Added configurable request body size limit

Added datapoint bulk insert

src/Server/Controllers/EntityController.cs

@@ -116,12 +116,14 @@ public class EntityController : ControllerBase
             else
             {
                 _logger.LogError("Unable to deserialize an entity");
+                ElmahCore.ElmahExtensions.RaiseError(new Exception("Unable to deserialize an entity"));
                 return Ok(new EntityIndexResult() { Success = false, Message = "Unable to deserialize an entity"});
             }
         } catch (Exception ex)
         {
             if (ex.InnerException is not null) ex = ex.InnerException;
             _logger.LogError("Unable to index the provided entities. {ex.Message} - {ex.StackTrace}", [ex.Message, ex.StackTrace]);
+            ElmahCore.ElmahExtensions.RaiseError(ex);
             return Ok(new EntityIndexResult() { Success = false, Message = ex.Message });
         }
 
@@ -142,6 +144,11 @@ public class EntityController : ControllerBase
         if (entity_ is null)
         {
             _logger.LogError("Unable to delete the entity {entityName} in {searchdomain} - it was not found under the specified name", [entityName, searchdomain]);
+            ElmahCore.ElmahExtensions.RaiseError(
+                new Exception(
+                    $"Unable to delete the entity {entityName} in {searchdomain} - it was not found under the specified name"
+                )
+            );
             return Ok(new EntityDeleteResults() {Success = false, Message = "Entity not found"});
         }
         searchdomain_.ReconciliateOrInvalidateCacheForDeletedEntity(entity_);

src/Server/Datapoint.cs

@@ -1,6 +1,3 @@
-using AdaptiveExpressions;
-using OllamaSharp;
-using OllamaSharp.Models;
 using Shared;
 
 namespace Server;
@@ -80,6 +77,10 @@ public class Datapoint
                     }
                 }
             }
+            if (toBeGenerated.Count == 0)
+            {
+                continue;
+            }
             IEnumerable<float[]> generatedEmbeddings = GenerateEmbeddings([.. toBeGenerated], model, aIProvider, embeddingCache);
             if (generatedEmbeddings.Count() != toBeGenerated.Count)
             {

src/Server/Helper/DatabaseHelper.cs

@@ -1,4 +1,3 @@
-using System.Configuration;
 using System.Data.Common;
 using System.Text;
 using System.Text.Json;
@@ -40,6 +39,19 @@ public class DatabaseHelper(ILogger<DatabaseHelper> logger)
         helper.ExecuteSQLNonQuery(query.ToString(), parameters);
     }
 
+    public static int DatabaseInsertEmbeddingBulk(SQLHelper helper, List<(string hash, string model, byte[] embedding)> data)
+    {
+        return helper.BulkExecuteNonQuery(
+            "INSERT INTO embedding (id_datapoint, model, embedding) SELECT d.id, @model, @embedding FROM datapoint d WHERE d.hash = @hash",
+            data.Select(element => new object[] {
+                new MySqlParameter("@model", element.model),
+                new MySqlParameter("@embedding", element.embedding),
+                new MySqlParameter("@hash", element.hash)
+            })
+        );
+    }
+
+
     public static int DatabaseInsertSearchdomain(SQLHelper helper, string name, SearchdomainSettings settings = new())
     {
         Dictionary<string, dynamic> parameters = new()
@@ -72,6 +84,32 @@ public class DatabaseHelper(ILogger<DatabaseHelper> logger)
         return helper.ExecuteSQLCommandGetInsertedID("INSERT INTO attribute (attribute, value, id_entity) VALUES (@attribute, @value, @id_entity)", parameters);
     }
 
