html\">ncview</a>, being slightly more general than <a href=\"http://meteora.ucsd.edu/~pierce/ncview_home_page.html\">ncview</a>, which targets maps. If <strong>ncvue</strong> is used with maps, it supports mostly structured grids, more precisely the grids supported by <a href=\"https://scitools.org.uk/cartopy/docs/latest/\">cartopy</a>.</p>\n\n<p><strong>ncvue</strong> is a Python script that can be called from within Python or as a command line tool. It is not supposed to produce publication-ready plots but rather provide a quick overview of the netCDF file.</p>\n\n<p>The complete documentation for <strong>ncvue</strong> is available from: <a href=\"https://mcuntz.github.io/ncvue/\">https://mcuntz.github.io/ncvue/</a></p>\n\n<p>The current version features three panels, that means three different plotting styles: Scatter/Line plots, Contour Plots, and Maps, with extensive tooltips on buttons, sliders, entry boxes, spinboxes, and menus. It includes the possibility to open and change netCDF files within <strong>ncvue</strong>. The netCDF file will be changed in all panels of the primary window and any secondary window once focus changes on the panel or window. Graphical documentation exists on <a href=\"https://mcuntz.github.io/ncvue/\">Github Pages</a>.</p>\n\n<p>The current version also provides two standalone applications on macOS and on Windows that come with all necessary libraries to run <strong>ncvue</strong>, including Python. Links to the installers are given on <a href=\"https://github.com/mcuntz/ncvue/\">Github</a> and in the <a href=\"https://mcuntz.github.io/ncvue/\">documentation</a>.</p>\n\n<p>Version 3.6 added different designs on macOS, Windows and Linux.</p>\n\n<p>Version 4.0 moved to a new project structure, using pyproject.toml, automatic versioning, src layout, and Github actions for continuous integration.</p>\n\n<p>Version 4.1 ported <strong>ncvue</strong> to conda-forge.</p>\n\n<p>Version 4.2 allowed groups in netcdf files.</p>\n\n<p>Version 4.3 allowed multiple netcdf files.</p>"
 }

CHANGELOG.rst

@@ -1,7 +1,13 @@
 Changelog
 ---------
 
+v4.3 (Jan 2024)
+    * Added conda and continuous integration badges.
+    * Allow multiple netcdf files.
+    * Squeeze output in `get_slice_miss` only if more than 1 dimension.
+
 v4.2 (Jan 2024)
+    * Updated versions of github actions.
     * Changed to sphinx_book_theme for documentation.
     * Use local copy of `tooltip.py` from idle.
     * Allow groups in netcdf files.

README.rst

@@ -21,12 +21,16 @@ replacement for ncview_ and panoply_.
    :target: https://badge.fury.io/py/ncvue
    :alt: PyPI version
 
+.. image:: https://img.shields.io/conda/vn/conda-forge/ncvue.svg
+   :target: https://anaconda.org/conda-forge/ncvue
+   :alt: Conda version
+
 .. image:: http://img.shields.io/badge/license-MIT-blue.svg?style=flat
    :target: https://github.com/mcuntz/ncvue/blob/master/LICENSE
    :alt: License
 
-.. image:: https://github.com/mcuntz/ncvue/workflows/Continuous%20Integration/badge.svg?branch=main
-   :target: https://github.com/mcuntz/ncvue/actions
+.. image:: https://github.com/mcuntz/ncvue/actions/workflows/main.yml/badge.svg?branch=main
+   :target: https://github.com/mcuntz/ncvue/actions/workflows/main.yml
    :alt: Build status
 
 About ncvue
@@ -56,54 +60,45 @@ Quick usage guide
 .. code-block:: bash
 
    ncvue netcdf_file.nc
+   ncvue netcdf_file1.nc netcdf_file2.nc
 
 or from within Python:
 
 .. code-block:: python
 
    from ncvue import ncvue
-   ncvue('netcdf_file.nc')
+   ncvue(['netcdf_file.nc'])
 
 where the netCDF file is optional. The latter can also be left out and
 a netCDF file can be opened with the "Open File" button from within
-``ncvue``.
+``ncvue``. The netCDF has to be given in a list because several netcdf
+files can be given as in the second example from the command line.
 
 Note, ``ncvue`` uses the `TkAgg` backend of `matplotlib`. It must be
 called before any other call to `matplotlib`. This also means that you
 cannot launch it from within `iPython` if it was launched with
 `--pylab`. It can be called from within a standard `iPython`, though,
 or using `ipython --gui tk`.
 
