Port IMPROVER to Python 3 #558

gavinevans · 2018-05-23T08:41:53Z

Addresses #488

Description:
Porting to Python 3.

Pycodestyle, pylintE and unit tests now seem to pass. Building the documentation works on Travis but not locally, although I haven't investigated this so far.

All CLI tests that don't require known good output pass. CLI tests that do require known good output currently all fail due to a change to hdf5libversion within each file, so all known good output needs regenerating. This can be done now that #538 has been merged.

Note that as this is a pull request to the Python 3 feature branch.

Testing:

[] Ran tests and they passed OK

fionaRust

A few initial comments to get you started.

fionaRust · 2018-05-30T12:42:30Z

bin/improver-spot-extract

@@ -233,7 +233,7 @@ def main():

    # Loop through the requested diagnostics and load data the essential data
    # and any additional data that is required.
-    for diagnostic_key in diagnostics.keys():
+    for diagnostic_key in list(diagnostics.keys()):


Does this need to be a list?

fionaRust · 2018-05-30T12:45:26Z

lib/improver/blending/weighted_blend.py

@@ -416,7 +416,7 @@ def __repr__(self):
        """Represent the configured plugin instance as a string."""
        description = ('<WeightedBlendAcrossWholeDimension:'
                       ' coord = {0:s}, weighting_mode = {1:s},'
-                       ' coord_adjust = {2:s}>')
+                       ' coord_adjust = {2}>')


Why has this lost :s when the line above hasn't.

This is because coord_adjust can be a None, rather than a string.

fionaRust · 2018-05-30T13:04:13Z

lib/improver/blending/weighted_blend.py

@@ -457,7 +457,7 @@ def process(self, cube, weights=None):
        if not isinstance(cube, iris.cube.Cube):
            msg = ('The first argument must be an instance of '
                   'iris.cube.Cube but is'
-                   ' {0:s}.'.format(type(cube)))
+                   ' {0:s}.'.format(str(type(cube))))


There are other ways of doing this:

potentially by removing the :s in the format specifier

putting !s in the format specifier, which calls str

These aren't necessarily better than your current version, which do you think is best?

I've removed the :s format specifier and removed enforcing type(cube) to be a string.

fionaRust · 2018-05-30T13:05:09Z

bin/improver-threshold

@@ -115,17 +115,17 @@ def main():
                thresholds_from_file = json.load(input_file)
            thresholds = []
            fuzzy_bounds = []
-            for key in thresholds_from_file.keys():
+            for key in list(thresholds_from_file.keys()):


Does this need to be a list?

fionaRust · 2018-05-30T13:09:37Z

lib/improver/ensemble_calibration/ensemble_calibration.py

@@ -876,7 +876,7 @@ def _apply_params(

            # If the coefficients are not available for the date, use the
            # raw ensemble forecast as the calibrated ensemble forecast.
-            if date not in optimised_coeffs.keys():
+            if date not in list(optimised_coeffs.keys()):


This can be changed to:
if date not in optimised_coeffs:
which is simpler

fionaRust · 2018-05-30T14:01:37Z

lib/improver/spotdata/main.py


        # Establish multiprocessing pool - each diagnostic processed on its
        # own thread.
        diagnostic_pool = mp.Pool(processes=n_diagnostic_threads)

        diagnostic_keys = [
-            diagnostic_name for diagnostic_name in diagnostics.keys()]
+            diagnostic_name for diagnostic_name in list(diagnostics.keys())]


only need diagnostic_keys = list(diagnostics.keys())

fionaRust · 2018-05-30T14:01:59Z

lib/improver/spotdata/main.py

@@ -192,7 +193,7 @@ def run_spotdata(diagnostics, ancillary_data, sites, config_constants,
        # Process diagnostics serially on one thread.
        resulting_cubes = CubeList()
        extrema_cubes = CubeList()
-        for key in diagnostics.keys():
+        for key in list(diagnostics.keys()):


for key in diagnostics:

fionaRust · 2018-05-30T14:03:16Z

lib/improver/spotdata/neighbour_finding.py

@@ -122,7 +122,7 @@ def process(self, cube, sites, ancillary_data,

        """
        if self.method == 'fast_nearest_neighbour':
-            if 'orography' in ancillary_data.keys():
+            if 'orography' in list(ancillary_data.keys()):


just need if 'orography' in ancillary_data

fionaRust · 2018-05-30T14:04:43Z

lib/improver/spotdata/read_input.py

@@ -63,7 +63,7 @@ def get_method_prerequisites(method, diagnostic_data_path):
        }

    additional_diagnostics = {}
-    if method in prereq.keys():
+    if method in list(prereq.keys()):


if method in prereq

fionaRust · 2018-05-30T14:05:11Z

lib/improver/spotdata/site_data.py

@@ -147,9 +147,9 @@ def parse_input(self, site_data):

        """
        latitude_entries = [i_site for (i_site, site) in enumerate(site_data)
-                            if 'latitude' in site.keys()]
+                            if 'latitude' in list(site.keys())]


if 'latitude' in site same thing below

fionaRust · 2018-05-30T14:20:54Z

lib/improver/tests/argparser/test_ArgParser.py

@@ -254,7 +257,7 @@ def test_adding_argument_with_defined_kwargs_dict(self):

        parser.add_arguments(args_to_add)
        result_args = parser.parse_args()
-        result_args = vars(result_args).keys()
+        result_args = list(vars(result_args).keys())


I think everywhere in this file you have something of the form list(vars(result_args).keys()) it can be simplified to just vars(result_args)

fionaRust · 2018-05-30T14:26:53Z

lib/improver/tests/argparser/test_ArgParser.py

@@ -311,6 +314,8 @@ class Test_wrong_args_error(unittest.TestCase):

    """Test the wrong_args_error method."""

+    @ManageWarnings(


Discussed trying an extra nested while, to try closing the file properly

fionaRust · 2018-05-30T14:35:17Z

...ensemble_calibration/ensemble_calibration/test_EstimateCoefficientsForEnsembleCalibration.py

@@ -399,7 +429,7 @@ def test_basic(self):
        optimised_coeffs, coeff_names = result
        self.assertIsInstance(optimised_coeffs, dict)
        self.assertIsInstance(coeff_names, list)
-        for key in optimised_coeffs.keys():
+        for key in list(optimised_coeffs.keys()):


for key in optimised_coeffs: and same below

fionaRust · 2018-05-30T14:35:40Z

...ensemble_calibration/ensemble_calibration/test_EstimateCoefficientsForEnsembleCalibration.py

@@ -335,6 +363,8 @@ def test_mean_predictor_estimate_coefficients_nans(self):

        current_forecast_predictor = cube.collapsed(
            "realization", iris.analysis.MEAN)
+        current_forecast_predictor.data = (


Why is this needed?

This has been added because otherwise you get this error:

test_EstimateCoefficientsForEnsembleCalibration.py:374: MaskedArrayFutureWarning: setting an item on a masked array which has a shared mask will not copy the mask and also change the original mask array in the future. Check the NumPy 1.11 release notes for more information. current_forecast_predictor.data[0][0][0] = np.nan

As the array is being specifically set to have NaN value, the masked values have been filled. This is due to a feature of cube.collapsed where the output cube seems to always be a masked array regardless of whether the input cube is a masked array.

fionaRust · 2018-05-30T14:37:07Z

...improver/tests/ensemble_copula_coupling/ensemble_copula_coupling/test_ResamplePercentiles.py

        upper_array = np.full(
-            forecast_at_percentiles[:, 0].shape, bounds_pairing[1])
+            forecast_at_percentiles[:, 0].shape, bounds_pairing[1],
+            dtype=np.int32)


Do you want this to be an int or a float?

