Implement a basic version of data generator

DbSeeder.sln

@@ -4,6 +4,7 @@ Project("{2150E333-8FDC-42A3-9474-1A3956D46DE8}") = "Solution Items", "Solution
 	ProjectSection(SolutionItems) = preProject
 		.gitignore = .gitignore
 	EndProjectSection
+EndProject
 Project("{FAE04EC0-301F-11D3-BF4B-00C04F79EFBC}") = "DbSeeder", "DbSeeder\DbSeeder.csproj", "{AA89F82A-3339-4168-B56C-22F5559700F8}"
 EndProject
 Project("{FAE04EC0-301F-11D3-BF4B-00C04F79EFBC}") = "DbSeeder.Tests", "DbSeeder.Tests\DbSeeder.Tests.csproj", "{4C3EA657-97AD-4E2E-AC54-57F4A05FA8CD}"

DbSeeder/Data/Bogus/BogusGenerator.cs

@@ -0,0 +1,7 @@
+namespace DbSeeder.Data.Bogus;
+
+public record BogusGenerator(
+    string Category,
+    string GeneratorIdentifier,
+    Type ReturnType,
+    Dictionary<string, Type> Params);

DbSeeder/Data/Bogus/BogusUtilities.cs

@@ -0,0 +1,117 @@
+using System.Reflection;
+using Bogus;
+
+namespace DbSeeder.Data.Bogus;
+
+public static class BogusUtilities
+{
+    private static readonly Dictionary<string, HashSet<Type>> AllowedTypes = new()
+    {
+        { "string", [typeof(string), typeof(char?), typeof(char[])] },
+        {
+            "number",
+            [
+                typeof(int?), typeof(float?), typeof(double?), typeof(byte?), typeof(int), typeof(float),
+                typeof(double), typeof(byte)
+            ]
+        }
+    };
+
+    public static Dictionary<string, List<BogusGenerator>> GetBogusGenerators()
+    {
+        var generators = new Dictionary<string, List<BogusGenerator>>();
+
+        var fakerType = typeof(Faker);
+
+        var props = fakerType.GetProperties()
+            .Where(p => p.CustomAttributes.Any(x => x.AttributeType.Name == "RegisterMustasheMethodsAttribute"))
+            .ToList();
+
+        foreach (var p in props)
+        {
+            var generatorCategory = p.Name.ToLower();
+            if (!generators.ContainsKey(generatorCategory))
+            {
+                generators[generatorCategory] = [];
+            }
+
+            var methods = p.PropertyType
+                .GetMethods(BindingFlags.Instance | BindingFlags.Public | BindingFlags.DeclaredOnly)
+                .Where(m => m.IsPublic &&
+                            !(m.Name.StartsWith("get_") || m.Name.StartsWith("set_")))
+                .ToList();
+
+            foreach (var m in methods)
+            {
+                var methodParams = m.GetParameters().ToDictionary(x => x.Name!, x => x.ParameterType);
+                var generator = new BogusGenerator(
+                    generatorCategory,
+                    generatorCategory + m.Name.ToLower(),
+                    m.ReturnType,
+                    methodParams);
+
+                generators[generatorCategory].Add(generator);
+            }
+        }
+
+        return generators;
+    }
+
+    public static Dictionary<string, List<BogusGenerator>> GetFiltersForReturnType(
+        this Dictionary<string, List<BogusGenerator>> src, string returnType)
+    {
+        var allowed = AllowedTypes[returnType];
+
+        var allowedGenerators = new Dictionary<string, List<BogusGenerator>>();
+        foreach (var (category, generators) in src)
+        {
+            allowedGenerators.Add(category, []);
+            foreach (var generator in generators)
+            {
+                if (allowed.Contains(generator.ReturnType))
+                {
+                    allowedGenerators[category].Add(generator);
+                }
+            }
+        }
+
+        return allowedGenerators;
+    }
+
+    public static dynamic? Generate(BogusGenerator generator)
+    {
+        var faker = new Faker();
+        var generationMethod = generator.GeneratorIdentifier[generator.Category.Length..];
+
+        var generatorProperty = faker.GetType().GetProperty(generator.Category,
+            BindingFlags.Instance | BindingFlags.Public | BindingFlags.IgnoreCase);
+        if (generatorProperty != null)
+        {
+            var categoryGenerator = generatorProperty.GetValue(faker);
+            var generatorMethod = categoryGenerator.GetType().GetMethod(generationMethod,
+                BindingFlags.Instance | BindingFlags.Public | BindingFlags.IgnoreCase);
+            if (generatorMethod != null)
+            {
+                var parameters = generatorMethod.GetParameters();
+                if (parameters.Length > 0)
+                {
+                    // TODO[#26]: Implement generators with params
+                    throw new NotImplementedException("This feature is currently not implemented. " +
+                                                      "Only parameterless generators can be used");
+                }
+
+                // TODO[#27]: Implement constraints handling
+                var result = generatorMethod.Invoke(categoryGenerator, null);
+                return Convert.ChangeType(result, generatorMethod.ReturnType)!;
+            }
+
+            Console.WriteLine($"Method '{generationMethod}' not found in '{generator.Category}' category");
+        }
+        else
+        {
+            Console.WriteLine($"Category '{generator.Category}' not found on Faker object");
+        }
+
+        return null;
+    }
+}

