Update for open class

Changed the levels and variables attributes to properties.  These
were causing some slowdown of file opening and indexing. Now
levels and variables will be generated when referenced.

src/grib2io/_grib2io.py

@@ -73,27 +73,27 @@ class `grib2io.open`, the file named `filename` is opened for reading (`mode
 
     Attributes
     ----------
+    closed : bool
+        `True` is file handle is close; `False` otherwise.
+    current_message : int
+        Current position of the file in units of GRIB2 Messages.
+    levels : tuple
+        Tuple containing a unique list of wgrib2-formatted level/layer strings.
+    messages : int
+        Count of GRIB2 Messages contained in the file.
     mode : str
         File IO mode of opening the file.
     name : str
         Full path name of the GRIB2 file.
-    messages : int
-        Count of GRIB2 Messages contained in the file.
-    current_message : int
-        Current position of the file in units of GRIB2 Messages.
     size : int
         Size of the file in units of bytes.
-    closed : bool
-        `True` is file handle is close; `False` otherwise.
     variables : tuple
         Tuple containing a unique list of variable short names (i.e. GRIB2
         abbreviation names).
-    levels : tuple
-        Tuple containing a unique list of wgrib2-formatted level/layer strings.
     """
-    __slots__ = ('_fileid', '_filehandle', '_hasindex', '_index', '_pos',
-                 'closed', 'current_message', 'levels', 'messages', 'mode',
-                 'name', 'size', 'variables')
+    __slots__ = ('_fileid', '_filehandle', '_hasindex', '_index', '_nodata',
+                 '_pos', 'closed', 'current_message', 'messages', 'mode',
+                 'name', 'size')
     def __init__(self, filename: str, mode: str="r", **kwargs):
         """
         Initialize GRIB2 File object instance.
@@ -109,6 +109,9 @@ def __init__(self, filename: str, mode: str="r", **kwargs):
         # Manage keywords
         if "_xarray_backend" not in kwargs:
             kwargs["_xarray_backend"] = False
+            self._nodata = False
+        else:
+            self._nodata = kwargs["_xarray_backend"]
         if mode in {'a','r','w'}:
             mode = mode+'b'
             if 'w' in mode: mode += '+'
@@ -139,15 +142,10 @@ def __init__(self, filename: str, mode: str="r", **kwargs):
         self.current_message = 0
         self.size = fstat.st_size
         self.closed = self._filehandle.closed
-        self.levels = None
-        self.variables = None
         self._fileid = hashlib.sha1((self.name+str(fstat.st_ino)+
                                      str(self.size)).encode('ASCII')).hexdigest()
         if 'r' in self.mode:
-            try:
-                self._build_index(no_data=kwargs['_xarray_backend'])
-            except(KeyError):
-                self._build_index()
+            self._build_index()
         # FIX: Cannot perform reads on mode='a'
         #if 'a' in self.mode and self.size > 0: self._build_index()
 
@@ -195,7 +193,7 @@ def __getitem__(self, key):
             raise KeyError('Key must be an integer, slice, or GRIB2 variable shortName.')
 
 
-    def _build_index(self, no_data=False):
+    def _build_index(self):
         """Perform indexing of GRIB2 Messages."""
         # Initialize index dictionary
         if not self._hasindex:
@@ -321,7 +319,7 @@ def _build_index(self, no_data=False):
                         msg._msgnum = self.messages-1
                         msg._deflist = deflist
                         msg._coordlist = coordlist
-                        if not no_data:
+                        if not self._nodata:
                             msg._data = Grib2MessageOnDiskArray((msg.ny,msg.nx), 2,
                                                                 TYPE_OF_VALUES_DTYPE[msg.typeOfValues],
                                                                 self._filehandle,
@@ -356,10 +354,21 @@ def _build_index(self, no_data=False):
                     self._filehandle.seek(0)
                 break
 
-        # Index at end of _build_index()
-        if self._hasindex and not no_data:
-             self.variables = tuple(sorted(set([msg.shortName for msg in self._index['msg']])))
-             self.levels = tuple(sorted(set([msg.level for msg in self._index['msg']])))
+
+    @property
+    def levels(self):
+        if self._hasindex and not self._nodata:
+            return tuple(sorted(set([msg.level for msg in self._index['msg']])))
+        else:
+            return None
+
+
+    @property
+    def variables(self):
+        if self._hasindex and not self._nodata:
+            return tuple(sorted(set([msg.shortName for msg in self._index['msg']])))
+        else:
+            return None
 
 
     def close(self):