For this unit test, I've changed this to be a float, as potentially this is more likely.

fionaRust · 2018-05-30T14:41:07Z

lib/improver/tests/nbhood/nbhood/test_BaseNeighbourhoodProcessing.py

@@ -125,7 +125,8 @@ def set_up_cube(zero_point_indices=((0, 0, 7, 7),), num_time_points=1,

    cube.add_dim_coord(
        DimCoord(
-            range(num_realization_points), standard_name='realization'), 0)
+            list(range(num_realization_points)),


This change isn't needed.

fionaRust · 2018-05-30T15:02:33Z

lib/improver/wxcode/weather_symbols.py

@@ -340,7 +340,7 @@ def find_all_routes(graph, start, end, route=None):
        route = route + [start]
        if start == end:
            return [route]
-        if start not in graph.keys():
+        if start not in list(graph.keys()):


Don't need list

fionaRust · 2018-05-30T15:02:49Z

lib/improver/wxcode/weather_symbols.py

@@ -394,12 +394,12 @@ def process(self, cubes):

        # Construct graph nodes dictionary
        graph = {key: [self.queries[key]['succeed'], self.queries[key]['fail']]
-                 for key in self.queries.keys()}
+                 for key in list(self.queries.keys())}


Don't need list

fionaRust · 2018-05-30T15:03:10Z

lib/improver/wxcode/wxcode_utilities.py

@@ -65,7 +65,7 @@
               29: 'Thunder_Shower_Day',
               30: 'Thunder'}

-WX_DICT = OrderedDict(sorted(_WX_DICT_IN.items(), key=lambda t: t[0]))
+WX_DICT = OrderedDict(sorted(list(_WX_DICT_IN.items()), key=lambda t: t[0]))


Might not need list

fionaRust · 2018-05-30T15:03:25Z

lib/improver/wxcode/wxcode_utilities.py

@@ -81,9 +81,9 @@ def add_wxcode_metadata(cube):
    cube.standard_name = None
    cube.var_name = None
    cube.units = "1"
-    wx_keys = np.array(WX_DICT.keys())
+    wx_keys = np.array(list(WX_DICT.keys()))


Not sure about these lists

I think this list is actually needed, as otherwise the attribute is:

array(odict_keys([0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30]), dtype=object)

rather than

array([ 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30])

fionaRust · 2018-05-30T15:03:57Z

tests/improver-ensemble-calibration/00_null.bats

@@ -1,4 +1,34 @@
 #!/usr/bin/env bats
+#!/usr/bin/env bats


You have this line twice now

fionaRust · 2018-05-30T15:05:50Z

lib/improver/utilities/mathematical_operations.py

@@ -165,11 +165,11 @@ def prepare_for_integration(self, cube):
        if self.direction_of_integration == "positive":
            integrated_cube = upper_bounds_cube.copy()
            integrated_cube.coord(self.coord_name_to_integrate).bounds = (
-                zip(lower_bounds, upper_bounds))
+                list(zip(lower_bounds, upper_bounds)))


Not sure if you need these lists in this file or not.

Removing this list causes the error:

ValueError: The shape of the bounds array should be points.shape + (n_bounds,)

… -E pass. Unit tests pass, although a few tests still generate warnings, which needs further investigation. CLI tests need further work, as all the known good output will change, as a result of an upgrade to the hdf5 version that is recorded within each netcdf file.

… Argparse unclosed file warning is now ignored, however, an unclosed socket warning will sometimes remains.

…ng that seems to arise based on the interaction between the ensemble calibration and the spot database tests.

… the ManageWarnings utility is only used within the unit tests.

…yCoefficientsForEnsembleCalibration.py and test_SpotDatabase.py, I've ignored Deprecation Warnings, which I think are coming from pandas that are causing Deprecation Warnings to be raised locally.

…leCalibration.py are trapped and other causes of errors are resolved.

…idoc. It seems like Sphinx should support the use of sphinx.apidoc, rather than sphinx.ext.apidoc, however, it seems like something about this interface doesn't seem to be working.

…UpClass.

