Merge pull request #730 from tmharty/recalculate_wyoming

Recalculate Wyoming

src/siphon/simplewebservice/wyoming.py

@@ -28,7 +28,7 @@ def __init__(self):
         super().__init__('http://weather.uwyo.edu/cgi-bin/sounding')
 
     @classmethod
-    def request_data(cls, time, site_id, **kwargs):
+    def request_data(cls, time, site_id, recalc=False, **kwargs):
         r"""Retrieve upper air observations from the Wyoming archive.
 
         Parameters
@@ -40,6 +40,11 @@ def request_data(cls, time, site_id, **kwargs):
             The three letter ICAO identifier of the station for which data should be
             downloaded.
 
+        recalc : bool
+            Whether to request the server recalculate the data (i.e. ignore its cache) before
+            returning it. Defaults to False. NOTE: This should be used sparingly because it
+            will increase the load on the service.
+
         kwargs
             Arbitrary keyword arguments to use to initialize source
 
@@ -50,10 +55,10 @@ def request_data(cls, time, site_id, **kwargs):
 
         """
         endpoint = cls()
-        df = endpoint._get_data(time, site_id)
+        df = endpoint._get_data(time, site_id, recalc=recalc)
         return df
 
-    def _get_data(self, time, site_id):
+    def _get_data(self, time, site_id, recalc=False):
         r"""Download and parse upper air observations from an online archive.
 
         Parameters
@@ -65,12 +70,15 @@ def _get_data(self, time, site_id):
             The three letter ICAO identifier of the station for which data should be
             downloaded.
 
+        recalc : bool
+            Returns recalculated data if True. Defaults to False.
+
         Returns
         -------
         `pandas.DataFrame`
 
         """
-        raw_data = self._get_data_raw(time, site_id)
+        raw_data = self._get_data_raw(time, site_id, recalc=recalc)
         soup = BeautifulSoup(raw_data, 'html.parser')
         tabular_data = StringIO(soup.find_all('pre')[0].contents[0])
         col_names = ['pressure', 'height', 'temperature', 'dewpoint', 'direction', 'speed']
@@ -129,7 +137,7 @@ def _get_data(self, time, site_id):
 
         return df
 
-    def _get_data_raw(self, time, site_id):
+    def _get_data_raw(self, time, site_id, recalc=False):
         """Download data from the University of Wyoming's upper air archive.
 
         Parameters
@@ -138,6 +146,8 @@ def _get_data_raw(self, time, site_id):
             Date and time for which data should be downloaded
         site_id : str
             Site id for which data should be downloaded
+        recalc : bool
+            Returns recalculated data if True. Defaults to False.
 
         Returns
         -------
@@ -148,6 +158,8 @@ def _get_data_raw(self, time, site_id):
         path = ('?region=naconf&TYPE=TEXT%3ALIST'
                 f'&YEAR={time:%Y}&MONTH={time:%m}&FROM={time:%d%H}&TO={time:%d%H}'
                 f'&STNM={site_id}')
+        if recalc:
+            path += '&REPLOT=1'
 
         resp = self.get_path(path)
         # See if the return is valid, but has no data

tests/test_wyoming.py

@@ -53,6 +53,45 @@ def test_wyoming():
     assert df.units['time'] is None
 
 
+@recorder.use_cassette('wyoming_sounding_recalculate')
+def test_wyoming_recalculate():
+    """Test that recalculation request returns the same data."""
+    df = WyomingUpperAir.request_data(
+        datetime(1999, 5, 4, 0), 'OUN', recalc=True)
+
+    assert df['time'][0] == datetime(1999, 5, 4, 0)
+    assert df['station'][0] == 'OUN'
+    assert df['station_number'][0] == 72357
+    assert df['latitude'][0] == 35.18
+    assert df['longitude'][0] == -97.44
+    assert df['elevation'][0] == 345.0
+
+    assert_almost_equal(df['pressure'][5], 867.9, 2)
+    assert_almost_equal(df['height'][5], 1219., 2)
+    assert_almost_equal(df['height'][30], 10505., 2)
+    assert_almost_equal(df['temperature'][5], 17.4, 2)
+    assert_almost_equal(df['dewpoint'][5], 14.3, 2)
+    assert_almost_equal(df['u_wind'][5], 6.60, 2)
+    assert_almost_equal(df['v_wind'][5], 37.42, 2)
+    assert_almost_equal(df['speed'][5], 38.0, 1)
+    assert_almost_equal(df['direction'][5], 190.0, 1)
+
+    assert df.units['pressure'] == 'hPa'
+    assert df.units['height'] == 'meter'
+    assert df.units['temperature'] == 'degC'
+    assert df.units['dewpoint'] == 'degC'
+    assert df.units['u_wind'] == 'knot'
+    assert df.units['v_wind'] == 'knot'
+    assert df.units['speed'] == 'knot'
+    assert df.units['direction'] == 'degrees'
+    assert df.units['latitude'] == 'degrees'
+    assert df.units['longitude'] == 'degrees'
+    assert df.units['elevation'] == 'meter'
+    assert df.units['station'] is None
+    assert df.units['station_number'] is None
+    assert df.units['time'] is None
+
+
 @recorder.use_cassette('wyoming_sounding_no_station')
 def test_wyoming_no_station():
     """Test that we handle stations with no ID from the Wyoming archive."""