Merge remote-tracking branch 'scclab/22-other-nc-files-in-folder-conflict-the-loading'

o3skim/sources.py

@@ -21,10 +21,10 @@ class Source:
 
     def skim(self):
         for name, model in self._models.items():
-            path = self._name + "_" + name
-            os.makedirs(path, exist_ok=True)
-            logger.info("Skim data from '%s'", path)
-            model.skim(path)
+            dirname = self._name + "_" + name
+            os.makedirs(dirname, exist_ok=True)
+            logger.info("Skim data from '%s'", dirname)
+            model.skim(dirname)
 
 
 class Model:
@@ -38,19 +38,18 @@ class Model:
             logger.debug("Load 'vmro3_zm' data")
             self.__get_vmro3_zm(**variables)
 
-    def skim(self, path):
+    def skim(self, dirname):
         if hasattr(self, '_tco3_zm'):
             logger.debug("Skim 'tco3_zm' data")
-            utils.to_netcdf(path, "tco3_zm", self._tco3_zm)
+            utils.to_netcdf(dirname, "tco3_zm", self._tco3_zm)
         if hasattr(self, '_vmro3_zm'):
             logger.debug("Skim 'vmro3_zm' data")
-            utils.to_netcdf(path, "vmro3_zm", self._vmro3_zm)
+            utils.to_netcdf(dirname, "vmro3_zm", self._vmro3_zm)
 
     @utils.return_on_failure("Error when loading 'tco3_zm'")
     def __get_tco3_zm(self, tco3_zm, **kwarg):
         """Gets and standarises the tco3_zm data"""
-        fnames = glob.glob(tco3_zm['dir'] + "/*.nc")
-        with xr.open_mfdataset(fnames) as dataset:
+        with xr.open_mfdataset(tco3_zm['paths']) as dataset:
             dataset = dataset.rename({
                 tco3_zm['name']: 'tco3_zm',
                 tco3_zm['coordinades']['time']: 'time',
@@ -62,8 +61,7 @@ class Model:
     @utils.return_on_failure("Error when loading 'vmro3_zm'")
     def __get_vmro3_zm(self, vmro3_zm, **kwarg):
         """Gets and standarises the vmro3_zm data"""
-        fnames = glob.glob(vmro3_zm['dir'] + "/*.nc")
-        with xr.open_mfdataset(fnames) as dataset:
+        with xr.open_mfdataset(vmro3_zm['paths']) as dataset:
             dataset = dataset.rename({
                 vmro3_zm['name']: 'vmro3_zm',
                 vmro3_zm['coordinades']['time']: 'time',

o3skim/utils.py

@@ -50,10 +50,10 @@ def create_empty_netCDF(fname):
     root_grp.close()
 
 
-def to_netcdf(path, name, dataset):
+def to_netcdf(dirname, name, dataset):
     """Creates or appends data to named netcdf files"""
     years, dsx = zip(*dataset.groupby("time.year"))
-    fnames = [path + "/" + name + "_%s.nc" % y for y in years]
+    fnames = [dirname + "/" + name + "_%s.nc" % y for y in years]
     logging.info("Save dataset into: %s", fnames)
     [create_empty_netCDF(fn) for fn in fnames if not os.path.isfile(fn)]
     xr.save_mfdataset(dsx, fnames, mode='a')

