Xarray v0.12.2 (29 June 2019) breaking metpy No module named 'xarray.core.accessors' #1077
Assignees
Labels
Milestone
Comments
|
Yeah I just saw this pop up on one our of workshop notebooks that I was updating. I'll get a PR up on this shortly. |
|
It looks like this was broken by pydata/xarray#2991, which moved the So, this will probably have to change to a conditional import for xarray <0.12.2? |
|
seems like this is an unwanted API break for xarray? cc'ing @shoyer |
|
Just found the same error when updating some climlab notebooks. Downgrading to xarray 0.12.1 resolved the problem. |
zbruick
added a commit
to zbruick/MetPy
that referenced
this issue
Jul 1, 2019
backend API of xarray v0.12.2. Fixes Unidata#1077.
|
Sorry for the trouble here! We should clarify that everything inside |
zbruick
added a commit
to zbruick/MetPy
that referenced
this issue
Jul 2, 2019
backend API of xarray v0.12.2. Fixes Unidata#1077. Updated import method per @jthielen's suggestion
zbruick
added a commit
to zbruick/MetPy
that referenced
this issue
Jul 2, 2019
backend API of xarray v0.12.2. Fixes Unidata#1077. Updated import method per @jthielen's suggestion
|
@shoyer No worries. Considering I put in the original PR, I should have seen this coming. Is |
dopplershift
added
Area: Xarray
No, it's just a loose convention, e.g., everything in |
shoyer
added a commit
to shoyer/xarray
that referenced
this issue
Jul 2, 2019
|
See pydata/xarray#3075 for the update to the xarray docs |
|
Looks good. Thanks @shoyer . |
shoyer
added a commit
to pydata/xarray
that referenced
this issue
Jul 4, 2019
mgrover1
pushed a commit
to mgrover1/MetPy
that referenced
this issue
Jul 15, 2019
backend API of xarray v0.12.2. Fixes Unidata#1077. Updated import method per @jthielen's suggestion
mgrover1
pushed a commit
to mgrover1/MetPy
that referenced
this issue
Jul 16, 2019
author Ryan May <[email protected]> 1506983621 -0600 committer Max Grover <[email protected]> 1563309011 -0600 parent cf7d532f6e926677ca58cea0c269485db9387a74 author Ryan May <[email protected]> 1506983621 -0600 committer Max Grover <[email protected]> 1563309007 -0600 parent cf7d532f6e926677ca58cea0c269485db9387a74 author Ryan May <[email protected]> 1506983621 -0600 committer Max Grover <[email protected]> 1563309000 -0600 parent cf7d532f6e926677ca58cea0c269485db9387a74 author Ryan May <[email protected]> 1506983621 -0600 committer Max Grover <[email protected]> 1563308996 -0600 parent cf7d532f6e926677ca58cea0c269485db9387a74 author Ryan May <[email protected]> 1506983621 -0600 committer Max Grover <[email protected]> 1563308940 -0600 MNT: Fix python 3-ism that snuck into conf.py Oops. MNT: Fix CONTRIBUTING.md Having it in the root as a link to the copy in docs breaks GitHub displaying it. Instead, keep the copy in the root and link to it in docs/. MNT: Fix malformed record in station test data CSV data had a comma in the remarks column. Remove that so that it parses better. MNT: Update and rework Wind SLP interpolation example Main thrust was to eliminate some missing data that wasn't properly flagged before (and was throwing a warning when run). Go ahead and use pandas to read the data. Also optimize to only read and project the data once. MNT: Update station plot examples a bit Use pandas to read in data. This is primarily so that we're more robust to all the missing values--we were missing some which was evoking warnings from get_wind_components. Also use clip_on as appropriate. MNT: Clean up copyright years Also update the list of authors. Add the Unidata Python Gallery link to the MetPy README Resolves: #581 Add Unidata Python gallery link (and brief description) to Related Projects in index.rst Resolves: #581 MNT: Fix errors from flake8-docstring pydocstyle 2.1.0 broke API, breaking flake8-docstrings. Pinning should fix this for now. MNT: Build examples with matplotlib < 2.1 for now Matplotlib 2.1.0 is causing problems with CartoPy maps. MNT: Fix erroring of builds when doc8 fails Apparently we were not catching this properly, since a bit of lint crept in. MNT: Fix link redirect DOC: Remove trailing whitespace [MNT] Cleanup storm_relative_helicity Add get_layer_heights function for retreiving layers without pressure data Add tests for get_layer_heights [MNT] Simplify storm relative helicity tests Refactor SRH Extend tests for SRH Add refactored version of original test. Fix unit handing during interpolation of segmented hodograph plots. Update tests for proper interpolation behavior. Add metpy logo for few plots. Add for few more examples. MNT: Move recommon dependency This should be in the doc dependencies in setup.py rather than in the Travis config, since it's needed to build the docs any time. ENH: Hide table of contents on main landing page This keeps the TOC in the side bar, but removes it from the landing page, elminating some duplication. MNT: A few minor cleanups for the contributing guide MNT: Fix examples with matplotlib 2.1 linewidth should never have been specified as a string, not sure how that happened. MNT: Silence some matplotlib 2.1 warnings from examples These caught incorrect behavior: mis-spelled/useless kwargs and wrong function calls, so that's good. MNT: Fix style lint from flake8 3.5 It now warns about variables named 'l', since they are easily confused with '1'. Changing all of these did seem to result in better code regardless. MNT: Add Travis build with docs and pre-releases (Fixes #583) This should help us do a better job of testing pre-releases by making sure we also run our examples against them. MNT: Add __repr__ for TextCollection Just use __str__. This overrides the behavior that Text gives with matplotlib 2.1, which was confusing to see in the output. Update the doc tests to reflect this, as well as silence a matplotlib warning. Add import alphabetically. Add import alphabetically. Turn on interpolation in CAPE/CIN shading by default. Update test image for interpolated filling. Increase mpl 1.4 threshold for shading test due to interpolation in 2.1 Update mpl version check to use string comparisons Cleanup precipitable water function. * Correct error in formula documentation. * Add bottom pressure argument. * Make default of top pressure None. Update test to have top of 400 hPa (old default value). Update reference to Salby 1996. Add test for precipitable water with no bounds given. BUG: Fix transformation of barbs with Cartopy (Fixes #598) ENH: Use our wx_code_map in the station plot examples MNT: Fix some warnings with missing data Also fix coding nan cloud coverage. MNT: Install proj4 regardless on Travis Since we're now using CartoPy for testing, we need proj for all builds. MNT: Need to move shapely spec on Travis Also needed to run CartoPy Add test case for issue #593 Sort pressure and heights before getting bounds. Add test for bounds failure, issue #596 Improve handling of floating point comparisons. * Use nanmin/nanmax for any data with nan values * Change comparison logic to our greater/less or close * Change interpolation point addition logic to check if the point is close * Closes #596 [MNT] Update import order Integrate all areas for CAPE/CIN, not only positive or negative. Improve logo positioning [MNT] Fix flake8 import order issues Update returns of figure annotations to be the annotations, not the figure or axis Add specific humidity from mixing ratio calculation Fix spelling error in mixing_ratio_from_specific_humidity docstring. Add test for specific humidity from mixing ratio calculation. Rename Convergence Functions to Divergence This commit is a simple rename of the kinematic functions labeled "convergence" with "divergence" instead, which is what is actually being calculated. It updates the corresponding tests along with that. It also corrects a few small comments where the string "convergence" happened to show up when it should not have. Add Myself to the Authors List Since the contributor's guide said to add my name to the list, I did. Hopefully future contributions will be more substantial that my first simple rename. Fix See Also Reference and Alignment As requested, fixes the changed "See Also" reference to refer to the stretching_deformation function, instead of just deleting it, as well as makes sure the hanging argument stays aligned with the shortened function name. Add tests for dry and moist static energy. Add calculations for dry and moist static energy. MNT: Clean up some docstring lint The short breaks were harmless, but the bad indent caused the docstring to render incorrectly. MNT: Update copyright year in docs MNT: Regenerate sphinx make files sphinx-build now has a make mode that handles all of the options, so we can have much smaller build scripts. Still need a bit more for clean and autogen, since there's more we need due to sphinx-gallery. MNT: Re-enable building docs on AppVeyor BUG: Fix filename pattern for sphinx-gallery This was not actually working correctly, as it's supposed to be a regex, not a glob. It just happened to match against the '/' in Unix paths, where as on windows the path '\' caused escaping of the '*' character, resulting in no scripts matching. MNT: Add pandas to environment file We use this in examples now. Add a Thickness Calculation This commit adds a thickness calculation using an integral form of the hypsometric equation with virtual temperature adjustment. This assumes hydrostatic balance. Also included are two tests of this calculation, one using matching data from other tests and one using a simple dry isothermal layer. Correct Thickness Formula Arguments Based on feedback, corrects the thickness formula to use d(ln(pres)) instead of d(pres)/pres. While these are equivalent in the continuous case, they do not give the same results in the discrete case here. Tests have likewise been updated to account for this change in formula (which should hopefully also resolve the legacy test failure). Updates to Optional Mixing Ratio and Depth Code Based on review feedback, updates the optional mixing ratio to be handled conditionally rather than with an array of zeros (if none is given), as well as conditionally calling get_layer only when a bottom/depth is specified. Adds two additional tests to cover all four scenarios (full layer-moist, subset-moist, full layer-dry, subset-dry). Also adds a citation for the thickness formula. Correction After Rebase with Upstream The recent addition of the static energy calculations unfortunately caused a messy rebase conflict where the new functions got interleaved. I thought I got it all cleaned up, but unfortunately missed this exporter statement. Improve supercell composite and sigtor equation representation. Add tests for scalar values to sigtor and supercell composite. Update sigtor and supercell composite to work with scalar values. Add input unit verification to supercell composite and sigtor Improve function signature Add test for segmented hodograph with differing bounds units. Convert bounds to base units before plotting segmented hodograph. DOC: check presence of patched rtd_theme prior to setting versions Allows docs to build on vanilla sphinx_rtd_theme, see #644 Update gridding_functions.py Changes interpolate function docstring to offer more specific explanation of 'hres' parameter Update gridding_functions.py Further generalizes docstring for hres parameter in interpolate function. Omits meters from default parameter description, as units are dependent on X and Y. Replace "View source" with "Improve this page" Modify the ReadTheDocs 'breadcrumbs.html' template to include this link. The link in the docs will point to the Github source in "edit" mode (on the master branch). If the page is autogenerated, it will point to a new issue instead with the page name included in the title and some instructions in the body. Add a Mixing Ratio from RH Calculation As a part of resolving issue #295, this commit adds a calculation to convert from RH to mixing ratio, given temperature and pressure data. Add the Thickness from RH Calculation As discussed in issue #295, adds in the thickness_hydrostatic_from_relative_humidity function, which is basically a wrapper for thickness_hydrostatic that takes RH as input and converts it to mixing ratio for thickness_hydrostatic to use. Dependent upon mixing_ratio_from_relative_humidity, hence included in same PR. Fix First Lines of Doc Strings in Tests to be Imperative As per discussion in #669, make sure that the first line of the doc string for all the tests are imperative to fit with convention. Update to Take RH as a Unitless Ratio Per discussion, updates the added mixing_ratio_from_relative_humidity and hydrostatic_thickness_from_relative_humidity functions to treat RH as a a unitless ratio between 0 and 1 inclusive. Applying this same update to other functions in thermo.py will happen in a separate PR. Updates mixing_ratio and saturation_mixing_ratio to force a true dimensionless return Updates precipitable water test to use assert_array_almost_equal to account for machine precision Added tests for dimensionless mixing ratio Add sphinx_rtd_theme to development environment DOC: Correct "the the" typos Update to Treat RH as a Unitless Ratio in Existing Functions As discussed in PR #669, for consistency with the rest of the library, relative humidity should be treated as a unitless ratio in [0, 1]. It was noticed that some previous functions in thermo.py treated it as a percentage. This aims to correct that. Fix Percent vs. Unitless Error in isentropic_example.py Added function for relative humidity from temperature and dewpoint Added test for relative humidity from temperature and dewpoint Removed imported constant no longer needed Fixed pep8 styleguide issues Fixed pep8 indentation problems Fixed rest of pep8 errors Added single blank line to satisfy styleguide Adjusted docstring to conform with PEP257 Failed test due to desired value being calculated with slightly different equation than what was used in the test. Updating desired value to conform with equation used in function Updated for use of existing functions with slightly different formula Updated docstrings Satisfy the linter Updated/Fixed Relative Humidity from Dewpoint Test Clean up isentropic interpolation example Changed extend value in colorbar call to 'max' Removed figure number 1 from the figure call Added call to plt.show at the end of script Changed unpacking interpolation results Clean up isentropic interpolation example Initial draft of GEMPAK Table Update RELH based on PR #681. Ensure that appropriate constants and function results are dimensionless (#625) * ensure dimensionless units in constants and mixing ratio / specific humidity calc BUG: Fix loading of custom javascript in docs The fixed relative path is no good. Instead, we can use the method suggested in the templating section of the sphinx docs: the script_files variable for the layout template. Also remove the additional loading of jquery, since that's guaranteed to happen with our theme. BUG: Fix relative paths for doc version handling (Fixes #675) This was not properly finding either the versions.json file nor properly linking to other versions if the user is anywhere but the main page. MNT: Fix title for "Say Thanks" button Clean up isentropic interpolation example Adds calculation for variable dx/dy added tests for calc_dx_dy Make calc_dx_dy take advantage of slicing and not return unknown dx dy values at the boundary. Add tests for 2D inputs and mismatched inputs Rename function to lat_lon_grid_spacing [MNT] Pin Freetype at 2.7 New functions for calculation of geopotential from height and vice versa (#678) Change unit attachment to LCL pressure (#694) Rename v_vorticity to vorticity and h_divergence to divergence Add wrappers for convergence, convergence_vorticity, and v_vorticity. Clean up isentropic interpolation example Changed extend value in colorbar call to 'max' Removed figure number 1 from the figure call Added call to plt.show at the end of script Changed unpacking interpolation results Create CODE_OF_CONDUCT.md Link to code of conduct. Add high contrast functionality to timestamp MNT: Remove hyphen from reference tag It doesn't break Sphinx, but it does break PyCharm. Just get rid of it since it's not necessary. MNT: Update matplotlib import in conftest For some reason pyplot needs to be imported explicitly. Easy enough just to fix. ENH: Make unit versions of at_least1d/atleast_2d