… and removal of the use of the warnings handler within test_ArgParser.py.

…nary, as this syntax is more explicit.

…propriate.

fionaRust

Looking good now :)
Tests pass

Katie-Howard

I have checked through all the changes and they generally look good. I've just made a few comments and have a couple of questions about the changes, which I have added. Once they have been addressed I think I'm happy with it. I've run all the unit and cli tests and they all seem fine, which is good!

Katie-Howard · 2018-05-31T15:34:32Z

lib/improver/tests/ensemble_calibration/ensemble_calibration/test_EnsembleCalibration.py

+                    "\nThe final iteration resulted in a percentage "
+                    "change that is greater than the"
+                    " accepted threshold "]
+WARNING_TYPES = [UserWarning, ImportWarning, FutureWarning, RuntimeWarning,


Is there a reason there are multiples of UserWarning, and ImportWarning?

The reason that there are multiple UserWarning and ImportWarning is because these need to match up with the messages in the IGNORED_MESSAGES list i.e.:

IGNORED_MESSAGES WARNING_TYPES

"Collapsing a non-contiguous coordinate." UserWarning

"Not importing directory .*sphinxcontrib'" ImportWarning

"The pandas.core.datetools module is deprecated" FutureWarning

Therefore there needs to be the same number of warnings as there are IGNORED_MESSAGES, so that's why it looks like the warnings are repeated, as the different messages are associated with different types of warning. The default warning is a UserWarning, so in some instances, no WARNING_TYPES list is provided, if all of the messages are UserWarnings.

You're right that it's not obvious that these two lists need to be the same length. Maybe a dictionary would be an alternative interface e.g.

MESSAGES_AND_WARNINGS = {
"Collapsing a non-contiguous coordinate.": UserWarning,
"Not importing directory .*sphinxcontrib'": ImportWarning,
...
}
or a list of tuples:
MESSAGES_AND_WARNINGS = [(UserWarning, "Collapsing a non-contiguous coordinate."), (ImportWarning, "Not importing directory .*sphinxcontrib'"), ...]

No I don't think it is obvious, so it might be useful if it was easier to understand that is the reason there are multiple similar warnings? You could just put a comment, or as you suggested do a dictionary?

Katie-Howard · 2018-05-31T15:38:35Z

...ensemble_calibration/ensemble_calibration/test_EstimateCoefficientsForEnsembleCalibration.py

@@ -49,9 +49,23 @@
                             _create_historic_forecasts, _create_truth)
 from improver.utilities.warnings_handler import ManageWarnings

+


This is super picky but you could probably just delete this extra line?

Katie-Howard · 2018-05-31T15:39:44Z

...ensemble_calibration/ensemble_calibration/test_EstimateCoefficientsForEnsembleCalibration.py

+                    "\nThe final iteration resulted in a percentage "
+                    "change that is greater than the"
+                    " accepted threshold "]
+WARNING_TYPES = [UserWarning, ImportWarning, FutureWarning, RuntimeWarning,


Again do you need multiples of UserWarning and ImportWarning?

Katie-Howard · 2018-05-31T15:40:56Z

...ensemble_calibration/ensemble_calibration/test_EstimateCoefficientsForEnsembleCalibration.py

+                          "\nThe final iteration resulted in a percentage "
+                          "change that is greater than the"
+                          " accepted threshold "],
+        warning_types=[UserWarning, ImportWarning, FutureWarning,


Do you need multiples of UserWarning and ImportWarning?

Katie-Howard · 2018-05-31T15:42:09Z

...ensemble_calibration/ensemble_calibration/test_EstimateCoefficientsForEnsembleCalibration.py

@@ -137,7 +165,7 @@ def test_statsmodels_members(self, warning_list=None):
            plugin = Plugin(distribution, desired_units,
                            predictor_of_mean_flag=predictor_of_mean_flag)
            self.assertTrue(len(warning_list) == 1)
-            self.assertTrue(any(item.category == UserWarning
+            self.assertTrue(any(item.category == ImportWarning


Why has this changed from UserWarning to ImportWarning?

