Use yield return/yield break for performance

Changed public API to use IEnumerable and yield return/yield break so that users do not have to load entire database query results to find matches. This should improve performance of matching.

Fixes #19

Author: TechnikEmpire

DistillNET/DistillNET/DistillNET.csproj

@@ -9,12 +9,12 @@
     <RepositoryUrl>https://github.com/TechnikEmpire/DistillNET</RepositoryUrl>
     <Description>DistillNET is a library for matching and filtering HTTP requests URLs using the Adblock Plus Filter format.</Description>
     <Copyright>Copyright © 2017 - 2018 Jesse Nicholson</Copyright>
-    <Version>1.5.1</Version>
+    <Version>1.6.0</Version>
     <Authors>Jesse Nicholson</Authors>
     <Company>Technik Empire</Company>
     <PackageTags>DistillNET Adblock AdblockPlus Adblock-Plus URL-Filter URL-Filtering Content-Filter Filter</PackageTags>
-    <PackageReleaseNotes>Rule loading is ~twice as fast as previous version now thanks to prepared statements.</PackageReleaseNotes>
-    <AssemblyVersion>1.5.1.0</AssemblyVersion>
+    <PackageReleaseNotes>Changed public API to use IEnumerable and yield return/yield break so that users do not have to load entire database query results to find matches. This should improve performance of matching.</PackageReleaseNotes>
+    <AssemblyVersion>1.6.0.0</AssemblyVersion>
   </PropertyGroup>
 
   <PropertyGroup Condition="'$(Configuration)|$(Platform)'=='Release|AnyCPU'">

DistillNET/DistillNET/DistillNET/FilterDbCollection.cs

@@ -424,7 +424,7 @@ public Tuple<int, int> ParseStoreRulesFromStream(Stream rawRulesStream, short ca
         /// <returns>
         /// A list of all compiled blacklisting URL filters for the given domain.
         /// </returns>
-        public List<UrlFilter> GetFiltersForDomain(string domain = "global")
+        public IEnumerable<UrlFilter> GetFiltersForDomain(string domain = "global")
         {
             return GetFiltersForDomain(domain, false);
         }
@@ -439,7 +439,7 @@ public List<UrlFilter> GetFiltersForDomain(string domain = "global")
         /// <returns>
         /// A list of all compiled whitelisting URL filters for the given domain.
         /// </returns>
-        public List<UrlFilter> GetWhitelistFiltersForDomain(string domain = "global")
+        public IEnumerable<UrlFilter> GetWhitelistFiltersForDomain(string domain = "global")
         {
             return GetFiltersForDomain(domain, true);
         }
@@ -456,14 +456,19 @@ public List<UrlFilter> GetWhitelistFiltersForDomain(string domain = "global")
         /// <returns>
         /// A list of either all whitelist or all blacklist filters for the given domain.
         /// </returns>
-        private List<UrlFilter> GetFiltersForDomain(string domain, bool isWhitelist)
+        private IEnumerable<UrlFilter> GetFiltersForDomain(string domain, bool isWhitelist)
         {
             var cacheKey = new Tuple<string, bool>(domain, isWhitelist);
 
 
             if (m_cache.TryGetValue(cacheKey, out List<UrlFilter> retVal))
             {
-                return retVal;
+                foreach (var elm in retVal)
+                {
+                    yield return elm;
+                }
+
+                yield break;
             }
 
             retVal = new List<UrlFilter>();
@@ -505,7 +510,9 @@ private List<UrlFilter> GetFiltersForDomain(string domain, bool isWhitelist)
                             while(reader.Read())
                             {
                                 short catId = reader.GetInt16(1);
-                                retVal.Add((UrlFilter)m_ruleParser.ParseAbpFormattedRule(reader.GetString(3), catId));
+                                var newRule = (UrlFilter)m_ruleParser.ParseAbpFormattedRule(reader.GetString(3), catId);
+                                retVal.Add(newRule);
+                                yield return newRule;                                
                             }
                         }
                     }
@@ -514,7 +521,7 @@ private List<UrlFilter> GetFiltersForDomain(string domain, bool isWhitelist)
 
             m_cache.Set(cacheKey, retVal);
 
-            return retVal;
+            yield break;
         }
 
         private List<string> GetAllPossibleSubdomains(string inputDomain)
@@ -532,7 +539,7 @@ private List<string> GetAllPossibleSubdomains(string inputDomain)
             return retVal;
         }
 
-        public List<Filter> GetFiltersForRequest(Uri requestString, string referer = "")
+        public IEnumerable<Filter> GetFiltersForRequest(Uri requestString, string referer = "")
         {
             return null;
         }

DistillNET/Tests/Program.cs

@@ -10,6 +10,7 @@
 using System.Collections.Specialized;
 using System.Diagnostics;
 using System.IO;
+using System.Linq;
 
 namespace DistillNET
 {
@@ -127,7 +128,7 @@ private static void Main(string[] args)
             sw.Restart();
             for(int i = 0; i < 1000; ++i)
             {
-                loadedFilters += filterCollection.GetFiltersForDomain().Count;
+                loadedFilters += filterCollection.GetFiltersForDomain().Count();
             }
             sw.Stop();