handle arrays They were assuming they always got Quantities. ENH: Add functions for derivatives on irregular grids Adds first and second derivatives, as well as gradient and laplacian functions. These work on grids that have uneven spacing between points. BUG: Fix incorrect deprecation The original, actually shipped name was `h_convergence`, not `convergence`. MNT: Elminate divergence_vorticity With the new derivative functions, there's no performance benefit to the combined function. Instead of adding the appropriate named version, just deprecate the old one entirely in favor of the individual vorticity and divergence functions. MNT: Deprecate shearing_stretching_deformation This will no longer give us any performance benefit, so remove it. MNT: Silence deprecation warnings in tests MNT: Fix typos ENH: Replace use of numpy.gradient wrapper Use our own first_derivative and gradient functions. Some tests needed to be updated to account for changes to how values at the edge of the grid are treated. MNT: Disable printing test log output by default ENH: Promote diff wrapper to part of the API MNT: Handle normalize_axis_index for old numpy It's just a nice helper for better error messages, so we can just backport as a noop. MNT: Change default dimension ordering for kinematics functions We already promised that we were going to do this, so we probably should. MNT: Update mailmap MNT: Work around flake8-docstrings problem 1.2.0 doesn't seem to like our code. MNT: Fix travis pinnings Need the latest flake8-docstrings AND pydocstyle. MNT: Update flake8 config to silence pydocstyle problems MNT: Update authors list MNT: Update doctests for NumPy 1.14 (Fixes #687) Since we really only need to run doctests on the latest numpy, just update the text output rather than try to use numpy's support for the "legacy" print mode. Force `mixing_ratio_from_relative_humidity` to return dimensionless Fixes issue #703 where `mixing_ratio_from_relative_humidity` returns percent when RH is supplied in percent. Adds a test to verify returned unit is indeed dimensionless. Clean up sigma interpolation example (#710) Export gradient ENH: Add config for CircleCI Try to get geos installed Install proj4 Fix the container version Add precipitation color table Add Batch of Functions for Brunt-Väisälä Frequency and Period This commit works towards implementing the functions discussed in issue #629: Brunt-Vaisala Frequency Squared, Frequency, and Period. This is based on previous commit 5119f72, but modified to use the new `first_derivative` function and to place the functions in thermo.py (which seems like the more fitting location). The previous `brunt_vaisala_frequency_layer` fucntion has also been removed. Add calls to 'add_timestamp' to examples Changed order of imports for metpy.plots Added 'add_timestamp' in the imports and at the end of the code for the sigma_to_pressure_interpolation.py file Added 'add_timestamp' to the isentropic_example.py file Added 'add_timestamp' to NEXRAD_Level_2_File.py file and change position of Metpy logo to lower left position Added 'add_timestamp' to NEXRAD_Level_3_File.py file and change position of Metpy logo to lower left position MNT: Adjust test tolerance Looks like the newest matplotlib tweaked the end of the lines being drawn, so slightly bump up the tolerance. Deprecate lat_lon_grid_spacing and add signed lat_lon_grid_deltas Port tests from lat_lon_grid_spacing to lat_lon_grid_deltas and parameterize [MNT] Remove landscape from code quality checker links. Add web bulb temperature calculation using Normand's rule. Add test for scalar values to wet_bulb_temperature Generalize for nD arrays Add tests for 1/2D inputs Add reference for wet bulb temperature calculation. Add values from NWS online calculator for reference. format fixups Add Dry Air Density Constant While working on some problems using MetPy, I tried using the density constant, and found that it wasn't actually there, despite being in the documentation! This commit adds it in. Ignore print statements in examples and tutorials. Add pretext and time_format options to add_timestamp. Add timestamp tests for pretext and time_format. Add test images for add_timestamp. ENH: Update Code Climate config First steps to trying to make this thing useful. BUG: Fix missing filename attribute So when we refactored the handling of files, we removed the filename attribute from Level3File. This attribute, it turns out, is used in all of the logging for the class, as well as __str__. More disturbingly, not one test screamed about it being broken. MNT: Disable a few more Code Climate checks Fix dependecies. #734 (#744) Fix dependecies. #734 Add AMS talk (#745) * MNT: Ignore new pytest cache directory * ENH: Add link to 2018 AMS talk Add ageostrophic wind calculation Add tests for ageostrophic wind Fix return types in docstrings of logo functions Add potential temperature to temperature conversion (#746) * Add potential temperature to temperature calculation Add county outline files MNT: Fix Travis testing with CartoPy Need to explicitly install Cython for CartoPy to build. This needs to be a global extra package since we use CartoPy for at least one test. MNT: Fix test with new CartoPy and old Matplotlib There's some kind of interaction here that changes the default gridlines. Fix this by explicitly specifying the gridlines. MNT: Tweak license text and update year Update to 2018. Also pull a fresh copy of the license text from GitHub to hopefully fix the license detection. MNT: Fix some missing copyright years It's 2018 now. MNT: Move some doc links to https to avoid redirects Rename testdata to staticdata Point get_test_data at staticdata folder [MNT] Update path to test data. Add unit conversion capability to plot_barbs. Add tests for barb unit conversion. Disable method length checks in codeclimate. Add wind barb unit conversion to skewt Add tests for skewt wind barb conversion. Bump test tolerance Add saturation equivalent potential temperature (#748) * Add saturation equivalent potential temperature Adds custom grid boundary to interpolate (#741) * Adds custom grid boundary to interpolate [MNT] Fix rendering issues in Montgomery Streamfunction MNT: use `mpcalc` as import alias for `metpy.calc` fixes #756 Add absolute vorticity calculation Add tests for absolute vorticity MNT: Update contributing instructions If we intend to update the AUTHORS.txt ourselves (which is good), we should probably stop telling people to do it themselves in the Contributor's Guide. Update state and coastline features for Cartopy 0.16 Add baroclinic and barotropic potential vorticity calculations. Add tests for baroclinic and barotropic potential vorticity. [MNT] Update maximum argument count to 10 Force coriolis parameter output to 1/s Test coriolis units are 1/s MNT: Fix up a couple of barb tests These fail due to changing freetype because they still have text, even though remove_text=True. This is from pytest-mpl, and really matplotlib, where it doesn't turn off axis labels. Let's turn them off manually since these tests are not meant to test axis labelling (due to units). This should allow these tests to be more reliable. MNT: Remove pinning to freetype 2.7 Replace with using testing packages for matplotlib from conda-forge. MNT: Slightly adjust some thresholds for new freetype Image diffs don't show any difference, so likely just a pixel or two. Good enough just to bump threshold. MNT: Install recommonmark from conda-forge Elminates the need to use pip separately. MNT: Blacklist problematic flake8-builtins version It doesn't seem to like loops where tuples are pulled out. Four panel example edit Added missing whitespaces as indicated by Stickler-ci. Modified subplot_kw parameter value to the proper dictionary format. Make any user kwargs to lat_lon_grid_deltas override the default geod ellps. Test user kwarg override on lat_lon_grid_deltas MNT: Remove unneeded blocks in turbulence tests Was trying to just remove use of "keys()", but it turns out that the blocks aren't necessary. MNT: Remove unnecessary use of keys(). Unless it adds clarity, it's best to just rely on the fact that iteration on a dictionary (including conversion