#24: added validation to DatasetURI's save method, and changed valida…

…tion function to only raise exception if the resource is not available

geospaas/catalog/models.py

@@ -185,6 +185,7 @@ def protocol(self):
         return self.uri.split(':')[0]
 
     def save(self, *args, **kwargs):
+        validate_uri(uri)
         # Validation is not usually done in the models but rather via form
         # validation. We should discuss if we want it or not. Presently, we
         # check only that the uri is valid but we may also check if it exists

geospaas/nansat_ingestor/managers.py

@@ -39,9 +39,9 @@ def get_or_create(self, uri, n_points=10, uri_filter_args=None, *args, **kwargs)
         if not uri_filter_args:
             uri_filter_args = {}
 
-        # Validate uri - this should fail if the uri doesn't point to a valid
+        # Validate uri - this should raise an exception if the uri doesn't point to a valid
         # file or stream
-        valid_uri = validate_uri(uri)
+        validate_uri(uri)
 
         # Several datasets can refer to the same uri (e.g., scatterometers and svp drifters), so we
         # need to pass uri_filter_args

geospaas/utils/utils.py

@@ -44,25 +44,27 @@ def product_path(module, filename):
 def validate_uri(uri):
     """ Validation of URI and its existence
 
-    URI should be either: file://localhost/some/path/filename.ext
-           or accessible: http://www.eee.rrr/some/path
+    URI conventions: URI = scheme:[//authority]path[?query][#fragment]
+    
+    Examples:
+        file://localhost/some/path/filename.ext
+        http://www.eee.rrr/some/path
 
-    If URI is not valid, the function rasies a ValueError or urrlib error
+    If URI is not valid, the function raises a ValueError or urrlib error
 
     """
     validation_result = False
     uri_parts = urlparse(uri)
     if uri_parts.scheme=='file' and uri_parts.netloc=='localhost':
-        if os.path.isfile(uri_parts.path):
-            validation_result = True
+        if not os.path.isfile(uri_parts.path):
+            raise FileNotFoundError(uri_parts.path)
     else:
         if URLLIB_VERSION == 2:
             request = urllibN.Request(uri)
         else:
             request = urllibN.PoolManager().request('GET', uri)
-        if request.status==200:
-            validation_result = True
-    return validation_result
+        if not request.status==200:
+            raise ConnectionError(uri)
 
 def nansat_filename(uri):
     # Check if data should be read as stream or as file? Or just: