Merge branch 'develop' into fix373

Author: janvanrijn

openml/datasets/functions.py

@@ -137,7 +137,7 @@ def _get_cached_dataset_arff(dataset_id):
                                    "cached" % dataset_id)
 
 
-def list_datasets(offset=None, size=None, tag=None):
+def list_datasets(offset=None, size=None, tag=None, status=None):
     """Return a list of all dataset which are on OpenML.
 
     Parameters
@@ -148,6 +148,10 @@ def list_datasets(offset=None, size=None, tag=None):
         The maximum number of datasets to show.
     tag : str, optional
         Only include datasets matching this tag.
+    status : str, optional
+        Should be {active, in_preparation, deactivated}. By
+        default active datasets are returned, but also datasets
+        from another status can be requested. 
 
     Returns
     -------
@@ -174,6 +178,9 @@ def list_datasets(offset=None, size=None, tag=None):
     if tag is not None:
         api_call += "/tag/%s" % tag
 
+    if status is not None:
+        api_call += "/status/%s" %status
+
     return _list_datasets(api_call)
 
 

openml/utils.py

@@ -118,9 +118,10 @@ def list_all(listing_call, batch_size=10000, *args, **filters):
     dict
     """
     page = 0
+    has_more = 1
     result = {}
 
-    while True:
+    while has_more:
         try:
             new_batch = listing_call(
                 *args,
@@ -135,5 +136,6 @@ def list_all(listing_call, batch_size=10000, *args, **filters):
                 break
         result.update(new_batch)
         page += 1
+        has_more = (len(new_batch) == batch_size)
 
     return result