to list) uses the keys anyway. Add Exner Function As discussed in #757, this commit adds the Exner function (as a function of pressure), and refactors `potential_temperature` and `temperature_from_potential_temperature` to use it. Closes #757. Update Equation and Add Reference Pressure Optional Argument Rename sample data to more meaningful names. Remove upper air functionality and add data method for testing. Change testing for newer test data retreival. Remove unused testing file ENH: Add Stack Overflow information (Fixes #782) Updates all of the support, contact, etc. to include information about posting on Stack Overflow. MNT: Use logging.basicConfig() This takes care of setting a default handler, formatter, and allows us to set the *default* logging level. The way we were doing it would make the module's level setting override any previous setting. This way allows the user to more easily override. MNT: Make NEXRAD decode WMO with ascii It doesn't make any sense to use utf8. MNT: Add some debug logging to NEXRAD This gives us some information about the search for the header and footer bytes. MNT: Improve logging in NEXRAD level 3 Replace an assert (about bytes remaining) with a warning message instead, since in some cases parsing can happen just fine. Also replace a commented out print statement with a debug log message. BUG: Fix get_layer_heights modifying heights in-place (Fixes #789) MNT: Remove unneeded comment from conf.py ENH: Add an issue template Add inertial advective wind calculation. Add test for inertial advective wind. MNT: Fix some doc http->https redirects MNT: Update codacy link Removes a redirect. MNT: Update link to WMO#8 report MNT: Remove unneeded use of matplotlib pyplot This started as cleaning up a warning from matplotlib that was using plt.axes() to return an existing axes. I ended up removing all uses of pyplot that were readily replaced by a call to the appropriate figure or axes method. Add utility to attach units to pandas dataframes. Add tests for pandas unit functionality. Add pandas to testing requirements. MNT: Add python_requires to setup.py This will be more important once we drop 2.7. ENH: Add Python 2.7 plans to docs (Fixes #803) ENH: Add information about semantic versioning (Fixes #802) MNT: Fix travis doc build With pip 10, we need an extra flag to make it blindly upgrade everything, which is what we want for our CI setup. Add wind vector plotting to the hodograph. Add test of hodograph vector plotting. Add feels like temperature. Add test for feels like temperature. Add reference PDF to NWS 10-201 Refactor colored line segment hodograph Add Q-vector and Static Stability Calculations Implements the Q-vector and static stability calculations as discussed in #661, based of formulas from Bluestein1992. Closes #661. Add new metpy/calc/tools parse_angle and its subfunction _abbrieviate_direction parse_angle allows users to easily convert directional text into a degree/radian value such as N becomes 0 deg, east becomes 90 deg, south_east becomes (135 deg) etc Adds calculation for dewpoint from specific humidity, temperature, and pressure Adds test for dewpoint from specific humidity Modify calculation of LFC and EL for arbitrary parcel profile Added tests for ML LFC and EL calculations Clarify get_layer doumentation. Clarify bulk shear documentation. Update development installation instructions. Make sure LFC is not below LCL. Add test for LFC not below LCL Pass parcel profile through cape_cin calculation. Add test for custom profile in cape_cin Removes grid size warning from generate_grid (#811) * Removes grid size warning from generate_grid This removes the grid size warning for grids with dimensions less than 10000 units. The message implies the use of meters, which means it could appear for users using other grid dimensions (like degrees). Make default_units work with lists or tuples and return dimensionless as the default unit. Add test for hodograph with united component range. Fails currently due to mpl bug Don't allow flake builtins 1.4.0 Add parse_angle() example to calculations page. Fixes issue that would error out parcel_profile if the profile did not reach the LCL. #827 Add test for parcel profile when the profile does not reach the LCL. [MNT] Pin at not netCDF4 1.4 for examples. Rename delete_masked_points to _delete_masked_points Update tests and skewt for renaming to _delete_masked_points. Import interp from calc instead of calc.tools Allow apparent temperature to take and return scalars. Add test for apparent temperature working with scalars. Add tests for gradient Add detail to gradient documentation. Add example of using gradient. Remove requirements text ENH: Support Unidata GINI NEXRAD composites (Fixes #795) Adds support for reading PNG compressed data, as well as fudging around the fact that year values are less than 100, but should be years since 1900. ENH: Enhance GINI tests Add a test of actual values in the data array, as well as the values of datetime in the netCDF interface. ENH: Add XArray support to GINI This makes GINIFile work as an XArray data store, so that it can be passed directly to xarray.open_dataset(). Port the example over to use it. MNT: Deprecate CDM and GINIFile.to_dataset() This deprecates the CDM layer, which is only used by the netCDF-like interface for GINIFile, in favor of the XArray support. MNT: Add a configuration for LGTM Reclassify/ignore a few files. MNT: Work around docstring formatting in xarray The formatting for the load docstring is incorrect in xarray, which causes sphinx to warn when building our own docs. Inheritance is wonderful. Add new functionality to GEMPAK table. Export wx_symbols to plots level. Update examples for wx_symbols being exported to plots level. Update testing for wx_symbols being exported to plots level. Export colortable registry and update examples. Export gridding functions. Rename colortables directory to colortable_files. ENH: Add simple registry implementation This adds a class with a register decorator that can be used to register callables under a string. ENH: Add framework for converting CF attributes to CartoPy ENH: Add xarray accessors to simplify getting projection ENH: Add example of using xarray projection info MNT: Bump up complexity threshold for CodeClimate ENH: Use xarray support to simplify GINI example MNT: Fix import on Python 2.7 Default behavior of Python 2.7 gets confused by our having our own xarray module and trying to import the main library. Some future imports solve this... MNT: Move pyproj import Right now metpy.calc won't import without having pyproj installed, which is a bit more of a pain in the but than I want for a hard dependency right now. MNT: Add xarray as a hard dependency MNT: Update codecov config Looks like we were missing some test lines MNT: Fix README link to install guide BUG: Fix interp with unit-ed masked arrays Because masked arrays and pint don't necessarily play nicely together (see hgrecco/pint#633), we need to make sure that the Quantity is on the left side of the multiply so that it controls the result--otherwise weird unit errors can result. MNT: Simplify isentropic interpolation code Simplify the interpolation of other variables, taking advantage of the fact that the linear interpolation can handle multiple variables at a time. MNT: Remove use of assert_almost_equal with .shape Should just be doing an explict assert arr.shape == truth. MNT: Simplify some of isentropic_interpolation This just eliminates some unneeded named temporaries, ensuring that their memory is released sooner. BUG: Fix isentropic_interpolation for masked data (Fixes #769) If an entire column of data was masked, the sorting would end up trying to interpolate off the bounds of the array. MNT: Ensure we handle columns of nans in isentropic_interpolation ENH: Add utility function to search for indices around a value ENH: Use linear search in isentropic_interpolation (Fixes #798) Previously, this was using numpy.searchsorted(), which is only strictly correct if the points are sorted/monotonic increasing. Theta can easily violate this condition. Instead, switch to a linear search, which allows also controlling which direction