Add EddyInfos to have some quick stat

Author: AntSimi

setup.py

@@ -40,6 +40,7 @@
             "EddyId = py_eddy_tracker.appli.grid:eddy_id",
             # eddies
             "MergeEddies = py_eddy_tracker.appli.eddies:merge_eddies",
+            "EddyInfos = py_eddy_tracker.appli.eddies:display_infos",
             # network
             "EddyNetworkGroup = py_eddy_tracker.appli.network:build_network",
             "EddyNetworkBuildPath = py_eddy_tracker.appli.network:divide_network",

src/py_eddy_tracker/appli/eddies.py

@@ -23,6 +23,7 @@
 from netCDF4 import Dataset
 from .. import EddyParser
 from ..observations.tracking import TrackEddiesObservations
+from ..observations.observation import EddiesObservations
 
 
 def merge_eddies():
@@ -54,3 +55,20 @@ def merge_eddies():
             other = other.add_rotation_type()
         obs = obs.merge(other)
     obs.write_file(filename=args.out)
+
+
+def display_infos():
+    parser = EddyParser("Display General inforamtion")
+    parser.add_argument("observations", nargs='+', help="Input observations to compute frequency")
+    args = parser.parse_args()
+    array_vars = [
+        "speed_contour_longitude",
+        "speed_contour_latitude",
+        "effective_contour_longitude",
+        "effective_contour_latitude",
+        "uavg_profile",
+    ]
+    for filename in args.observations:
+        print(f"-- {filename} --")
+        e = EddiesObservations.load_file(filename, remove_vars=array_vars)
+        print(e.__repr__())

src/py_eddy_tracker/observations/observation.py

@@ -199,7 +199,66 @@ def shape(self):
         return self.observations.shape
 
     def __repr__(self):
-        return "%d observations" % len(self.observations)
+        """
+        return general informations on dataset as a string
+
+        :return: informations on datasets
+        :rtype: str
+        """
+        t0, t1 = self.period
+        period = t1 - t0 + 1
+        bins_lat = (-90, -60, -15, 15, 60, 90)
+        bins_amplitude = array((0, 1, 2, 3, 4, 5, 10, 500))
+        bins_radius = array((0, 15, 30, 45, 60, 75, 100, 200, 2000))
+        nb_obs = self.observations.shape[0]
+
+        def hist(varname, x="lat", bins=bins_lat, percent=False, mean=False, nb=False):
+            """
+            :param str varname: variable to use to compute stat
+            :param str x: variable to use to know in which bins
+            :param array bins:
+            :param bool percent: normalize by sum of all bins
+            :param bool mean: compute mean by bins
+            :param nb mean: only count by bins
+            :return: value by bins
+            :rtype: array
+            """
+            if nb:
+                v = histogram(self[x], bins=bins)[0]
+            else:
+                v = histogram(self[x], bins=bins, weights=self[varname])[0]
+            if percent:
+                v = v.astype("f4") / v.sum() * 100
+            elif mean:
+                v /= histogram(self[x], bins=bins)[0]
+            return v
+
+        def box_display(value):
+            """Return value evenly spaced with few numbers"""
+            return "".join([f"{v_:10.2f}" for v_ in value])
+
+        return f"""{nb_obs} observations from {t0} to {t1} ({period} days, ~{nb_obs / period:.0f} obs/day)
+    Speed area      : {self["speed_area"].sum() / period / 1e12:.2f} Mkm²/day
+    Effective area  : {self["effective_area"].sum() / period / 1e12:.2f} Mkm²/day
+
+    Distribution in Amplitude:
+    Amplitude bounds (cm)  {box_display(bins_amplitude)}
+    Percent of eddies         : {box_display(hist('time', 'amplitude', bins_amplitude / 100., percent=True, nb=True))}
+
+    Distribution in Radius:
+    Speed radius (km)      {box_display(bins_radius)}
+    Percent of eddies         : {box_display(hist('time', 'radius_s', bins_radius * 1000., percent=True, nb=True))}
+
+    Distribution in Latitude
+    Latitude bounds        {box_display(bins_lat)}
+    Percent of eddies         : {box_display(hist('time', percent=True, nb=True))}
+    Percent of speed area     : {box_display(hist('speed_area', percent=True))}
+    Percent of effective area : {box_display(hist('effective_area', percent=True))}
+
+    Mean speed radius (km)    : {box_display(hist('radius_s', mean=True) / 1000.)}
+    Mean effective radius (km): {box_display(hist('radius_e', mean=True) / 1000.)}
+    Mean amplitude (cm)       : {box_display(hist('amplitude', mean=True) * 100.)}
+"""
 
     def __getitem__(self, attr):
         if attr in self.elements:
@@ -1366,6 +1425,16 @@ def interp_grid(
         else:
             raise Exception(f'method "{method}" unknown')
 
+    @property
+    def period(self):
+        """
+        Give time coverage
+
+        :return: first and last date
+        :rtype: (int,int)
+        """
+        return self.time.min(), self.time.max()
+
 
 @njit(cache=True)
 def grid_count_(grid, i, j):

src/py_eddy_tracker/observations/tracking.py

@@ -192,14 +192,6 @@ def extract_with_period(self, period, **kwargs):
             mask *= self.time <= (dataset_period[1] + p_max)
         return self.__extract_with_mask(mask, **kwargs)
 
-    @property
-    def period(self):
-        """
-        Give time coverage
-        Returns: 2 date
-        """
-        return self.time.min(), self.time.max()
-
     def get_mask_from_id(self, tracks):
         mask = zeros(self.tracks.shape, dtype=bool_)
         compute_mask_from_id(tracks, self.index_from_track, self.nb_obs_by_track, mask)