DbSeeder/Data/Bogus/DataGeneratorsEvaluator.cs

@@ -0,0 +1,65 @@
+using DbSeeder.Schema;
+
+namespace DbSeeder.Data.Bogus;
+
+public static class DataGeneratorsEvaluator
+{
+    public static List<BogusGenerator> FindBestNGenerators(
+        this Dictionary<string, List<BogusGenerator>> allGenerators,
+        Column column,
+        int n = 1)
+    {
+        var weights = new Dictionary<int, List<BogusGenerator>>();
+
+        foreach (var (_, generatorsCategory) in allGenerators)
+        {
+            foreach (var generator in generatorsCategory)
+            {
+                if (generator.Params.Count != 0)
+                {
+                    // TODO[#26]: Implement generators with params
+                    continue;
+                }
+
+                var weight = CalculateLevenshteinDistance(column.Name, generator.GeneratorIdentifier);
+                if (!weights.ContainsKey(weight))
+                {
+                    weights[weight] = [];
+                }
+
+                weights[weight].Add(generator);
+            }
+        }
+
+        return weights.OrderBy(x => x.Key)
+            .SelectMany(x => x.Value)
+            .Take(n)
+            .ToList();
+    }
+
+    private static int CalculateLevenshteinDistance(string word1, string word2)
+    {
+        var n = word1.Length;
+        var m = word2.Length;
+        var distance = new int[n + 1, m + 1];
+
+        for (var i = 0; i <= n; i++) distance[i, 0] = i;
+        for (var j = 0; j <= m; j++) distance[0, j] = j;
+
+        for (var i = 1; i <= n; i++)
+        {
+            for (var j = 1; j <= m; j++)
+            {
+                var cost = (word2[j - 1] == word1[i - 1]) ? 0 : 1;
+
+                distance[i, j] = Math.Min(
+                    Math.Min(
+                        distance[i - 1, j] + 1, // Deletion
+                        distance[i, j - 1] + 1), // Insertion
+                    distance[i - 1, j - 1] + cost); // Substitution
+            }
+        }
+
+        return distance[n, m];
+    }
+}

DbSeeder/Data/GeneratorFactory.cs

@@ -0,0 +1,57 @@
+using System.Data;
+using DbSeeder.Data.Bogus;
+using DbSeeder.Schema;
+
+namespace DbSeeder.Data;
+
+internal static class GeneratorFactory
+{
+    private static readonly HashSet<string> StringTypes = ["text", "varchar", "nvarchar", "char", "nchar", "ntext"];
+
+    private static readonly HashSet<string> NumeralTypes =
+        ["bigint", "int", "smallint", "tinyint", "bit", "decimal", "money", "smallmoney", "float", "real"];
+
+    private static readonly HashSet<string> BinaryTypes = ["binary", "varbinary", "image"];
+
+    private static readonly HashSet<string> DateTimeTypes =
+        ["datetime", "smalldatetime", "date", "time", "datetime2", "datetimeoffset"];
+
+    private static readonly HashSet<string> OtherTypes =
+            ["uniqueidentifier", "timestamp", "xml", "udt", "structured", "variant"];
+
+    private static readonly Dictionary<string, List<BogusGenerator>> Generators = BogusUtilities.GetBogusGenerators();
+
+    public static object? GetGeneratorByColumnV2(Column col)
+    {
+        // type filter
+        var generators = Generators;
+        if (StringTypes.Contains(col.DataType, StringComparer.OrdinalIgnoreCase))
+        {
+            generators = generators.GetFiltersForReturnType("string");
+        }
+        else if (NumeralTypes.Contains(col.DataType, StringComparer.OrdinalIgnoreCase))
+        {
+            generators = generators.GetFiltersForReturnType("number");
+        }
+
+        // semantic filter
+        var generator = generators.FindBestNGenerators(col, n: 1).First();
+        var generatedValue = (object?)BogusUtilities.Generate(generator);
+        return generatedValue;
+    }
+
+    public static (Type, Func<object>) GetGeneratorByColumn(Column col)
+    {
+        if (StringTypes.Contains(col.DataType, StringComparer.OrdinalIgnoreCase))
+        {
+            return (typeof(string), () => Guid.NewGuid().ToString("N"));
+        }
+
+        if (NumeralTypes.Contains(col.DataType, StringComparer.OrdinalIgnoreCase)) // add type limits check
+        {
+            return (typeof(long), () => Random.Shared.NextInt64() * 17);
+        }
+
+        throw new NotImplementedException($"{col.DataType} is not currently supported");
+    }
+}

