Merge pull request #3 from ContinuumIO/new_intake

updates for intake 0.2

conda/meta.yaml

@@ -15,7 +15,7 @@ requirements:
     - python
     - jinja2
   run:
-    - intake
+    - intake>=0.2
     - python
     - requests
 

docs/source/api.rst

@@ -2,11 +2,7 @@ API Reference
 =============
 
 .. autosummary::
-   intake_splunk.Plugin
    intake_splunk.core.SplunkSource
 
-.. autoclass:: intake_splunk.Plugin
-   :members:
-
 .. autoclass:: intake_splunk.core.SplunkSource
    :members:

docs/source/conf.py

@@ -59,9 +59,10 @@
 # built documents.
 #
 # The short X.Y version.
-version = u'0.0.2'
+import intake_splunk
+version = intake_splunk.__version__
 # The full version, including alpha/beta/rc tags.
-release = u'0.0.2'
+release = intake_splunk.__version__
 
 # The language for content autogenerated by Sphinx. Refer to documentation
 # for a list of supported languages.

docs/source/index.rst

@@ -1,6 +1,9 @@
 Welcome to intake-splunk's documentation!
 ================================================
 
+This package accesses tabular data in Splunk, and can be used by Intake to
+load that into pandas dataframes.
+
 .. toctree::
    :maxdepth: 2
    :caption: Contents:

docs/source/quickstart.rst

@@ -38,13 +38,7 @@ Three parameters are of interest when defining a data source:
 Creating Catalog Entries
 ~~~~~~~~~~~~~~~~~~~~~~~~
 
-To include in a catalog, the plugin must be listed in the plugins of the catalog::
-
-   plugins:
-     source:
-       - module: intake_splunk
-
-and entries must specify ``driver: splunk``.
+To use, catalog entries must specify ``driver: splunk``.
 
 
 

intake_splunk/__init__.py

@@ -1,29 +1,5 @@
-from intake.source import base
 from ._version import get_versions
 __version__ = get_versions()['version']
 del get_versions
 
-
-class Plugin(base.Plugin):
-    """Plugin for Splunk reader"""
-
-    def __init__(self):
-        super(Plugin, self).__init__(name='splunk',
-                                     version=__version__,
-                                     container='dataframe',
-                                     partition_access=True)
-
-    def open(self, query, url, auth, chunksize=5000, **kwargs):
-        """
-        Create SplunkSource instance
-
-        Parameters
-        ----------
-        query, url, auth, chunksize
-            See ``SplunkSource``.
-        """
-        from intake_splunk.core import SplunkSource
-        base_kwargs, source_kwargs = self.separate_base_kwargs(kwargs)
-        return SplunkSource(query, url, auth,
-                            metadata=base_kwargs['metadata'],
-                            **source_kwargs)
+from .core import SplunkSource

intake_splunk/core.py

@@ -9,6 +9,7 @@
 
 # because Splunk connections are against a self-signed cert, all connections
 # would raise a warning
+from . import __version__
 warnings.filterwarnings('ignore', module='urllib3.connectionpool')
 
 
@@ -28,24 +29,28 @@ class SplunkSource(base.DataSource):
 
     """
     container = 'dataframe'
+    version = __version__
+    name = 'splunk'
+    partition_access = True
 
     def __init__(self, query, url, auth, chunksize=5000,
                  metadata=None):
-        self.splunk = SplunkConnect(url)
-        if isinstance(auth, (tuple, list)):
-            self.splunk.auth(*auth)
-        else:
-            self.splunk.auth_head(key=auth)
+        self.url = url
+        self.auth = auth
         self.query = query
         self.chunksize = chunksize
         self._df = None
-        super(SplunkSource, self).__init__(container=self.container,
-                                           metadata=metadata)
+        super(SplunkSource, self).__init__(metadata=metadata)
 
     def _get_schema(self):
         if self._df is None:
             # this waits until query is ready, but Splunk has results_preview
             # end-point which can be fetched while query is running
+            self.splunk = SplunkConnect(self.url)
+            if isinstance(self.auth, (tuple, list)):
+                self.splunk.auth(*self.auth)
+            else:
+                self.splunk.auth_head(key=self.auth)
             self._df = self.splunk.read_dask(self.query, self.chunksize)
         self.npartitions = self._df.npartitions
         return base.Schema(datashape=None,