sources_example.yaml

@@ -23,7 +23,7 @@
 # expected to have only one folder output named "CCMI-1_IPSL".
 #
 # This model has 2 variables (tco3_zm and vmro3_zm) which datasets are 
-# located in different directories. Therefore the key 'dir' is different
+# located in different directories. Therefore the key 'paths' is different
 # in both of them. Therefore, the output expected at "CCMI-1_IPSL" is 
 # 2 type of files: 
 #   - tco3_zm_[YEAR].nc: With tco3 skimmed data
@@ -44,7 +44,7 @@ CCMI-1:
             name: toz
             # Path where to find the netCDF files
             # [FIXED_KEY -- MANDATORY]: [CORRECT_VALUE -- MANDATORY]
-            dir: Ccmi/mon/toz
+            paths: Ccmi/mon/toz/*.nc
             # Coordinades description for tco3 data. 
             # [FIXED_KEY -- MANDATORY]:
             coordinades:
@@ -62,7 +62,7 @@ CCMI-1:
             name: vmro3
             # Path where to find the netCDF files
             # [FIXED_KEY -- MANDATORY]: [CORRECT_VALUE -- MANDATORY]
-            dir: Ccmi/mon/vmro3
+            paths: Ccmi/mon/vmro3/*_????.nc
             # Coordinades description for vmro3 data. 
             # [FIXED_KEY -- MANDATORY]: 
             coordinades:
@@ -85,7 +85,7 @@ CCMI-1:
 #
 # This case of ERA-i indeed has 2 variables (tco3_zm and vmro3_zm) but in
 # this case, are located inside the same dataset files, therefore the 
-# key 'dir' is the same in both variables. The output expected at 
+# key 'paths' is the same in both variables. The output expected at 
 # "ECMWF_ERA-5" is 2 type of files: 
 #   - tco3_zm_[YEAR].nc: With tco3 skimmed data
 #   - vmro3_zm_[YEAR].nc: With vmro3 skimmed data
@@ -105,7 +105,7 @@ ECMWF:
             name: tco3
             # Path where to find the netCDF files
             # [FIXED_KEY -- MANDATORY]: [CORRECT_VALUE -- MANDATORY]
-            dir: Ecmwf/Era5
+            paths: Ecmwf/Era5/*.nc
             # Coordinades description for tco3 data. 
             # [FIXED_KEY -- MANDATORY]:
             coordinades:
@@ -123,7 +123,7 @@ ECMWF:
             name: toz
             # Path where to find the netCDF files
             # [FIXED_KEY -- MANDATORY]: [CORRECT_VALUE -- MANDATORY]
-            dir: Ecmwf/Erai
+            paths: Ecmwf/Erai/*.nc
             # Coordinades description for tco3 data. 
             # [FIXED_KEY -- MANDATORY]:
             coordinades:
@@ -141,7 +141,7 @@ ECMWF:
             name: vmro3
             # Path where to find the netCDF files
             # [FIXED_KEY -- MANDATORY]: [CORRECT_VALUE -- MANDATORY]
-            dir: Ecmwf/Erai
+            paths: Ecmwf/Erai/*.nc
             # Coordinades description for vmro3 data. 
             # [FIXED_KEY -- MANDATORY]: 
             coordinades:

tests/mockup_data.py

@@ -2,9 +2,8 @@
 
 import xarray as xr
 import numpy as np
-import netCDF4
-import os.path
 import datetime
+import os
 from o3skim import utils
 
 
@@ -37,7 +36,7 @@ def dataset(name, coordinades):
     )
 
 
-def netcdf(path, name, coordinades, **kwarg):
+def netcdf(dirname, name, coordinades, **kwarg):
     """Creates or appends data to a mock netcdf file"""
     ds = dataset(name, coordinades)
-    utils.to_netcdf(path, name, ds)
+    utils.to_netcdf(dirname, name, ds)

tests/mockup_noise.py

+"""This modules creates mockup data for testing"""
+
+import xarray as xr
+import numpy as np
+import datetime
+import random
+
+
+base = datetime.datetime(2000, 1, 1)
+indexes = {
+    'time': [base + datetime.timedelta(days=9*i) for i in range(99)],
+    'plev': [x for x in range(1, 1000, 100)],
+    'lat':  [x for x in range(-90, 90, 10)],
+    'lon':  [x for x in range(-180, 180, 20)]
+}
+
+
+def data_vars():
+    """Creates a mock n-array with coordinade values"""
+    dim = [len(axis) for _, axis in indexes.items()]
+    return tuple(indexes), np.ones(dim),
+
+
+def data_coord(): 
+    """Creates a mock coordinades"""
+    return indexes
+
+
+def dataset(name):
+    """Creates a dataset acording to the global module indexes"""
+    return xr.Dataset(
+        {name: data_vars()},
+        coords = data_coord()
+    )
+
+
+def netcdf(path, name, **kwarg):
+    """Creates or appends data to a noise netcdf file"""
+    ds=dataset(name)
+    ds.to_netcdf(path)

tests/noise_files.yaml

+- name: toz
+  path: Ccmi/mon/toz/toz_noise.nc
+- name: tco3
+  path: Ecmwf/Era5/tco3_noise.nc
+- name: toz
+  path: Ecmwf/Erai/toz_noise.nc
+- name: vmro3
+  path: Ecmwf/Erai/vmro3_noise.nc

