re-org package and define public api

config docs

Author: haileyajohnson

.gitignore

@@ -4,7 +4,9 @@
 .venv
 .vscode
 **/.coverage
+uv.lock
 
 # build files
 **/*.egg-info
-**/__pycache__
+**/__pycache__
+**/build

cdippy/__init__.py

@@ -0,0 +1,8 @@
+# import plots library
+from . import plots
+
+# import public top-level modules
+from . import cdipnc, nchashes, ncstats, ndbc, spectra, stndata
+
+# public API (i.e. "from cdippy import *")
+__all__ = ["plots", "cdipnc", "nchashes", "ncstats", "ndbc", "spectra", "stndata"]

cdippy/cdipnc.py

@@ -1,15 +1,19 @@
 from datetime import datetime, timedelta, timezone
 import os
 
+import logging
 import netCDF4
 
 import numpy as np
 import numbers
 from bisect import bisect_left, bisect_right
 
 import cdippy.ndbc as ndbc
-import cdippy.utils as cu
-import cdippy.url_utils as uu
+import cdippy.utils.utils as cu
+import cdippy.utils.urls as uu
+
+
+logger = logging.getLogger(__name__)
 
 
 class CDIPnc:
@@ -144,7 +148,7 @@ def set_request_info(
             is set to True, only nonpub records will be returned.
         """
         if start is None:
-            start = datetime(1975, 1, 1)
+            start = datetime(1975, 1, 1).replace(tzinfo=timezone.utc)
         if end is None:
             end = datetime.now(timezone.utc)
         self.set_timespan(start, end)
@@ -156,14 +160,19 @@ def set_request_info(
     def set_timespan(self, start, end):
         """Sets request timespan"""
         if isinstance(start, str):
-            self.start_dt = datetime.strptime(start, "%Y-%m-%d %H:%M:%S")
+            self.start_dt = datetime.strptime(start, "%Y-%m-%d %H:%M:%S").replace(
+                tzinfo=timezone.utc
+            )
         else:
             self.start_dt = start
         if isinstance(end, str):
-            self.end_dt = datetime.strptime(end, "%Y-%m-%d %H:%M:%S")
+            self.end_dt = datetime.strptime(end, "%Y-%m-%d %H:%M:%S").replace(
+                tzinfo=timezone.utc
+            )
         else:
             self.end_dt = end
         self.start_stamp = cu.datetime_to_timestamp(self.start_dt)
+
         self.end_stamp = cu.datetime_to_timestamp(self.end_dt)
 
     def get_request(self) -> dict:
@@ -365,23 +374,25 @@ def __get_indices(self, times: list, start_stamp: int, end_stamp: int) -> tuple:
         e_idx = bisect_right(times, end_stamp, s_idx)
         return s_idx, e_idx
 
-    def get_nc(self, url: str = None) -> netCDF4.Dataset:
-        if url is None:
+    def get_nc(self, url: str = None, retry: bool = False) -> netCDF4.Dataset:
+        if not url:
             url = self.url
-        # Check if the html page or file exists
-        if (
-            url[0:4] == "http" and not uu.url_exists(url + ".html")
-        ) and not os.path.isfile(url):
-            return None
         try:
-            nc = netCDF4.Dataset(url)
-        except Exception:
+            return netCDF4.Dataset(url)
+        except Exception as e:
             # Try again if unsuccessful (nc file not ready? THREDDS problem?)
-            try:
-                nc = netCDF4.Dataset(url)
-            except Exception:
-                nc = None
-        return nc
+            if retry:
+                logger.warning(
+                    msg=f"Retrying to open dataset at {url} due to an unexpected exception: {e}"
+                )
+                try:
+                    return netCDF4.Dataset(url)
+                except Exception:
+                    pass
+            logger.exception(
+                msg=f"Failed to open dataset at {url} due to an unexpected exception: {e}"
+            )
+            return None
 
     def byte_arr_to_string(self, b_arr: np.ma.masked_array) -> str:
         if np.ma.is_masked(b_arr):
@@ -1027,7 +1038,3 @@ def __init__(self, stn, data_dir=None, org=None):
         """For parameters see CDIPnc.set_dataset_info."""
         CDIPnc.__init__(self, data_dir)
         self.set_dataset_info(stn, org, "realtimexy")
-
-
-if __name__ == "__main__":
-    pass

cdippy/mobile.py

@@ -2,8 +2,8 @@
 from bisect import bisect_left
 
 from cdippy.cdipnc import CDIPnc
-import cdippy.timestamp_utils as tsu
-import cdippy.utils as cu
+import cdippy.utils.utils as tsu
+import cdippy.utils.utils as cu
 
 
 class MopData(CDIPnc):
@@ -297,7 +297,3 @@ def get_target_timespan(
 
         # If we get to here there's a problem
         return (None, None, None)
-
-
-if __name__ == "__main__":
-    pass

cdippy/mopdata.py

@@ -1,5 +1,5 @@
-import cdippy.url_utils as uu
-import cdippy.utils as cu
+import cdippy.utils.urls as uu
+import cdippy.utils.utils as cu
 
 
 class NcHashes:
@@ -49,7 +49,3 @@ def save_new_hashes(self):
 
     def get_old_hashes(self):
         return cu.pkl_load(self.hash_pkl)
-
-
-if __name__ == "__main__":
-    pass

cdippy/nchashes.py

@@ -116,12 +116,3 @@ def make_categorical_flag_var(self, flag_name: str):
             self.data[flag_name], categories=self.meta.get_flag_values(flag_name)
         )
         return cat.rename_categories(self.meta.get_flag_meanings(flag_name))
-
-
-if __name__ == "__main__":
-    # - Tests
-    def t1():
-        av = NcStats("100p1")
-        print(av.make_stats())
-
-    # t1()

cdippy/ncstats.py

@@ -3,8 +3,8 @@
 import os
 from datetime import datetime, timezone
 
-import cdippy.url_utils as uu
-import cdippy.utils as cu
+import cdippy.utils.urls as uu
+import cdippy.utils.utils as cu
 
 sos_base = "https://sdf.ndbc.noaa.gov/sos/server.php"
 request = "request=DescribeSensor"
@@ -42,7 +42,3 @@ def get_wmo_id(stn):
         return ids[stn]
     else:
         return None
-
-
-if __name__ == "__main__":
-    pass

cdippy/ndbc.py

@@ -0,0 +1,3 @@
+from . import annual_hs_boxplot, compendium, sst_climatology
+
+__all__ = ["annual_hs_boxplot", "compendium", "sst_climatology"]

cdippy/plots/__init__.py

@@ -1,5 +1,5 @@
-import cdippy.utils as ut
-import cdippy.plots.plot_utils as pu
+import cdippy.utils.utils as ut
+import cdippy.plots.utils as pu
 from datetime import datetime
 import numpy as np
 import calendar