Katie-Howard · 2018-05-31T15:48:57Z

lib/improver/tests/spotdata/spotdata/test_WriteOutput.py

            Plugin(method, self.data_directory).process(self.cube)

-    def test_invalid_output_path(self):


Do we not need this test anymore?

The reason that I removed this test was because it was causing a number of low-level warnings that looked like they were coming from the underlying C code:

.HDF5-DIAG: Error detected in HDF5 (1.10.1) thread 140639791306496: #000: H5F.c line 491 in H5Fcreate(): unable to create file major: File accessibilty minor: Unable to open file #001: H5Fint.c line 1247 in H5F_open(): unable to open file: time = Fri Jun 1 08:45:57 2018 , name = '/test_data.nc', tent_flags = 13 major: File accessibilty minor: Unable to open file #002: H5FD.c line 809 in H5FD_open(): open failed major: Virtual File Layer minor: Unable to initialize object #003: H5FDsec2.c line 346 in H5FD_sec2_open(): unable to open file: name = '/test_data.nc', errno = 13, error message = 'Permission denied', flags = 13, o_flags = 242 major: File accessibilty minor: Unable to open file

I don't think there would be a way of ignoring these messages. Also, I thought that this test was a bit unusual because essentially it's testing what happens when you try to save a file to the root directory, so it's more of a test of the file system, rather than plugin. Therefore, I thought that the simplest option would be to remove this test.

Ok, that sounds fine to me.

* Modifications to upgrade IMPROVER to Python 3. Pycodestyle and pylint -E pass. Unit tests pass, although a few tests still generate warnings, which needs further investigation. CLI tests need further work, as all the known good output will change, as a result of an upgrade to the hdf5 version that is recorded within each netcdf file. * Edits to ignore relevant warnings arising when importing statsmodels. Argparse unclosed file warning is now ignored, however, an unclosed socket warning will sometimes remains. * Update python version used in Travis. * Further minor edits for warnings. There is still one DeprecationWarning that seems to arise based on the interaction between the ensemble calibration and the spot database tests. * Reconfigure the managing of warnings for ensemble calibration, so the the ManageWarnings utility is only used within the unit tests. * Make unit test output verbose. * As a result of some interaction between the warnings within test_ApplyCoefficientsForEnsembleCalibration.py and test_SpotDatabase.py, I've ignored Deprecation Warnings, which I think are coming from pandas that are causing Deprecation Warnings to be raised locally. * Edits to ensure warnings raised in test_EstimateCoefficientsForEnsembleCalibration.py are trapped and other causes of errors are resolved. * Print statement to try to debug Travis error. * Modify warnings that are managed by test_statsmodels_members. * Remove verbosity from unit test output. * Sphinx documentation fix, as sphinx.apidoc has moved to sphinx.ext.apidoc. It seems like Sphinx should support the use of sphinx.apidoc, rather than sphinx.ext.apidoc, however, it seems like something about this interface doesn't seem to be working. * Remove unused import of ManageWarnings. * Include an explicit closing of the file that is opened within the setUpClass. * Removal of unnecessary lists added by the 2to3 tool. * Further edits to remove unnecessary conversion of variables to a list and removal of the use of the warnings handler within test_ArgParser.py. * Edits to remove enforcing of string formatting. * Update to the type of error raised in CLI tests for a missing land mask. * Re-added usage of keys() when iterating over the keys within a dictionary, as this syntax is more explicit. * Further edits to explicitly access the keys on a dictionary, where appropriate. * Removed extra blank line.