to search. This also seems to be faster since we're no longer using apply_along_axis, but can rely on indexing operations. Added a function to calculate critical angle as well as a reference and a test for it. (#843) MNT: Move location of xarray helpers This moves to a top-level module, which is imported by the top-level __init__. This causes any import of metpy to activate the xarray accessors, which simplifies using them in a few places. ENH: Add xarray preprocessing decorator This currently handles converting xarray DataArray's passed to functions into pint Quantity instances. ENH: Apply xarray decorator to calculations This changes all exported functions to have DataArray instances passed to them automatically converted to pint Quantity instances. I added one test to ensure this works; ideally all calculations would have such tests, but that seems overly repetitive. MNT: Fix close test functions with mixed units and not This was triggering a warning about numpy.float64 and a lack of a _units attribute. The root cause was comparing a value with units with one that had them. One fix is to make sure neither have units in parcel profile (since both values have been coerced already to have the same). The other change is to tweak _greater_or_close and _less_or_close to use > and < rather than numpy.greater and numpy.less, so that an error results if a Quantity and an array are compared. MNT: Use strftime as a method in add_timestamp Rather than hardcoding the use of the strftime function from the datetime module, use it as a method. This way a passed in object can have their own implementation of strftime, which paves the way to using the '.dt' accessor from xarray. ENH: Monkey patch in strftime for xarray This is going to be a PR, but without it things are far too ugly, so we'll monkey patch one in for now. This requires moving to xarray 0.9.6 which is when the .dt accessor was added. MNT: Clean up isentropic example This passes in CartoPy transforms to plotting functions instead of manually tranforming everything. It also uses xarray instead of netcdf4 to handle data parsing. MNT: Fix up sigma interpolation example This now passes CartoPy projections as transforms rather than manually projecting. It also works around the masked array and pint issue by hard-coding units rather than parsing them. We cannot use xarray for this right now because there are problems with how the file does lat/lons. MNT: Update Four Panel Map example Moves to use xarray. MNT: Remove pinning netcdf4 We've addressed all of the issues in the examples. Optimize parse_angle (#850) Optimized parse_angle() for long list of directional strings. BUG: Fix mask propagation in gradient (Fixes #642) The calculation propagates masked values fine, in that any calculation involving one more more masked values ends up as masked (similar to NaN propagation). The problem was using numpy's concatenate, which stripped masks. Use np.ma.concatentate instead, dropping to a regular array if there are no masked values. MNT: Extend Travis build time. MNT: Clean up CircleCI build Reduce log output so we can figure out what's going on, and fix the build. MNT: Bump CartoPy requirement for tests We need some of the newer projection options. MNT: Update AUTHORS Also update the .mailmap for that. MNT: Fix some non-standard naming Capitalized names are reserved for classes. Picked up by pep8-naming 0.7. Update thermo.py Fix docstring for wet bulb temperature function. Drop 3.4 from the .travis.yml matrix, setpy.py, and README Change x to coordinates in gradient and laplacian interface Fixes #842 by replacing the x keyword argument in _process_gradient_args with coordinates, and including x as a deprecated option. Also adds tests to verify that use of x as a keyword argument raises the deprecation warning. Additional changes to drop python 3.4 Rename wind functions to remove get_ Renames: - `get_wind_speed` -> `wind_speed` - `get_wind_dir` -> `wind_direction` - `get_wind_components` -> `wind_components` Also, deprecates the old functions, and updates the usage of these functions in the examples/tutorials. Fixes #865. Specify handling for calm and north winds This commit modifies `wind_direction` so that calm winds return 0 and north winds return 360, according to the meteorological convention. Also forces `wind_direction` to return in degrees (some past results had it returning as dimensionless). Fixes #794. Add conversions between omega and w under hydrostatic conditions This commit adds two basic equations that assume hydrostatic balence in order to convert between omega (Dp/Dt) and w (Dz/Dt). Initial move of gridding to interpolate This (rather substantial) commit refactors the current gridding package to become the interpolate package. This includes making gridding exist as a wrapper module (that deprecation warns whenever used), pulling out some of the internals of the old gridding functions to work on a point basis, and renaming/reorganizing many of the functions to make them fit better within a more general interpolation package. This does not yet include moving over interpolation functionality that currently resides in calc. Move interpolation functionality in calc to interpolate This commit moves the interpolation functionality in metpy.calc.tools (all forms of interpolation along a single axis) to a new module in interpolate. This also contains a series of changes to fix syntax errors and other mistakes based on CI and code review Build in identification of spatiotemporal coordinates Adds coordinate mapping utilities to the Dataset accessor, and uses them in `parse_cf` to relabel the `axis` attributes appropriately. Then, adds in methods to the DataArray accessor so that the spatiotemporal coordinates can be easily accessed. Add xarray tutorial This commit adds a new tutorial to demonstrate the extent of MetPy's xarray integration, and how to perform basic operations. This tutorial should be kept up to date as integration with xarray improves. Add tests This commit adds tests for systematic coordinate identification. Clean up DataArray equality tests Instead of checking equality of values and attributes, use the `indentical` method, which is much cleaner. Let dimension coordinates take precedence In working with the GOES16 data, the coordinate identification in parse_cf fails because there are two coordinates with "projection_x_coordinate" as their standard name. But, only one of them is actually a dimension coordinate, and we would want that one to be selected. This commit implements and tests for this. Expose cartopy globe in CFProjection and xarray accessor This commit changes the protected method _make_cartopy_globe in CFProjection into a property, and exposes that property in our xarray DataArray accessor as well. permit usage of temperature in Kelvin for heat_index calc "Improve page" links for examples and tutorials Adds the sphinx-gallery configuration to the template so we can link to the Github edit page for the gallery .py files. The template also identifies the index.html files and directs them to the README.txt files. fix apparent_temperature when only windchill is calculated workaround numpy=1.10.0 issue? Modify lat_lon_grid_deltas to allow for >2D arrays In the course of working on xarray examples in metpy, the current implementation of lat_lon_grid_deltas only allowing 1D or 2D input was a hinderance. This commit makes a simple change to allow >2D arrays, so long as the trailing dimensions are y, x. Fix remnants from lat_lon_grid_spacing deprecation The deprecation warnings for lat_lon_grid_spacing were not being tested/handled, and some of the comments for lat_lon_grid_deltas still referenced lat_lon_grid_spacing instead. This commit fixes those. Remove deprecated kinematics functionality from 0.7 This commit removes the deprecated kinematics functions slated for removal in 0.9, as well as stopping the warning for using the default dim_order, as outlined in #690. Allow extra leading dimensions in gradient and kinematics calculations Based on issues with some kinematics calculations generalizing to data that has extra leading dimensions (such as time or vertical as commonly encountered when