-When using ``ncvue`` with `jupyter` notebooks, one has to set
-`%matplotlib inline` before the import and call of ``ncvue``. You have
-set `%matplotlib inline` again if you want to continue having inline
-plots in `jupyter` afterwards.
-
-.. code-block:: python
-
-   %matplotlib inline
-   from ncvue import ncvue
-   ncvue('netcdf_file.nc')
-   %matplotlib inline
-
-One can also install standalone macOS or Windows applications that
-come with everything needed to run ``ncvue`` including Python:
+..
+   One can also install standalone macOS or Windows applications that
+   come with everything needed to run ``ncvue`` including Python:
 
-- `macOS app`_ (macOS > 10.13 [High Sierra] on Intel)
-- `Windows executable`_ (Windows 10)
+   - `macOS app`_ (macOS > 10.13 [High Sierra] on Intel)
+   - `Windows executable`_ (Windows 10)
 
-The macOS app should work from macOS 10.13 (High Sierra) onward on
-Intel processors. There is no standalone application for macOS on
-Apple Silicon (M1) chips because I do not have a paid Apple
-Developer ID. Other installation options work, though.
+   The macOS app should work from macOS 10.13 (High Sierra) onward on
+   Intel processors. There is no standalone application for macOS on
+   Apple Silicon (M1) chips because I do not have a paid Apple
+   Developer ID. Other installation options work, though.
 
-A dialog box might pop up on macOS saying that the ``ncvue.app`` is
-from an unidentified developer. This is because ``ncvue`` is an
-open-source software.  Depending on the macOS version, it offers to
-open it anyway. In later versions of macOS, this option is only given
-if you right-click (or control-click) on the ``ncvue.app`` and choose
-`Open`. You only have to do this once. It will open like any other
-application the next times.
+   A dialog box might pop up on macOS saying that the ``ncvue.app`` is
+   from an unidentified developer. This is because ``ncvue`` is an
+   open-source software.  Depending on the macOS version, it offers to
+   open it anyway. In later versions of macOS, this option is only given
+   if you right-click (or control-click) on the ``ncvue.app`` and choose
+   `Open`. You only have to do this once. It will open like any other
+   application the next times.
 
 General layout
 ^^^^^^^^^^^^^^
@@ -220,9 +215,9 @@ Installation
 ------------
 
 ``ncvue`` is an application written in Python. If you have Python
- installed, then the best is to install ``ncvue`` within the Python
- universe. The easiest way to install ``ncvue`` is thence via `pip` if
- you have cartopy_ installed already:
+installed, then the best is to install ``ncvue`` within the Python
+universe. The easiest way to install ``ncvue`` is thence via `pip` if
+you have cartopy_ installed already:
 
 .. code-block:: bash
 
@@ -236,14 +231,15 @@ installing, for example, Miniconda_:
 
    conda install -c conda-forge ncvue
 
-We also provide a standalone `macOS app`_ and a `Windows executable`_
-that come with everything needed to run ``ncvue`` including
-Python. The macOS app should work from macOS 10.13 (High Sierra)
-onward. It is, however, only tested on macOS 10.15 (Catalina). Drop me
-a message if it does not work on newer operating systems.
+..
+   We also provide a standalone `macOS app`_ and a `Windows executable`_
+   that come with everything needed to run ``ncvue`` including
+   Python. The macOS app should work from macOS 10.13 (High Sierra)
+   onward. It is, however, only tested on macOS 10.15 (Catalina). Drop me
+   a message if it does not work on newer operating systems.
 
 See the installation instructions_ in the documentation_ for more
-information.
+information on installing `Cartopy` and ``ncvue with pip``.
 
 License
 -------
@@ -255,8 +251,9 @@ Copyright (c) 2020-2024 Matthias Cuntz
 
 ``ncvue`` uses the Azure_ 2.0 theme by rdbende_ on Linux and Windows.
 
-Standalone applications are produced with `cx_Freeze`_, currently
-maintained by `Marcelo Duarte`_.
+..
+   Standalone applications are produced with `cx_Freeze`_, currently
+   maintained by `Marcelo Duarte`_.
 
 The project structure of ``ncvue`` was very originally based on a
 template_ provided by `Sebastian Müller`_ but has evolved

cx_freeze_setup.py

@@ -3,25 +3,33 @@
 Make stand-alone version of ncvue with cx_Freeze.
 