* latest: (1367 commits) Move vicinity processing to threshold CLI (metoppv#581) Refactored to allow for separate filling functions (metoppv#587) Add warning if too many zero values for advection velocities (metoppv#585) Impro 621 improve snow (metoppv#586) metoppv#442: Add profiling support to CLIs (metoppv#554) Update readthedocs config to include netcdftime Add test for licence information (metoppv#582) IMPRO-639: GenerateProbabilitiesFromMeanAndVariance plugin (metoppv#578) Removal of renaming of coordinates within ensemble calibration, as this functionality is not used, as no input is provided with ensemble_member_id or ensemble_realization_id as a coordinate. (metoppv#579) Add coverage badge to README Try out coverage (metoppv#565) Impro-672 use realizations rather than members (metoppv#577) Modifications to allow neighbourhood processing of complex numbers (metoppv#570) Add support for providing a verbose option for the unit tests. (metoppv#569) IMPRO-618: Grid to grid interpolation (metoppv#566) IMPRO-212: A faster percentile method. (metoppv#568) Impro 624 add coordinate collapse (metoppv#567) Add Python version to README.md Port IMPROVER to Python 3 (metoppv#558) (metoppv#564) Reordered KGO creation call position so it gets made even when test fails. (metoppv#562) ...

* Modifications to upgrade IMPROVER to Python 3. Pycodestyle and pylint -E pass. Unit tests pass, although a few tests still generate warnings, which needs further investigation. CLI tests need further work, as all the known good output will change, as a result of an upgrade to the hdf5 version that is recorded within each netcdf file. * Edits to ignore relevant warnings arising when importing statsmodels. Argparse unclosed file warning is now ignored, however, an unclosed socket warning will sometimes remains. * Update python version used in Travis. * Further minor edits for warnings. There is still one DeprecationWarning that seems to arise based on the interaction between the ensemble calibration and the spot database tests. * Reconfigure the managing of warnings for ensemble calibration, so the the ManageWarnings utility is only used within the unit tests. * Make unit test output verbose. * As a result of some interaction between the warnings within test_ApplyCoefficientsForEnsembleCalibration.py and test_SpotDatabase.py, I've ignored Deprecation Warnings, which I think are coming from pandas that are causing Deprecation Warnings to be raised locally. * Edits to ensure warnings raised in test_EstimateCoefficientsForEnsembleCalibration.py are trapped and other causes of errors are resolved. * Print statement to try to debug Travis error. * Modify warnings that are managed by test_statsmodels_members. * Remove verbosity from unit test output. * Sphinx documentation fix, as sphinx.apidoc has moved to sphinx.ext.apidoc. It seems like Sphinx should support the use of sphinx.apidoc, rather than sphinx.ext.apidoc, however, it seems like something about this interface doesn't seem to be working. * Remove unused import of ManageWarnings. * Include an explicit closing of the file that is opened within the setUpClass. * Removal of unnecessary lists added by the 2to3 tool. * Further edits to remove unnecessary conversion of variables to a list and removal of the use of the warnings handler within test_ArgParser.py. * Edits to remove enforcing of string formatting. * Update to the type of error raised in CLI tests for a missing land mask. * Re-added usage of keys() when iterating over the keys within a dictionary, as this syntax is more explicit. * Further edits to explicitly access the keys on a dictionary, where appropriate. * Removed extra blank line.

gavinevans force-pushed the impro_591_python_3 branch from f078920 to 05bdd6b Compare May 23, 2018 14:40

metoppv deleted a comment May 30, 2018

fionaRust self-requested a review May 30, 2018 12:18

gavinevans force-pushed the impro_591_python_3 branch from 9499430 to cd78a8b Compare May 30, 2018 13:20

fionaRust reviewed May 30, 2018

View reviewed changes

gavinevans force-pushed the impro_591_python_3 branch from cd78a8b to 3aabaf8 Compare May 30, 2018 15:20

gavinevans added 15 commits May 30, 2018 16:25

Edits to ignore relevant warnings arising when importing statsmodels.…

8521746

… Argparse unclosed file warning is now ignored, however, an unclosed socket warning will sometimes remains.

Update python version used in Travis.

2d0589c

Further minor edits for warnings. There is still one DeprecationWarni…

bfb8e1f

…ng that seems to arise based on the interaction between the ensemble calibration and the spot database tests.

