interface for deleting duplicates

src/main/java/com/commafeed/backend/dao/FeedDAO.java

@@ -7,10 +7,15 @@ import javax.ejb.Stateless;
 import javax.persistence.Query;
 import javax.persistence.TypedQuery;
 import javax.persistence.criteria.CriteriaQuery;
+import javax.persistence.criteria.Expression;
 import javax.persistence.criteria.JoinType;
+import javax.persistence.criteria.Path;
 import javax.persistence.criteria.Predicate;
 import javax.persistence.criteria.Root;
 import javax.persistence.criteria.SetJoin;
+import javax.xml.bind.annotation.XmlAccessType;
+import javax.xml.bind.annotation.XmlAccessorType;
+import javax.xml.bind.annotation.XmlRootElement;
 
 import org.apache.commons.codec.digest.DigestUtils;
 import org.apache.commons.lang.StringUtils;
@@ -26,6 +31,13 @@ import com.google.common.collect.Lists;
 @Stateless
 public class FeedDAO extends GenericDAO<Feed> {
 
+	@XmlRootElement
+	@XmlAccessorType(XmlAccessType.FIELD)
+	public static class FeedCount {
+		public String normalizedUrlHash;
+		public List<Feed> feeds;
+	}
+
 	private List<Predicate> getUpdatablePredicates(Root<Feed> root,
 			Date threshold) {
 
@@ -123,4 +135,29 @@ public class FeedDAO extends GenericDAO<Feed> {
 		return deleted;
 
 	}
+
+	public List<FeedCount> findDuplicates(int offset, int limit) {
+		CriteriaQuery<String> query = builder.createQuery(String.class);
+		Root<Feed> root = query.from(getType());
+
+		Path<String> hashPath = root.get(Feed_.normalizedUrlHash);
+		Expression<Long> count = builder.count(hashPath);
+
+		query.select(hashPath);
+
+		query.groupBy(hashPath);
+		query.having(builder.greaterThan(count, 1l));
+
+		TypedQuery<String> q = em.createQuery(query);
+		limit(q, offset, limit);
+		List<String> normalizedUrlHashes = q.getResultList();
+
+		List<FeedCount> result = Lists.newArrayList();
+		for (String hash : normalizedUrlHashes) {
+			FeedCount fc = new FeedCount();
+			fc.normalizedUrlHash = hash;
+			fc.feeds = findByField(Feed_.normalizedUrlHash, hash);
+		}
+		return result;
+	}
 }

src/main/java/com/commafeed/frontend/rest/resources/AdminREST.java

@@ -21,6 +21,7 @@ import com.commafeed.backend.DatabaseCleaner;
 import com.commafeed.backend.MetricsBean;
 import com.commafeed.backend.StartupBean;
 import com.commafeed.backend.dao.FeedDAO;
+import com.commafeed.backend.dao.FeedDAO.FeedCount;
 import com.commafeed.backend.dao.UserDAO;
 import com.commafeed.backend.dao.UserRoleDAO;
 import com.commafeed.backend.feeds.FeedRefreshTaskGiver;
@@ -38,8 +39,8 @@ import com.commafeed.frontend.SecurityCheck;
 import com.commafeed.frontend.model.UserModel;
 import com.commafeed.frontend.model.request.FeedMergeRequest;
 import com.commafeed.frontend.model.request.IDRequest;
-import com.google.api.client.util.Lists;
 import com.google.common.base.Preconditions;
+import com.google.common.collect.Lists;
 import com.google.common.collect.Maps;
 import com.google.common.collect.Sets;
 import com.wordnik.swagger.annotations.Api;
@@ -268,17 +269,36 @@ public class AdminREST extends AbstractResourceREST {
 		return Response.ok(map).build();
 	}
 
+	@Path("/cleanup/findDuplicateFeeds")
+	@GET
+	@ApiOperation(value = "Find duplicate feeds")
+	public Response findDuplicateFeeds(@QueryParam("page") int page,
+			@QueryParam("limit") int limit) {
+		List<FeedCount> list = feedDAO.findDuplicates(limit * page, limit);
+		return Response.ok(list).build();
+	}
+
 	@Path("/cleanup/merge")
 	@POST
 	@ApiOperation(value = "Merge feeds", notes = "Merge feeds together")
-	public Response mergeFeeds(@ApiParam(required = true) FeedMergeRequest request) {
+	public Response mergeFeeds(
+			@ApiParam(required = true) FeedMergeRequest request) {
 		Feed into = feedDAO.findById(request.getIntoFeedId());
+		if (into == null) {
+			return Response.status(Status.BAD_REQUEST)
+					.entity("'into feed' not found").build();
+		}
 
 		List<Feed> feeds = Lists.newArrayList();
 		for (Long feedId : request.getFeedIds()) {
 			Feed feed = feedDAO.findById(feedId);
 			feeds.add(feed);
 		}
+		
+		if (feeds.isEmpty()) {
+			return Response.status(Status.BAD_REQUEST)
+					.entity("'from feeds' empty").build();
+		}
 
 		cleaner.mergeFeeds(into, feeds);
 		return Response.ok().build();

src/main/webapp/js/controllers.js

@@ -1204,6 +1204,34 @@ function($scope, $state, $stateParams,	$dialog, AdminUsersService) {
 	};
 }]);
 
+module.controller('ManageDuplicateFeedsCtrl', [
+	'$scope', 'AdminCleanupService',
+	function($scope, AdminCleanupService) {
+	
+	$scope.limit = 10;
+	$scope.page = 0;
+	$scope.mergeData = {};
+	$scope.refreshData = function() {
+		AdminCleanupService.findDuplicateFeeds({
+			limit : $scope.limit,
+			page : $scope.page
+		}, function(data) {
+			$scope.counts = data;
+		});
+	};
+	
+	$scope.focus = function(count) {
+		$scope.current = count;
+	};
+	
+	$scope.merge = function() {
+		AdminCleanupService.mergeFeeds($scope.mergeData, function() {
+			alert('done!');
+		});
+	};
+	
+}]);
+
 module.controller('SettingsCtrl', ['$scope', '$location', 'SettingsService', 'AnalyticsService', 'ServerService',
 function($scope, $location, SettingsService, AnalyticsService, ServerService) {
 	

src/main/webapp/js/main.js

@@ -91,6 +91,11 @@ app.config([ '$routeProvider', '$stateProvider', '$urlRouterProvider', '$httpPro
 				templateUrl : 'templates/admin.useredit.html',
 				controller : 'ManageUserCtrl'
 			});
+			$stateProvider.state('admin.duplicate_feeds', {
+				url : '/feeds/duplicates/',
+				templateUrl : 'templates/admin.duplicate_feeds.html',
+				controller : 'ManageDuplicateFeedsCtrl'
+			});
 			$stateProvider.state('admin.settings', {
 				url : '/settings',
 				templateUrl : 'templates/admin.settings.html',

src/main/webapp/js/services.js

@@ -272,6 +272,26 @@ module.factory('AdminSettingsService', ['$resource', function($resource) {
 	return res;
 }]);
 
+module.factory('AdminCleanupService', ['$resource', function($resource) {
+	var actions = {
+		findDuplicateFeeds : {
+			method : 'GET',
+			isArray: true,
+			params : {
+				_method : 'findDuplicateFeeds'
+			}
+		},
+		mergeFeeds : {
+			method : 'POST',
+			params : {
+				_method : 'merge'
+			}
+		},
+	};
+	var res = $resource('rest/admin/cleanup/:_method', {}, actions);
+	return res;
+}]);
+
 module.factory('ServerService', ['$resource', function($resource) {
 	var res =  $resource('rest/server/get');
 	return res;

src/main/webapp/templates/admin.duplicate_feeds.html

@@ -0,0 +1,34 @@
+<div class="row">
+	
+	Limit <input type="number" ng-model="limit" />
+	Page <input type="number" ng-model="page" />
+	<input type="button" class="btn" ng-click="refreshData()" value="Refresh" />
+	<table class="table table-condensed table-hover" ui-if="counts">
+		<thead>
+			<tr>
+				<th>url</th>
+				<th>count</h>
+			</tr>
+		</thead>
+		<tbody>
+			<tr ng-repeat="count in counts" ng-click="focus(count)" class="pointer">
+				<td>{{count.normalizedUrlHash}}</td>
+				<td>{{count.feeds.length}}</td>
+			</tr>
+		</tbody>
+	</table>
+	
+	<div ui-if="current">
+		<div>
+			Merge
+			<select ng-model="mergeData.feedIds" size="15" multiple="multiple" 
+				ng-options="feed.id as feed.url for feed in current.feeds">
+			</select>
+			
+			Into
+			<select ng-model="mergeData.intoFeedId" ng-options="feed.id as feed.url for feed in current.feeds">
+			</select>
+		</div>
+		<input type="button" class="btn" ng-click="merge()" value="Merge" />
+	</div>
+</div>