working with xarray examples), this commit modifies the default derivative axes in kinematics calculations to be based on trailing axes (-2 and -1) rather than leading (0 and 1), to follow our convention (and CF option) of (..., Y, X) dimension ordering. Numerous tests are added to make sure that the affected calculations work with extra dimensions (tested on 4-D data). Additionally, the gradient function previously only supported the full gradient of an array according to its description, but a useful case is just obtaining the horizontal gradient along all levels or times. Handling of the "axes" kwarg and was added to the _process_gradient_args function to resolve this, so that "axes" becomes a kwarg of gradient and laplacian. It must be less than or equal to in length compared to the coordinates/deltas. Add netcdf4 to testing requirements MNT: Replace assert with log message in NEXRAD We're just checking a size here, so be less aggressive and warn rather than error out. It's completely possible to be off a little in size and still parse through the file fine. MNT: Update NEXRAD for new message size With latest RDA Build 18.0, the size of the status message has increased with some spare bytes. Just include the spares in the size check for now. Add xarray to intersphinx mapping In conf.py, adds xarray to intersphinx mapping so that we can easily link from our documentation to xarray's. Add API doc module override, and use it with metpy.calc This commit modifies the module template to allow it to be overridden by a custom module page, implements a make checker for that override, and implements this to better organize the docs for metpy.calc. Also updates the Travis and AppVeyor doc builds to use the overridecheck option. Remove removed functions from doc page override After #889 and #882 were merged, there was a conflict where the override doc page created in #889 had functions that were removed in #882. This commit removes those and should restore the doc builds to working order. Also cleans up a line from code review in #889. Fix NEP 14 broken link Bump minimum xarray to v0.10.7, numpy to v1.11.0 Add cross section function and supporting calculations This commit adds a `cross_section` function to `interpolate`. It allows interpolation to a cross-sectional slice through gridded data along a geodesic path between two lat/lon points. Associated changes are also made to some xarray functionality. Additionally, this adds a number of calculations to the calc subpackage in support of cross section functionality: - unit vectors - tangential and normal components - absolute momentum (pseudoangular momentum) NARR Example CF Parsing While I was working through examples, I noticed that the parse_cf function, when applied to the full dataset of the NARR example, would fail...the x and y coordinates would double in size, and many NaN's appeared. This was because the lon and lat variables lacked a grid_mapping attribute, and so, their x and y coordinates did not get scaled from km to m like the other variables that did. Thus, they did not line up when xarray tried to combine them back together in a non-unit-aware fashion. This commit fixes this by changing when _fixup_coords is called, so that all x/y coordinates get scaled accordingly. Fixes based on CI, code coverage analysis, and review Update docstring for vorticity Documentation for the vorticity function hadn't been updated to reflect the fact that the function can handle varying grid spacings. This commit updates the documentation accordingly. Add NARR Cross Section Example This commit adds a basic example of cross section analysis using the NARR example dataset. It plots potential temperature, relative humidity, and projected winds, with an inset 500 hPa map with the cross section path depicted. Make derivative calculations work with xr.DataArray This commit modifies the current derivative calculations (first_derivative, second_derivative, gradient, and laplacian) to work with DataArrays themselves, rather than by conversion to (and output as) pint Quantities. In order to support lat/lon non-uniform coordinates, the lat_lon_grid_deltas function has moved from kinematics to tools to be alongside the derivative calculations, and a helper function to make it work nicely with a DataArray was added. Some small fixes are also made to the xarray tutorial along the way. Add additional units to the pint registry Units such as 'degrees_north' are part of CF standard for latitude and longitude coordinates, and unless they are renamed, the xarray to pint quantity conversion breaks with these units. This commit defines these as new units in our registry, with aliases/abbriviations as needed. Clean up exisiting workarounds In order to work with data that had 'degrees_north' (or similar), workarounds were needed in the past. With the unit update, these are no longer needed. Part of the remaining discussion involved how exactly to handle 'gpm' non-standard units. A comment referencing the relevant issue has been added to the tutorial section on the existing workaround, since the issue needs further implementation discussion. Fix non-tuple NumPy indexing in MetPy Since NumPy 1.15 came out, we have been getting pages and pages of warnings about our use of non-tuple based indexing. This commit converts all of our previous list-based indexing to tuple-based indexing. This doesn't yet clear out all the warnings (some of them are upstream in Pint and scipy), but it does drop the number down from 600-something to 10-something. Add GWFS function from GEMPAK This commit wraps the gaussian_filter function from SciPy so that it emulates the GWFS function from GEMPAK as closely as possible. The key difference is that GEMPAK adds the leftover weights from outside the averaging window to the center point, but there is no option for gaussian_filter to do this. Left unfinished is restoring metadata to the result. For instance, an xarray.DataArray provided as input becomes a lowly numpy.ndarray on output. Add tests and docs for gwfs() This commit adds a few basic tests for the metpy.calc.gwfs function, and also updates the GEMPAK compatibility table. I'm not sure how to add new documentation to the MetPy API section. Fix whitespace issues re pep8 Extend functionality of gwfs and update docs gwfs is extended to preserve units and ensure that only horizontal directions are smoothed (assuming last two axes are horizontal). Documentation is updated so that this function appears. Rename gwfs() to smooth_gaussian() Update GEMPAK Conversion Guide One place was overlooked when changing gwfs() to smooth_gaussian(). This corrects the oversight. Add link to SciPy 2018 talk Add links for scipy 2018 poster Respell doc page names for clarity. MNT: List matplotlib >= 2.2 in requirements for examples (Fixes #867) This may help people to not run into problems with older matplotlibs. MNT: Add some implemented functions to GEMPAK table (Fixes #857) These got missed for 0.8 MNT: Update some redirecting links BUG: Increase default x clipping radius for SkewT The previous default of 0.08 was just a little too tight. Add gempak_color() and related tests (#916) * Add convert_gempak_color() and related tests This commit adds a new convert_gempak_color function to metpy.plots to facilitate conversion between GEMPAK color numbers and corresponding Matplotlib colors. All known GEMPAK quirks are accomodated (such as 101 representing the background color). Add unit tutorial MNT: Add warning filter for numpy messages The "dtype size has changed" warning is harmless enough, and is really becoming verbose of late, at least in terms of the number of places it pops up. Silence that for now here. MNT: Add additional assert to check If there are no log messages, before this change you just get an IndexError. Now you get a failed assert and it's quicker to see what's going on. BUG: Don't use basicConfig for logging (Fixes #866) What basicConfig does is install a root handler if none is already installed. This breaks applications that also try to use it to set a different level. It's completely unnecessary on Python 3, since by default there is a handler that logs anything WARNING or higher