tests/sources_base.yaml

@@ -2,14 +2,14 @@ CCMI-1:
     IPSL:
         tco3_zm:
             name: toz
-            dir: Ccmi/mon/toz
+            paths: Ccmi/mon/toz/toz_????.nc
             coordinades:
                 time: time
                 lat: lat
                 lon: lon
         vmro3_zm:
             name: vmro3
-            dir: Ccmi/mon/vmro3
+            paths: Ccmi/mon/vmro3/*.nc
             coordinades:
                 time: time
                 plev: plev
@@ -19,7 +19,7 @@ ECMWF:
     ERA-5:
         tco3_zm:
             name: tco3
-            dir: Ecmwf/Era5
+            paths: Ecmwf/Era5/tco3_????.nc
             coordinades:
                 lon: longitude 
                 lat: latitude
@@ -27,14 +27,14 @@ ECMWF:
     ERA-i:
         tco3_zm:
             name: toz
-            dir: Ecmwf/Erai
+            paths: Ecmwf/Erai/toz_????.nc
             coordinades:
                 time: time
                 lat: latitude
                 lon: longitude 
         vmro3_zm:
             name: vmro3
-            dir: Ecmwf/Erai
+            paths: Ecmwf/Erai/vmro3_????.nc
             coordinades:
                 time: time
                 plev: level

tests/sources_err.yaml

@@ -2,7 +2,7 @@ ECMWF:
     ERA-i:
         vmro3_zm: # Correct variable
             name: vmro3
-            dir: Ecmwf/Erai
+            paths: Ecmwf/Erai/vmro3_????.nc
             coordinades:
                 time: time
                 plev: level
@@ -10,7 +10,7 @@ ECMWF:
                 lon: longitude
         tco3_zm: # Incorrect variable
             name: non_existing_var
-            dir: Ecmwf/Erai
+            paths: Ecmwf/Erai/toz_????.nc
             coordinades:
                 time: time
                 lat: latitude

tests/test_o3skim.py

@@ -10,6 +10,7 @@ import glob
 from o3skim import sources, utils
 # from pyfakefs.fake_filesystem_unittest import TestCase
 from . import mockup_data
+from . import mockup_noise
 
 
 class TestO3SKIM_sources(unittest.TestCase):
@@ -25,22 +26,31 @@ class TestO3SKIM_sources(unittest.TestCase):
         self.create_mock_datasets()
         self.backup_datasets()
         self.assert_with_backup()
+        self.create_noise_datasets()
 
     def tearDown(self):
         """Tear down test fixtures, if any."""
 
     def create_mock_datasets(self):
         """Creates mock data files according to the loaded configuration"""
+        with utils.cd('data'):
             for _, collection in self.config_base.items():
                 for _, variables in collection.items():
                     for _, vinfo in variables.items():
-                    path = "data/" + vinfo["dir"]
-                    os.makedirs(path, exist_ok=True)
-                    mockup_data.netcdf(path, **vinfo)
+                        dirname = os.path.dirname(vinfo['paths'])
+                        os.makedirs(dirname, exist_ok=True)
+                        mockup_data.netcdf(dirname, **vinfo)
+
+    def create_noise_datasets(self):
+        """Creates noise data files according to the noise configuration"""
+        config_noise = utils.load("tests/noise_files.yaml")
+        with utils.cd('data'):
+            for ninfo in config_noise:
+                mockup_noise.netcdf(**ninfo)
 
     def clean_output(self):
         """Cleans output removing all folders at output"""
-        with utils.cd("output"):
+        with utils.cd('output'):
             directories = (d for d in os.listdir() if os.path.isdir(d))
             for directory in directories:
                 shutil.rmtree(directory) 
@@ -48,22 +58,22 @@ class TestO3SKIM_sources(unittest.TestCase):
     def backup_datasets(self):
         """Loads the mock datasets into an internal variable"""
         self.ds_backup = {}
+        with utils.cd('data'): 
             for source, collection in self.config_base.items():
                 self.ds_backup[source] = {}
                 for model, variables in collection.items():
                     self.ds_backup[source][model] = {}
                     for v, vinfo in variables.items(): 
-                    paths = "data/" + vinfo["dir"] + "/*.nc"
-                    with xr.open_mfdataset(paths) as ds:
+                        with xr.open_mfdataset(vinfo['paths']) as ds:
                             self.ds_backup[source][model][v] = ds
 
     def assert_with_backup(self):
         """Asserts the dataset in the backup is equal to the config load"""
+        with utils.cd('data'): 
             for source, collection in self.config_base.items():
                 for model, variables in collection.items():
                     for v, vinfo in variables.items():
-                    paths = "data/" + vinfo["dir"] + "/*.nc"
-                    with xr.open_mfdataset(paths) as ds:
+                        with xr.open_mfdataset(vinfo['paths']) as ds:
                             xr.testing.assert_identical(
                                 self.ds_backup[source][model][v], ds)