add 查询Edges 的Source和Target 重复数据

AIDotNet · Aug 13, 2024 · af028f3 · af028f3
1 parent 5816593
commit af028f3
Show file tree

Hide file tree

Showing 3 changed files with 34 additions and 1 deletion.
diff --git a/Directory.Build.props b/Directory.Build.props
@@ -1,7 +1,7 @@
 <Project>
   <!-- See https://aka.ms/dotnet/msbuild/customize for more details on customizing your build -->
   <PropertyGroup>
-    <Version>0.1.14</Version>
+    <Version>0.1.15</Version>
     <SKVersion>1.17.1</SKVersion>  
   </PropertyGroup>
 </Project>
diff --git a/src/GraphRag.Net/Domain/Service/GraphService.cs b/src/GraphRag.Net/Domain/Service/GraphService.cs
@@ -195,6 +195,7 @@ public async Task InsertGraphDataAsync(string index, string input)
                             }
                             if (!_edges_Repositories.IsAny(p => p.Target == relationShip.Edge.Target && p.Source == relationShip.Edge.Source))
                             {
+                                relationShip.Edge.Id=Guid.NewGuid().ToString();
                                 relationShip.Edge.Index = index;
                                 _edges_Repositories.Insert(relationShip.Edge);
                             }
@@ -233,13 +234,40 @@ public async Task InsertGraphDataAsync(string index, string input)
                 {
                     Edges edge = new Edges()
                     {
+                        Id=Guid.NewGuid().ToString(),
                         Index = index,
                         Source = nodeDic[e.Source],
                         Target = nodeDic[e.Target],
                         Relationship = e.Relationship
                     };
                     _edges_Repositories.Insert(edge);
                 }
+
+                //查询Edges 的Source和Target 重复数据
+                var repeatEdges = _edges_Repositories.GetDB().Queryable<Edges>()
+                .GroupBy(p => new { p.Source, p.Target })
+                .Select(p => new { p.Source, p.Target, Count = SqlFunc.AggregateCount(p.Source) })
+                .ToList().Where(p => p.Count > 1).ToList();
+                //合并查询Edges 的Source和Target 重复数据
+                foreach (var edge in repeatEdges)
+                { 
+                    var edges = _edges_Repositories.GetList(p => p.Source == edge.Source && p.Target == edge.Target);
+                    var firstEdge=edges.First();
+
+                    for (int i = 1; i < edges.Count(); i++)
+                    {
+                        if (firstEdge.Relationship == edges[i].Relationship)
+                        {
+                            //相同的边进行合并
+                            _edges_Repositories.Delete(edges[i]);
+                            continue;
+                        }
+                        var newDesc=await _semanticService.MergeDesc(firstEdge.Relationship, edges[i].Relationship);
+                        firstEdge.Relationship= newDesc;
+                        _edges_Repositories.Update(firstEdge);
+                        _edges_Repositories.Delete(edges[i]);
+                    }       
+                }
             }
             catch (Exception ex)
             {

diff --git a/src/GraphRag.Net/Repositories/Graph/Edges/Edges.cs b/src/GraphRag.Net/Repositories/Graph/Edges/Edges.cs
@@ -11,6 +11,11 @@ namespace GraphRag.Net.Repositories
     [SugarIndex("i_edges_target", nameof(Edges.Target), OrderByType.Asc)]
     public class Edges
     {
+        /// <summary>
+        /// 主键
+        /// </summary>
+        [SugarColumn(IsPrimaryKey = true)]
+        public string Id { get; set; }
         /// <summary>
         /// 索引
         /// </summary>