to stderr, which is really what we want. This isn't the case on Python 2, so what we can do in that case is put a StreamHandler (which defaults to stderr) for all of metpy (all submodules will defer to it by default). WARNING is already the default logging level anyway. Add gpm unit as an alias to meters Use Pooch to download sample data Add it to `metpy/cbook.py` and include the dependency on setup.py. Create the registry file using `pooch.make_registry` and stored it in the `metpy` package. Remove 32-bit windows builds A Few Doc Fixes This commit contains a few documentation fixes: - Fix the description of the cross_section/interpolate_to_slice functions to reference xarray's interpolation, not metpy's - Clear up some usage details with Dataset.metpy.parse_cf() (see Issue #909) - Update dx/dy handling wording (as in #901) for all applicable functions in kinematics - Add notes to kinematics functions about handling of >2 dimensions (similar what ageostrophic_wind had in parameter descriptions) - Update the GEMPAK table for closed issues BUG: Fix SRH unexpectedly returning masked values BUG: Fix edge cases in profiles calculations (Fixes #902) By not including the LCL in the profiles used in calculations, some odd corner cases result where, e.g., the intersection points with the ambient profile change. It also has an impact in cape/cin calculations. Refactor profile handling to explicitly work with a profile that includes the LCL and expose this as an API function. MNT: Tweak test profile With changes to LFC calculation to use a profile that includes the LCL, the old data does not match the test case of an uncapped sounding. Tweak a temperature to get back to that case, since that's what we want to test here. MNT: Adjust CIN value in test Now that sfc cape/cin uses a profile that includes the LCL, this increases the amount of negative area in the profile, so we need to adjust the value here. Add rest of shapefile associated data Add county features Add example of using counties with different resolutions. Remove cartopy dependence in cross-sections Add v to beginning of version number for pooch search path. Add instructions for adding a file to the cache Adds US State borders at same scales as counties refactor class to MetPyMapFeature Access constants via module (#930) Access all constants via module MNT: Only pin to minor version of pooch Rather than pinning to exact version including bug fix. Turns out we were already doing this for conda. MNT: Fixup tests for matplotlib 3.0 Tests were broken by matplotlib 3.0 due to our tests' use of old matplotlib style parameters. Main code was note affected. MNT: Eliminate some unnecessary uses of pyplot This specifically fixes a problem between matplotlib 3.0 and CartoPy. MNT: Add workaround for CartoPy/Matplotlib incompatibility Replace deprecated interp with interpolate_1d Add decorator to ignore metpy deprecation warnings. Turn metpy deprecation warnings into errors with pytest. Decorate tests of deprecated functionality. Remove edgecolor from example address pytest fixture warnings emitted by pytest==3.8.1 MNT: Remove unused warnings.simplefilter Really shouldn't be filtering out our own deprecation warnings anyways--this has been made an explicit error unless otherwise flagged. MNT: Replace use of warnings.simplefilter with pytest pytest has the ability to mark individual tests to filter, so that's better anyway. MNT: Change wrapping around binary operators Recommended by pep8 for new code, and I kinda prefer it, so went ahead and adjusted. MNT: Remove use of numpy.asscalar Numpy 1.16 will deprecate numpy.asscalar. remove inaccessible/noop code in metpy.deprecation increase coverage of metpy.interpolate.points increase coverage of metpy.units MNT: Make io module docstring conform to standards Picked up by pydocstyle 3.0. MNT: Fix up copyright year MNT: Silence numpy warning about using a tuple The warning was blaming Pint, but it was actually coming from this code. MNT: Clean up code for nexrad spec parser flake8 is picking up this file now, might as well make it compliant. MNT: Correctly escape regular expression groups Can't use raw strings since we need \n and \r, so need to use \\ for the regular expression groups. MNT: Remove hacks for Matplotlib 3.0.0/CartoPy Matplotlib 3.0.1 has been released. Add Skew-T axes at arbitrary locations (#959) Add support for drawing SkewT on an axes at an arbitrary location. Adds n point smoothing function Update docs for dropping 3.5 and adding 3.7 Add 3.7 to CI builds Update for dropping 3.5 and adding 3.7 Add sudo required Use xenial [MNT] Remake test image with cartopy 0.17 [MNT] Use https links Bump to mpl 1.5.1 for cartopy Bump test tolerances for mpl 2.0 Remove old mpl patches. Update PV calc for isobaric Fix interpolation tests for SciPy 1.2 Scipy 1.2 fixed a bug in 2D cubic interpolation, so we need to adjust the values in the sample files. We can avoid shipping mulitple versions of the static data by skipping the tests on older Scipy versions. Using log.warning insted of warnings.warn Add test to verify warning is logged Add a reference pressure to moist_lapse and dry_lapse Clean up empty array creation. relax pooch requirement to >=0.1 closes #977 Pin pooch at <0.3 Clarify wind_components docstring and close #969 fixes issue #960 and add test function Reattach grid_mapping after cross-section calculations. Add unit- and axis-aware selection to the metpy accessors This commit adds the .loc indexer and .sel method to both the DataArray and Dataset metpy accessors to allow for unit-aware indexing/selection. For DataArrays (for which axis types are uniquely determined), this also allows selection via the axis name ('time', 'vertical', etc.) instead of the coordinate name. Also, adds this new unit-aware selection to the xarray tutorial. Change axis-tracking attribute from axis to _metpy_axis Previously, any existing axis attribute was removed to enforce uniqueness of any given axis, but this modified the user's data. Now, the axis is tracked in _metpy_axis to avoid collisions. [MNT] Update requirements Add link to install guide. Make necessary coordinate information for cross sections clearer As discussed in Unidata/MetPy#949, the error message from the cross section function could be made clearer in the case where coordinate information is missing. This catches the AttributeError that was raised and instead raises a more informative error. The docstrings have also been updated. ENH: Implement simplified plotting interface Adds a declarative interface for creating plots, similar in nature to GEMPAK. DOC: Add examples for simplified plotting ENH: Add GEMPAK area definitions ENH: Add tests for simplified plotting interface ENH: Add traitlets to dependencies ENH: Rename maps to layers Seems a bit clearer. [MNT] Pin pint <= 0.8.1 Mark noqa on case in metpy deprecations. Switch from recommonmark to m2r [MNT] Bump to numpy 1.12 [MNT] Pin flake8-comprehensions for 2.7 builds MNT: Update mailmap ENH: Add roadmap (fixes #941) MNT: Disallow pyproj 2.0.0 This has completely busted us. updates docs for latitude units Adds more combinations for wx symbols Adds a function to linearly interpolate to level MNT: Update doc8 configuration Sphinx-gallery changed their output and it completely buggers doc8. Probably shouldn't be running doc8 on generated files anyway. This does eliminate checking some of our index.rst files, but it's a small price to pay to have things working again. MNT: Update some redirecting URLs MNT: Fix broken link to Miniconda Fix rst style links in the Contributing Guide Replace some rst style links with markdown syntax. Use https for PEP8 link Fix LFC calculation when positive area exists below LCL Soundings can exist where there is positive area only below the LCL. This commonly occurs when there is a superadiabatic lapse rate near the ground. Other soundings may have positive area all the way from the surface to an EL above the LCL. The current code conflated these two situations, assuming that any positive area in the sounding must mean the second case is at play. This commit makes sure that any positive area that is found is above the LCL. If so, the LFC is set