Reconfigure the managing of warnings for ensemble calibration, so the…

5f34cc7

… the ManageWarnings utility is only used within the unit tests.

Make unit test output verbose.

8b35d50

As a result of some interaction between the warnings within test_Appl…

4a5e529

…yCoefficientsForEnsembleCalibration.py and test_SpotDatabase.py, I've ignored Deprecation Warnings, which I think are coming from pandas that are causing Deprecation Warnings to be raised locally.

Edits to ensure warnings raised in test_EstimateCoefficientsForEnsemb…

925b75e

…leCalibration.py are trapped and other causes of errors are resolved.

Print statement to try to debug Travis error.

d12b5bc

Modify warnings that are managed by test_statsmodels_members.

8a332dd

Remove verbosity from unit test output.

b2d9950

Sphinx documentation fix, as sphinx.apidoc has moved to sphinx.ext.ap…

cb56e9d

…idoc. It seems like Sphinx should support the use of sphinx.apidoc, rather than sphinx.ext.apidoc, however, it seems like something about this interface doesn't seem to be working.

Remove unused import of ManageWarnings.

3623265

Include an explicit closing of the file that is opened within the set…

d36d57a

…UpClass.

Removal of unnecessary lists added by the 2to3 tool.

df46464

gavinevans force-pushed the impro_591_python_3 branch 2 times, most recently from 44b4119 to df46464 Compare May 30, 2018 15:36

metoppv deleted a comment May 30, 2018

gavinevans added 2 commits May 30, 2018 17:34

Further edits to remove unnecessary conversion of variables to a list…

bc81ece

… and removal of the use of the warnings handler within test_ArgParser.py.

Edits to remove enforcing of string formatting.

0c0cc8b

metoppv deleted a comment May 30, 2018

Update to the type of error raised in CLI tests for a missing land mask.

65018b5

metoppv deleted a comment May 31, 2018

gavinevans added 2 commits May 31, 2018 10:17

Re-added usage of keys() when iterating over the keys within a dictio…

6db85dc

…nary, as this syntax is more explicit.

Further edits to explicitly access the keys on a dictionary, where ap…

4971791

…propriate.

fionaRust approved these changes May 31, 2018

View reviewed changes

metoppv deleted a comment May 31, 2018

Katie-Howard requested changes May 31, 2018

View reviewed changes

Removed extra blank line.

72acb8e

Katie-Howard approved these changes Jun 1, 2018

View reviewed changes

fionaRust merged commit 7719345 into metoppv:feature_branch_python_3 Jun 1, 2018

This was referenced Jun 5, 2018

Add support for providing a verbose option for the unit tests #569

Merged

Update to remove warnings when running the Iris 2 stack #540

Closed

This was referenced Jun 15, 2018

Scope porting to Python 3 #265

Closed

Port IMPROVER to Python 3 #488

Closed

Start porting to Python 3 #487

Closed

		@@ -311,6 +314,8 @@ class Test_wrong_args_error(unittest.TestCase):

		"""Test the wrong_args_error method."""

		@ManageWarnings(

IGNORED_MESSAGES	WARNING_TYPES
"Collapsing a non-contiguous coordinate."	UserWarning
"Not importing directory .*sphinxcontrib'"	ImportWarning
"The pandas.core.datetools module is deprecated"	FutureWarning

		@@ -49,9 +49,23 @@
		_create_historic_forecasts, _create_truth)
		from improver.utilities.warnings_handler import ManageWarnings

		Plugin(method, self.data_directory).process(self.cube)

		def test_invalid_output_path(self):

Port IMPROVER to Python 3 #558

Port IMPROVER to Python 3 #558

Conversation

gavinevans commented May 23, 2018

fionaRust left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

fionaRust left a comment

Choose a reason for hiding this comment

Katie-Howard left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Katie-Howard Jun 1, 2018 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Katie-Howard Jun 1, 2018 •

edited

Loading