 On macOS, use minimal virtual environment
-    pyenv virtualenv 3.12.1 install-ncvue
-    pyenv rehash
-    pyenv local install-ncvue
-    pip install --upgrade cython numpy pyshp six
-    isshapely=$(pip freeze | grep shapely)
-    if [[ -z ${isshapely} ]] ; then pip uninstall -y shapely ; fi
-    pip install shapely --no-binary shapely
-    pip install scipy
-    pip install matplotlib
-    pip install pykdtree
-    pip install netcdf4
-    pip install cartopy
-    pip install flake8  # for Emacs
-    pip install cx_Freeze
-    # libtiff.5.dylib version of PIL too old for pyproj -> use current from homebrew
-    mv ~/.pyenv/versions/3.12.1/envs/install-ncvue/lib/python3.8/site-packages/PIL/.dylibs/libtiff.5.dylib \
-       ~/.pyenv/versions/3.12.1/envs/install-ncvue/lib/python3.8/site-packages/PIL/.dylibs/libtiff.5.dylib.save
-    \cp /usr/local/lib/libtiff.5.dylib \
-        ~/.pyenv/versions/3.12.1/envs/install-ncvue/lib/python3.8/site-packages/PIL/.dylibs/libtiff.5.dylib
+   # essential subset
+   if [[ "$(uname -m)" == "arm64" ]] ; then
+       export OPENBLAS="$(brew --prefix openblas)"
+       export HDF5_DIR="$(brew --prefix hdf5)"
+       export GEOS_DIR="$(brew --prefix geos)"
+       export GEOS_CONFIG="$(brew --prefix geos)/bin/geos-config"
+   fi
+   pyenv virtualenv 3.11.7 install-ncvue
+   pyenv rehash
+   pyenv local install-ncvue
+   pyenv rehash
+   # test if install works
+   python -m pip install --upgrade cython numpy pyshp six
+   [[ -z $(python -m pip freeze | grep shapely) ]] && \
+           python -m pip uninstall -y shapely
+   python -m pip install shapely --no-binary shapely
+   python -m pip install scipy matplotlib pykdtree
+   python -m pip install netcdf4 cartopy flake8
+   python -m pip install GDAL==$(gdal-config --version) \
+       --global-option=build_ext --global-option="-I${HOMEBREW_PREFIX}/include"
+   python -m pip install cartopy
+   python -m pip install cx_Freeze
+   # libtiff.5.dylib version of PIL too old for pyproj -> use current from homebrew
+   mv ~/.pyenv/versions/3.11.7/envs/install-ncvue/lib/python3.8/site-packages/PIL/.dylibs/libtiff.5.dylib \
+       ~/.pyenv/versions/3.11.7/envs/install-ncvue/lib/python3.8/site-packages/PIL/.dylibs/libtiff.5.dylib.save
+   \cp /usr/local/lib/libtiff.5.dylib \
+       ~/.pyenv/versions/3.11.7/envs/install-ncvue/lib/python3.8/site-packages/PIL/.dylibs/libtiff.5.dylib
 
 On Windows, use conda-forge for everything because more up-to-date
     # Do not use mkl for smaller executable with PyInstaller/cx_Freeze
@@ -41,6 +49,7 @@
     python cx_freeze_setup.py bdist_dmg
 Windows installer
     python.exe cx_freeze_setup.py bdist_msi
+
 '''
 import os
 import codecs
@@ -51,7 +60,7 @@
 import shutil
 
 from cx_Freeze import setup, Executable
-from cx_Freeze.dist import build as _build
+# from cx_Freeze.dist import build as _build
 
 
 # find __version__
@@ -99,24 +108,24 @@ def _post_build_m1(exedir):
             shutil.copy2(aa, exedir)
 
 
-# post build hook
-# https://stackoverflow.com/questions/17806485/execute-a-python-script-post-install-using-distutils-setuptools
-class build(_build):
-    def run(self):
-        _build.run(self)
-        if sys.platform == 'win32':
-            self.execute(_post_build_win, (self.build_exe,),
-                         msg='Post-build on Windows')
-        elif sys.platform == 'darwin':
-            if platform.machine() == 'arm64':
-                # self.execute(_post_build_m1, (self.build_exe,),
-                #              msg='Post-build on macOS Apple Silicon (M1)')
-                pass
-            else:
-                self.execute(_post_build_mac, (self.build_exe,),
-                             msg='Post-build on macOS')
-        else:
-            pass
+# # post build hook
+# # https://stackoverflow.com/questions/17806485/execute-a-python-script-post-install-using-distutils-setuptools
+# class build(_build):
+#     def run(self):
+#         _build.run(self)
+#         if sys.platform == 'win32':
+#             self.execute(_post_build_win, (self.build_exe,),
+#                          msg='Post-build on Windows')
+#         elif sys.platform == 'darwin':
+#             if platform.machine() == 'arm64':
+#                 # self.execute(_post_build_m1, (self.build_exe,),
+#                 #              msg='Post-build on macOS Apple Silicon (M1)')
+#                 pass
+#             else:
+#                 self.execute(_post_build_mac, (self.build_exe,),
+#                              msg='Post-build on macOS')
+#         else:
+#             pass
 
 
 package   = 'ncvue'
@@ -128,7 +137,7 @@ def run(self):
 
 version = _find_version('src/' + package, '_version.py')
 
-script        = 'bin/ncvue'
+script        = 'bin.save/ncvue'  # 'src/ncvue/__main__.py'
 packages      = ['scipy', 'netCDF4']  # others detected automatically
 excludes      = ['pyflakes', 'mccabe', 'pycodestyle', 'flake8',  # flake8
                  'gtk', 'PyQt4', 'PyQt5', 'wx']                  # matplotlib
@@ -213,7 +222,7 @@ def run(self):
 setup(name=package,
       version=version,
       description=doclines,
-      cmdclass={'build': build},
+      # cmdclass={'build': build},
       options={'build_exe': build_exe_options,
                'bdist_mac': bdist_mac_options,
                'bdist_dmg': bdist_dmg_options,