to the LCL as before, but if not, there is no LFC (it is set np.nan). Changes suggested by @dopplershift are incorporated. Add test for LFC fix Fix crash with mixed-layer LFC calculation This commit fixes a crash when computing the LFC of a mixed-layer parcel and adds a corresponding test for this situation. Fix incorrect results with LFC calculation Currently, some of the logic in the lfc() function assumes a surface parcel is being considered, but that is no longer necessarily the case. This commit addresses the two places where a surface parcel has been assumed. First, when looking for intersections, the code assumed the first point of the parcel trajectory had an identical temperature to the environment. This is only true if a surface-based parcel is being used. Second, the LCL of the parcel is computed using surface values. This, too, is not appropriate if the parcel is not a surface parcel. The lfc_ml2 testcase was giving wrong results because of the second problem. This commit addresses both problems and corrects the associated testcase. Add additional test for latest LFC fix This test gives 990 hPa for the LFC (the LCL value in this case) if the fix to the find_intersections() call is not present. Use NumPy's isclose() so that Python 2.7 is supported Update Advanced Sounding doc to explain why 0 index is chosen Add line break Update potential_vorticity_baroclinic() docstring Fixes #1035 MNT: fix flake8 errors fix doc issue remove link to ejssm article add functionality to declarative plotting Fix Scipy 1.3 not-C-contiguous error As pointed out by @akrherz, using `np.vstack` followed by transposition causes problems (the array not being C-contiguous, which scipy needs). This stacks the points in a safer way. Fixes #1043. migrate mpl.cbook.iterable to numpy.iterable Due to lack of support for the '%' as a unit within pint, this update converts '%' units to 'percent' units within the MetPy xarray accessor. Fixes #1038. Docstrings improved for thermo calculations requiring pressure arrays to be from high to low pressure. Update to test_mpl.py to update deprecated function Update EL calculation to prevent an EL existing below a LCL. Completes the fix to #1003. first pass at better docs for declarative Make parse_cf() not keep attributes in parsed dataset Rename xarray accessor classes and move check_axis outside of accessor Refactor xarray coordinate parsing to be on-demand on the DataArray Update xarray tutorial for on-demand coordinate parsing Improves documtation for interpolate_to_grid to specify that linear distance or degrees can be the units of the input coordinates (and resultantly, the resolution of the grid). Fixes #1060. Remove lat_lon_grid_spacing and the cdm module and related components Add warning if lat/lon is assumed for projection Fix docstring for vertical velocity functions to render equations This update to the lfc() function allows for multiple LFCs to be identified within a sounding. This fixes #992 and begins to address Add in altimeter_to_station_pressure and altimeter_to_sea_level_pressure Updated tests to include METAR calculations Removed changes Removed all additional changes Update metpy/calc/basic.py Co-Authored-By: Zach Bruick <[email protected]> Update metpy/calc/basic.py Co-Authored-By: Zach Bruick <[email protected]> Update metpy/calc/basic.py Co-Authored-By: Zach Bruick <[email protected]> Update metpy/calc/basic.py Co-Authored-By: Zach Bruick <[email protected]> Update metpy/calc/basic.py Co-Authored-By: Zach Bruick <[email protected]> Update metpy/calc/basic.py Co-Authored-By: Zach Bruick <[email protected]> Update metpy/calc/basic.py Co-Authored-By: Zach Bruick <[email protected]> Update metpy/calc/basic.py Co-Authored-By: Zach Bruick <[email protected]> Removed whitespace Corrected sea-level pressure Remove conversion within the function since it supports units, update documentation removed check units lines Consolidate standard atmosphere variables removed trailing whitespace Changed incorrect variable, added documentation Fixed indentation error Fixed units issue with station pressure calculation Fixed incorrect value for gamma Added additional placeholders Adds a conditional import of the DatetimeAccessor due to a change in the backend API of xarray v0.12.2. Fixes #1077. Updated import method per @jthielen's suggestion Added note that x-axis can be any variable, due to feedback received from a user BUG: Fixup geopotential<->height calculation (Fixes #1075) Previous implementations was a naive implementation that directly translated formulas from source material. It included a lot of division and subtraction, resulting in catastrophic cancellation and severe loss of floating point precision. New formulation is just an algebraic reformulation that avoids these problems. Vendor functionality from xarray.core Add tests for otherwised unused portions of xarray utilities Add formulas to geopotential functions in docs and correct reference MNT: Update ticks for SkewT (Fixes #987) This updates the implementation for ticks in SkewT based on changes in matplotlib. This greatly cleans this up, and for some reason, is necessary with matplotlib 3.1. MNT: Update some skewT image test thresholds This gets the test suite passing with Matplotlib 3.1. Not sure what changed, but a deep look at the differences only reveals some subtle changes to just how black the individual tick marks are. Added 3.7-dev to the matrix Updated year in LICENSE Replace jrleeman with zbruick as a codeowner Updated contributors in AUTHORS.txt and .mailmap allow any pooch >= 0.1 to be used Update install guide for new pooch requirement Fix pint 0.9 errors from units.wraps and iterable Work around matplotlib issue with pint 0.9 Pint 0.9 fails to setup matplotlib on Python 2.7. We can work around this by catching the resulting ImportError and falling back to our internal matplotlib unit code. Test issues with pull request More precise decimal Seperated gamma value into individual calculations Remove space between header in new functions Add in additional line Remove extra line Updated documentation Updated docs Updated documentation update test Removed files Add in indentation Added note that x-axis can be any variable, due to feedback received from a user Vendor functionality from xarray.core Add tests for otherwised unused portions of xarray utilities Add formulas to geopotential functions in docs and correct reference MNT: Update some skewT image test thresholds This gets the test suite passing with Matplotlib 3.1. Not sure what changed, but a deep look at the differences only reveals some subtle changes to just how black the individual tick marks are. Updated year in LICENSE Replace jrleeman with zbruick as a codeowner Updated contributors in AUTHORS.txt and .mailmap allow any pooch >= 0.1 to be used Update install guide for new pooch requirement Fix pint 0.9 errors from units.wraps and iterable Work around matplotlib issue with pint 0.9 Pint 0.9 fails to setup matplotlib on Python 2.7. We can work around this by catching the resulting ImportError and falling back to our internal matplotlib unit code. Add in indentation Add in indentation Edit indentation Add in indentation Edit indentation Remove space Edit indentation Edited indentation Edit indentation Edit equation documentation Edit equations Update documentation Fixed equation Add in extra space Realigned Equations Fixed Indentation Edited equations Add in extra space Removed blank line Remove indent Update documentation Added in underscore for documentation Updated references Edit Documentation
ahuang11
pushed a commit
to ahuang11/MetPy
that referenced
this issue
Aug 30, 2019
backend API of xarray v0.12.2. Fixes Unidata#1077. Updated import method per @jthielen's suggestion
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
I just created a new environment and installed the latest versions of xarray and metpy. The latest version of xarray seems to break metpy and when importing metpy I got this error:
xarray version: 0.12.2 py_0 conda-forge
metpy version: 0.10.0 py37_1001 conda-forge
Rolling back to xarray v0.12.1 fixes this issue.
The text was updated successfully, but these errors were encountered: