[ckan harvester] Support for default tags and groups

README.rst

@@ -113,6 +113,30 @@ to add the `ckan_harvester` plugin to your options file:
 
 After adding it, a 'CKAN' option should appear in the 'New harvest source' form.
 
+The CKAN harvesters support a number of configuration options to control their
+behaviour. Those need to defined as a JSON object in the configuration form
+field. The currently supported configuration options are:
+
+* api_version: You can force the harvester to use eithoer version '1' or
+    '2' of the CKAN API. Default is '2'.
+
+* default_tags: A list of tags that will be added to all harvested datasets.
+    Tags don't need to previously exist.
+
+* default_groups: A list of groups to which the harvested datasets will be
+    added to. The groups must exist. Note that you must use ids or names to
+    define the groups according to the API version you defined (names for
+    version '1', ids for version '2')
+
+Here is an example of a configuration object (the one that must be entered in
+the configuration field)::
+
+    {
+     "api_version":"1",
+     "default_tags":["new-tag-1","new-tag-2"],
+     "default_groups":["my-own-group"]
+    }
+
 
 The harvesting interface
 ========================

ckanext/harvest/harvesters/base.py

@@ -106,11 +106,17 @@ class HarvesterBase(SingletonPlugin):
             schema = default_package_schema()
             schema["id"] = [ignore_missing, unicode]
 
+            # Check API version
+            if self.config:
+                api_version = self.config.get('api_version','2')
+            else:
+                api_verion = '2'
+
             context = {
                 'model': model,
                 'session': Session,
                 'user': u'harvest',
-                'api_version':'2',
+                'api_version': api_version,
                 'schema': schema,
             }
 

ckanext/harvest/harvesters/ckanharvester.py

@@ -1,7 +1,9 @@
 import urllib2
 
+from ckan.lib.base import c
+from ckan import model
 from ckan.model import Session, Package
-from ckan.logic import ValidationError, NotFound
+from ckan.logic import ValidationError, NotFound, get_action
 from ckan.lib.helpers import json
 
 from ckanext.harvest.model import HarvestJob, HarvestObject, HarvestGatherError, \
@@ -65,6 +67,16 @@ class CKANHarvester(HarvesterBase):
 
         try:
             config_obj = json.loads(config)
+
+            if 'default_groups' in config_obj:
+                # Check if default groups exist
+                context = {'model':model,'user':c.user}
+                for group_name in config_obj['default_groups']:
+                    try:
+                        group = get_action('group_show')(context,{'id':group_name})
+                    except NotFound,e:
+                        raise ValueError('Default group not found')
+
         except ValueError,e:
             raise e
 
@@ -196,6 +208,24 @@ class CKANHarvester(HarvesterBase):
 
         try:
             package_dict = json.loads(harvest_object.content)
+
+            # Set default tags if needed
+            default_tags = self.config.get('default_tags',[])
+            if default_tags:
+                if not 'tags' in package_dict:
+                    package_dict['tags'] = []
+                package_dict['tags'].extend([t for t in default_tags if t not in package_dict['tags']])
+
+            # Ignore remote groups for the time being
+            del package_dict['groups']
+
+            # Set default groups if needed
+            default_groups = self.config.get('default_groups',[])
+            if default_groups:
+                if not 'groups' in package_dict:
+                    package_dict['groups'] = []
+                package_dict['groups'].extend([g for g in default_groups if g not in package_dict['groups']])
+
             return self._create_or_update_package(package_dict,harvest_object)
         except ValidationError,e:
             self._save_object_error('Invalid package with GUID %s: %r' % (harvest_object.guid, e.error_dict),