+    public static int DatabaseInsertAttributes(SQLHelper helper, List<(string attribute, string value, int id_entity)> values) //string[] attribute, string value, int id_entity)
+    {
+        return helper.BulkExecuteNonQuery(
+            "INSERT INTO attribute (attribute, value, id_entity) VALUES (@attribute, @value, @id_entity)",
+            values.Select(element => new object[] {
+                new MySqlParameter("@attribute", element.attribute),
+                new MySqlParameter("@value", element.value),
+                new MySqlParameter("@id_entity", element.id_entity)
+            })
+        );
+    }
+
+    public static int DatabaseInsertDatapoints(SQLHelper helper, List<(string name, ProbMethodEnum probmethod_embedding, SimilarityMethodEnum similarityMethod, string hash)> values, int id_entity)
+    {
+        return helper.BulkExecuteNonQuery(
+            "INSERT INTO datapoint (name, probmethod_embedding, similaritymethod, hash, id_entity) VALUES (@name, @probmethod_embedding, @similaritymethod, @hash, @id_entity)",
+            values.Select(element => new object[] {
+                new MySqlParameter("@name", element.name),
+                new MySqlParameter("@probmethod_embedding", element.probmethod_embedding),
+                new MySqlParameter("@similaritymethod", element.similarityMethod),
+                new MySqlParameter("@hash", element.hash),
+                new MySqlParameter("@id_entity", id_entity)
+            })
+        );
+    }
+
     public static int DatabaseInsertDatapoint(SQLHelper helper, string name, ProbMethodEnum probmethod_embedding, SimilarityMethodEnum similarityMethod, string hash, int id_entity)
     {
         Dictionary<string, dynamic> parameters = new()
@@ -144,7 +182,7 @@ public class DatabaseHelper(ILogger<DatabaseHelper> logger)
 
         helper.ExecuteSQLNonQuery("DELETE embedding.* FROM embedding JOIN datapoint dp ON id_datapoint = dp.id JOIN entity ON id_entity = entity.id WHERE entity.id_searchdomain = @searchdomain", parameters);
         helper.ExecuteSQLNonQuery("DELETE datapoint.* FROM datapoint JOIN entity ON id_entity = entity.id WHERE entity.id_searchdomain = @searchdomain", parameters);
-        helper.ExecuteSQLNonQuery("DELETE attribute.* FROM attribute JOIN entity ON id_entity = entity.id WHERE entity.id_searchdomain = @searchdomain", parameters);
+        helper.ExecuteSQLNonQuery("DELETE FROM attribute WHERE id_entity IN (SELECT entity.id FROM entity WHERE id_searchdomain = @searchdomain)", parameters);
         return helper.ExecuteSQLNonQuery("DELETE FROM entity WHERE entity.id_searchdomain = @searchdomain", parameters);
     }
 

src/Server/Helper/SQLHelper.cs

@@ -80,6 +80,33 @@ public class SQLHelper:IDisposable
         }
     }
 
+    public int BulkExecuteNonQuery(string sql, IEnumerable<object[]> parameterSets)
+    {
+        lock (connection)
+        {
+            EnsureConnected();
+            EnsureDbReaderIsClosed();
+
+            using var transaction = connection.BeginTransaction();
+            using var command = connection.CreateCommand();
+
+            command.CommandText = sql;
+            command.Transaction = transaction;
+
+            int affectedRows = 0;
+
+            foreach (var parameters in parameterSets)
+            {
+                command.Parameters.Clear();
+                command.Parameters.AddRange(parameters);
+                affectedRows += command.ExecuteNonQuery();
+            }
+
+            transaction.Commit();
+            return affectedRows;
+        }
+    }
+
     public bool EnsureConnected()
     {
         if (connection.State != System.Data.ConnectionState.Open)

src/Server/Helper/SearchdomainHelper.cs

@@ -245,18 +245,29 @@ public class SearchdomainHelper(ILogger<SearchdomainHelper> logger, DatabaseHelp
         else
         {
             int id_entity = DatabaseHelper.DatabaseInsertEntity(helper, jsonEntity.Name, jsonEntity.Probmethod, _databaseHelper.GetSearchdomainID(helper, jsonEntity.Searchdomain));
+            List<(string attribute, string value, int id_entity)> toBeInsertedAttributes = [];
             foreach (KeyValuePair<string, string> attribute in jsonEntity.Attributes)
             {
-                DatabaseHelper.DatabaseInsertAttribute(helper, attribute.Key, attribute.Value, id_entity); // TODO implement bulk insert to reduce number of queries
+                toBeInsertedAttributes.Add(new() {
+                    attribute = attribute.Key,
+                    value = attribute.Value,
+                    id_entity = id_entity
+                });
             }
+            DatabaseHelper.DatabaseInsertAttributes(helper, toBeInsertedAttributes);
 
