Process dataset elements with access element children

Fixes Unidata#115

siphon/catalog.py

@@ -85,12 +85,28 @@ def __init__(self, catalog_url):
         self.services = []
         self.catalog_refs = OrderedDict()
         self.metadata = {}
+        self.ds_with_access_elements_to_process = []
         service_skip_count = 0
         service_skip = 0
+        current_dataset = None
+        previous_dataset = None
         for child in root.iter():
             tag_type = child.tag.split('}')[-1]
             if tag_type == "dataset":
+                current_dataset = child.attrib['name']
                 self._process_dataset(child)
+
+                if previous_dataset:
+                    # see if the previously processed dataset has access elements as children
+                    # if so, these datasets need to be processed specially when making access_urls
+                    has_access_element_info = any(self.datasets[previous_dataset].access_element_info)
+                    if has_access_element_info:
+                        self.ds_with_access_elements_to_process.append(previous_dataset)
+
+                previous_dataset = current_dataset
+
+            elif tag_type == "access":
+                self.datasets[current_dataset].add_access_element_info(child)
             elif tag_type == "catalogRef":
                 self._process_catalog_ref(child)
             elif (tag_type == "metadata") or (tag_type == ""):
@@ -114,12 +130,13 @@ def __init__(self, catalog_url):
         self._process_datasets()
 
     def _process_dataset(self, element):
+        catalog_url = ''
         if "urlPath" in element.attrib:
             if element.attrib["urlPath"] == "latest.xml":
-                ds = Dataset(element, self.catalog_url)
-            else:
-                ds = Dataset(element)
-            self.datasets[ds.name] = ds
+                catalog_url = self.catalog_url
+
+        ds = Dataset(element, catalog_url=catalog_url)
+        self.datasets[ds.name] = ds
 
     def _process_catalog_ref(self, element):
         catalog_ref = CatalogRef(self.catalog_url, element)
@@ -132,8 +149,14 @@ def _process_metadata(self, element, tag_type):
 
     def _process_datasets(self):
         for dsName in list(self.datasets.keys()):
-            self.datasets[dsName].make_access_urls(
-                self.base_tds_url, self.services, metadata=self.metadata)
+            # check to see if dataset needs to have access urls created, if not, remove the dataset
+            has_url_path = not self.datasets[dsName].url_path is None
+            is_ds_with_access_elements_to_process = dsName in self.ds_with_access_elements_to_process
+            if has_url_path or is_ds_with_access_elements_to_process:
+                self.datasets[dsName].make_access_urls(
+                    self.base_tds_url, self.services, metadata=self.metadata)
+            else:
+                self.datasets.pop(dsName)
 
 
 class CatalogRef(object):
@@ -208,8 +231,12 @@ def __init__(self, element_node, catalog_url=""):
 
         """
         self.name = element_node.attrib['name']
-        self.url_path = element_node.attrib['urlPath']
+        if ("urlPath" in element_node.attrib):
+            self.url_path = element_node.attrib['urlPath']
+        else:
+            self.url_path = None
         self.catalog_name = ''
+        self.access_element_info = {}
         self._resolved = False
         self._resolverUrl = None
         # if latest.xml, resolve the latest url
@@ -297,8 +324,27 @@ def make_access_urls(self, catalog_url, all_services, metadata=None):
                     access_urls[service.service_type] = server_url + \
                         service.base + self.url_path
 
+        if any(self.access_element_info):
+
+            for service_type in self.access_element_info:
+                url_path = self.access_element_info[service_type]
+                found_service = None
+                for service in all_services:
+                    if service.name == service_type:
+                        found_service = service
+                        break
+
+                if found_service:
+                    access_urls[service_type] = server_url + \
+                                                    found_service.base + url_path
+
         self.access_urls = access_urls
 
+    def add_access_element_info(self, access_element):
+        service_name = access_element.attrib["serviceName"]
+        url_path = access_element.attrib["urlPath"]
+        self.access_element_info[service_name] = url_path
+
 
 class SimpleService(object):
     r"""