Traceback (most recent call last):
  File "cal_sounding_data.py", line 360, in <module>
    newset_dic[k] = specStationSounding(k,dataset_dic)
  File "cal_sounding_data.py", line 284, in specStationSounding
    result_dic = soundingCalculation(num,p,t,td,ws,wd)
  File "cal_sounding_data.py", line 170, in soundingCalculation
    lfc_p, lfc_T = mpcalc.lfc(P,T,Td)
  File "/anaconda2/envs/py3/lib/python3.6/site-packages/metpy/xarray.py", line 138, in wrapper
    return func(*args, **kwargs)
  File "/anaconda2/envs/py3/lib/python3.6/site-packages/metpy/units.py", line 290, in wrapper
    return func(*args, **kwargs)
  File "/anaconda2/envs/py3/lib/python3.6/site-packages/metpy/calc/thermo.py", line 378, in lfc
    return x[0], y[0]
  File "/anaconda2/envs/py3/lib/python3.6/site-packages/pint/quantity.py", line 1281, in __getitem__
    value = self._magnitude[key]
IndexError: index 0 is out of bounds for axis 0 with size 0
I debugged it with pdb and inserted following code to catch the error:
try:
    lfc_p, lfc_T = mpcalc.lfc(P,T,Td)
except IndexError:
    pdb.set_trace()




    

I ran this program again and then checked variables of P, T and Td when it encounter that error.
It shows that:
(Pdb) P
<Quantity([ 1004.  1000.   943.   928.   925.   850.   839.   749.   700.   699.
   603.   500.   404.   400.   363.   306.   300.   250.   213.   200.
   176.   150.   120.   105.   100.    70.    66.    58.    50.    40.
    33.    30.    23.    20.    16.], 'hectopascal')>
(Pdb) T
<Quantity([ 24.2  24.   20.2  21.6  21.4  20.4  20.2  14.4  13.2  13.    6.8  -3.3
 -13.1 -13.7 -17.9 -25.5 -26.9 -37.9 -46.7 -48.7 -52.1 -58.9 -67.3 -66.5
 -66.7 -65.1 -66.1 -60.9 -60.5 -57.7 -50.1 -50.3 -50.1 -47.9 -43.1], 'degC')>
(Pdb) Td
<Quantity([  2.19000000e+01   2.21000000e+01   1.92000000e+01   2.05000000e+01
   2.04000000e+01   1.84000000e+01   1.74000000e+01   8.40000000e+00
  -2.80000000e+00  -3.00000000e+00  -1.52000000e+01  -2.03000000e+01
  -2.91000000e+01  -2.77000000e+01  -2.49000000e+01  -3.95000000e+01
  -4.19000000e+01  -5.19000000e+01  -6.07000000e+01  -6.27000000e+01
  -6.51000000e+01  -7.19000000e+01   9.99999000e+05   9.99999000e+05
   9.99999000e+05   9.99999000e+05   9.99999000e+05   9.99999000e+05
   9.99999000e+05   9.99999000e+05   9.99999000e+05   9.99999000e+05
   9.99999000e+05   9.99999000e+05   9.99999000e+05], 'degC')>
So it seems that P, T and Td are all correct. After all, they are actually not axis with size 0. Most of the time, this function can run well.
Expected output :

Correct lfc_p and lfc_T.
Versions :
Python version : 3.6.6
MetPy version : 0.8.0
          
Save data file and script file (script file name can be "cal_sounding_data.py").

Keep data file and script file in the same directory and run

python cal_sounding_data.py
          
For this data, the LFC is 971.04 hPa, but the LCL is 970.70 hPa. This triggers the else block here and the variable idx is set to a NumPy array with one element containing False. x and y become "empty" quantities (e.g., <Quantity([], 'hectopascal')>), leading to the crash.
It would make more sense to me for the function to return the value of the LCL in this case.
          
Right, I've got a bug fix going on this. Essentially, it comes from the fact that parcel_profile does not include the LCL (for good reasons); this means the assumed shape of the profile jumps directly from the point below the LCL to the point above, which causes an intersection with the ambient temperature profile that doesn't occur if you include the LCL. This breaks the assumption in lfc.
I've got a fix locally I just need to finish up.
          
Given that the plot referred to previously looks like this:

should a separate issue be raised regarding the fact that the black line (parcel profile) does not pass through the LCL (the black point)? You mentioned earlier there is a good reason this isn't the case, but it would be surprising to see a plot like this, I would think.
The code used to make the plot:
import matplotlib.pyplot as plt
import numpy as np
import metpy.calc as mpcalc
from metpy.plots import SkewT
from metpy.units import units
p = np.array([1004.,  1000.,   943.,   928.,   925.,   850.,   839.,   749.,   700.,   699.,
              603.,   500.,   404.,   400.,   363.,   306.,   300.,   250.,   213.,   200.,
              176.,   150.,   120.,   105.,   100.,    70.,    66.,    58.,    50.,    40.,
              33.,    30.,    23.,    20.,    16.]) * units('hPa')
T = np.array([24.2,  24.,   20.2,  21.6,  21.4,  20.4,  20.2,  14.4,  13.2,  13.,    6.8,  -3.3,
              -13.1, -13.7, -17.9, -25.5, -26.9, -37.9, -46.7, -48.7, -52.1, -58.9, -67.3, -66.5,
              -66.7, -65.1, -66.1, -60.9, -60.5, -57.7, -50.1, -50.3, -50.1, -47.9, -43.1])  \
              * units('degC')
Td = np.array([2.19000000e+01,   2.21000000e+01,   1.92000000e+01,   2.05000000e+01,
               2.04000000e+01,   1.84000000e+01,   1.74000000e+01,   8.40000000e+00,
               -2.80000000e+00,  -3.00000000e+00,  -1.52000000e+01,  -2.03000000e+01,
               -2.91000000e+01,  -2.77000000e+01,  -2.49000000e+01,  -3.95000000e+01,
               -4.19000000e+01,  -5.19000000e+01,  -6.07000000e+01,  -6.27000000e+01,
               -6.51000000e+01,  -7.19000000e+01,   9.99999000e+05,   9.99999000e+05,
               9.99999000e+05,   9.99999000e+05,   9.99999000e+05,   9.99999000e+05,
               9.99999000e+05,   9.99999000e+05,   9.99999000e+05,   9.99999000e+05,
               9.99999000e+05,   9.99999000e+05,   9.99999000e+05]) * units('degC')
Td[Td.m>9e5] = -68 * units('degC')
# Calculate the LCL
lcl_pressure, lcl_temperature = mpcalc.lcl(p[0], T[0], Td[0])
print(lcl_pressure, lcl_temperature)
# Calculate the parcel profile.
parcel_prof = mpcalc.parcel_profile(p, T[0], Td[0]).to('degC')
# Create a new figure. The dimensions here give a good aspect ratio
fig = plt.figure(figsize=(9, 9))
skew = SkewT(fig, rotation=30)
# Plot the data using normal plotting functions, in this case using
# log scaling in Y, as dictated by the typical meteorological plot
skew.plot(p, T, 'r')
skew.plot(p, Td, 'g')
skew.ax.set_ylim(1050, 900)
skew.ax.set_xlim(20, 30)
# Plot LCL temperature as black dot
skew.plot(lcl_pressure, lcl_temperature, 'ko', markerfacecolor='black')
# Plot the parcel profile as a black line
skew.plot(p, parcel_prof, 'k', linewidth=2)
# Plot a zero degree isotherm
skew.ax.axvline(0, color='c', linestyle='--', linewidth=2)
# Add the relevant special lines
skew.plot_dry_adiabats()
skew.plot_moist_adiabats()
skew.plot_mixing_lines()
# Show the plot
plt.show()
Of course one could manually insert the LCL point into the profile list before plotting, but that seems like an unnecessary complication.




    

          
So the "good" reason is for simplicity of interface. You give the function a set of pressures (the first of which is the starting point) as well as the starting temp and dew point, and it gives you back temperatures at all the levels. It's simple to use, and simple to do things like plot and fill the area between that profile and temperature.
One of the things I'm adding as a result of this issue is a new function parcel_profile_with_lcl, which will take pressure, temperature, and (currently) starting dewpoint, and it gives you back pressure, ambient temperature, and parcel profile, all of which include the LCL. I'm wondering now if it would be better to make dew point the same for this function.
          
Including the LCL point in the dew point profile would make it easier to calculate the virtual temperature of the environment at the LCL. That in turn would make computing CAPE with the virtual temperature correction easier in cases where the LFC is at the LCL.

Edited to add: And in cases where the LFC is not at the LCL, it makes computing CIN with the virtual temperature correction easier.
      
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.
      
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.
      
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.
      
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
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
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
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
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
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