             List<Datapoint> datapoints = [];
+            List<(JSONDatapoint datapoint, string hash)> toBeInsertedDatapoints = [];
             foreach (JSONDatapoint jsonDatapoint in jsonEntity.Datapoints)
             {
                 string hash = Convert.ToBase64String(SHA256.HashData(Encoding.UTF8.GetBytes(jsonDatapoint.Text)));
-                Datapoint datapoint = DatabaseInsertDatapointWithEmbeddings(helper, searchdomain, jsonDatapoint, id_entity, hash);
-                datapoints.Add(datapoint);
+                toBeInsertedDatapoints.Add(new()
+                {
+                    datapoint = jsonDatapoint,
+                    hash = hash
+                });
             }
+            List<Datapoint> datapoint = DatabaseInsertDatapointsWithEmbeddings(helper, searchdomain, toBeInsertedDatapoints, id_entity);
             
             var probMethod = Probmethods.GetMethod(jsonEntity.Probmethod) ?? throw new ProbMethodNotFoundException(jsonEntity.Probmethod);
             Entity entity = new(jsonEntity.Attributes, probMethod, jsonEntity.Probmethod.ToString(), datapoints, jsonEntity.Name)
@@ -270,6 +281,38 @@ public class SearchdomainHelper(ILogger<SearchdomainHelper> logger, DatabaseHelp
         }
     }
 
+    public List<Datapoint> DatabaseInsertDatapointsWithEmbeddings(SQLHelper helper, Searchdomain searchdomain, List<(JSONDatapoint datapoint, string hash)> values, int id_entity)
+    {
+        List<Datapoint> result = [];
+        List<(string name, ProbMethodEnum probmethod_embedding, SimilarityMethodEnum similarityMethod, string hash)> toBeInsertedDatapoints = [];
+        List<(string hash, string model, byte[] embedding)> toBeInsertedEmbeddings = [];
+        foreach ((JSONDatapoint datapoint, string hash) value in values)
+        {
+            Datapoint datapoint = BuildDatapointFromJsonDatapoint(value.datapoint, id_entity, searchdomain, value.hash);
+            toBeInsertedDatapoints.Add(new()
+            {
+                name = datapoint.name,
+                probmethod_embedding = datapoint.probMethod.probMethodEnum,
+                similarityMethod = datapoint.similarityMethod.similarityMethodEnum,
+                hash = value.hash
+            });
+            foreach ((string, float[]) embedding in datapoint.embeddings)
+            {
+                toBeInsertedEmbeddings.Add(new()
+                {
+                    hash = value.hash,
+                    model = embedding.Item1,
+                    embedding = BytesFromFloatArray(embedding.Item2)
+                });
+            }
+            result.Add(datapoint);
+        }
+        
+        DatabaseHelper.DatabaseInsertDatapoints(helper, toBeInsertedDatapoints, id_entity);
+        DatabaseHelper.DatabaseInsertEmbeddingBulk(helper, toBeInsertedEmbeddings);
+        return result;
+    }
+
     public Datapoint DatabaseInsertDatapointWithEmbeddings(SQLHelper helper, Searchdomain searchdomain, JSONDatapoint jsonDatapoint, int id_entity, string? hash = null)
     {
         if (jsonDatapoint.Text is null)

src/Server/Program.cs

@@ -35,6 +35,12 @@ EmbeddingSearchOptions configuration = configurationSection.Get<EmbeddingSearchO
 builder.Services.Configure<EmbeddingSearchOptions>(configurationSection);
 builder.Services.Configure<ApiKeyOptions>(configurationSection);
 
+// Configure Kestrel
+builder.WebHost.ConfigureKestrel(options =>
+{
+    options.Limits.MaxRequestBodySize = configuration.MaxRequestBodySize ?? 50 * 1024 * 1024;
+});
+
 // Migrate database
 var helper = new SQLHelper(new MySql.Data.MySqlClient.MySqlConnection(configuration.ConnectionStrings.SQL), configuration.ConnectionStrings.SQL);
 DatabaseMigrations.Migrate(helper);

src/Server/Searchdomain.cs

@@ -219,7 +219,7 @@ public class Searchdomain
 
     public void UpdateModelsInUse()
     {
-        modelsInUse = GetModels([.. entityCache]);
+        modelsInUse = GetModels(entityCache.ToList());
     }
 
     private static float EvaluateEntityAgainstQueryEmbeddings(Entity entity, Dictionary<string, float[]> queryEmbeddings)

src/Server/appsettings.json

@@ -16,5 +16,8 @@
         "Application": "Embeddingsearch.Server"
     }
   },
+  "Embeddingsearch": {
+    "MaxRequestBodySize": 524288000
+  },
   "AllowedHosts": "*"
 }