DbSeeder/Data/IDataGenerator.cs

@@ -0,0 +1,23 @@
+using DbSeeder.Schema;
+
+namespace DbSeeder.Data;
+
+public interface IDataGenerator<out TOut>
+{
+    /// <summary>
+    /// Generates a single instance of an object that matches the SQL schema's entry for the specified table.
+    /// </summary>
+    /// <param name="table">The table schema for which to generate an instance.</param>
+    /// <returns>An object representing an instance of the SQL schema's entry.</returns>
+    /// <remarks>Any constraints encountered are only applicable within the scope of this method call.</remarks>
+    TOut Generate(Table table);
+
+    /// <summary>
+    /// Generates the requested number of records for the specified table.
+    /// </summary>
+    /// <param name="table">The table schema for which to generate records.</param>
+    /// <param name="count">The number of records to generate.</param>
+    /// <returns>An asynchronous stream of objects representing the generated records.</returns>
+    /// <remarks>Any constraints encountered are only applicable within the scope of this method call.</remarks>
+    IAsyncEnumerable<TOut> GenerateMultiple(Table table, int count);
+}

DbSeeder/Data/Sql/BulkInsertOptions.cs

@@ -0,0 +1,8 @@
+namespace DbSeeder.Data.Sql;
+
+/// <summary>
+/// Represents options for generating bulk insert SQL queries.
+/// </summary>
+/// <param name="TotalRecords">The total number of records to generate.</param>
+/// <param name="RecordsPerInsert">The number of records per insert query.</param>
+public record BulkInsertOptions(int TotalRecords, int RecordsPerInsert);

DbSeeder/Data/Sql/BulkInsertSqlQuery.cs

@@ -0,0 +1,8 @@
+namespace DbSeeder.Data.Sql;
+
+public record BulkInsertSqlQuery(
+    string Table,
+    IReadOnlyList<IReadOnlyDictionary<string, string>> Values) // [col: value]
+{
+
+}

DbSeeder/Data/Sql/ISqlDataGenerator.cs

@@ -0,0 +1,15 @@
+using DbSeeder.Schema;
+
+namespace DbSeeder.Data.Sql;
+
+public interface ISqlDataGenerator : IDataGenerator<InsertSqlQuery>
+{
+    /// <summary>
+    /// Generates bulk insert SQL queries for inserting multiple instances of objects that match the SQL schema's entry for the specified table.
+    /// </summary>
+    /// <param name="table">The table schema for which to generate bulk insert queries.</param>
+    /// <param name="options">Options specifying the total number of records and the number of records per insert.</param>
+    /// <returns>An asynchronous stream of bulk insert SQL queries.</returns>
+    /// <remarks>Any constraints encountered are only applicable within the scope of this method call.</remarks>
+    IAsyncEnumerable<BulkInsertSqlQuery> GenerateBulk(Table table, BulkInsertOptions options);
+}

DbSeeder/Data/Sql/InsertSqlQuery.cs

@@ -0,0 +1,41 @@
+using System.Text;
+
+namespace DbSeeder.Data.Sql;
+
+public record InsertSqlQuery(
+    string Table,
+    IReadOnlyDictionary<string, string> Value,
+    IReadOnlyList<InsertSqlQuery>? RelatedQueries = null)
+{
+    public override string ToString()
+    {
+        var sb = new StringBuilder();
+
+        if (RelatedQueries is not null && RelatedQueries.Any())
+        {
+            foreach (var query in RelatedQueries)
+            {
+                sb.AppendLine(query.ToString());
+            }
+        }
+
+        sb.AppendLine(BuildSqlQuery());
+        return sb.ToString();
+
+        string BuildSqlQuery()
+        {
+            return $"insert into {Table}({GetInsertCols()}) " +
+                   $"value ({GetInsertValues()});";
+        }
+    }
+
+    private string GetInsertCols()
+    {
+        return string.Join(",", Value.Keys);
+    }
+
+    private string GetInsertValues()
+    {
+        return string.Join(",", Value.Values);
+    }
+}