/// <summary>
/// Given a type and name for a variable, returns whether this appears to be a vector type,
/// and also the associated data type for this type. If a valid data type could not
/// be determined, this will throw.
/// </summary>
/// <param name="name">The name of the variable to inspect.</param>
/// <param name="rawType">The type of the variable to inspect.</param>
/// <param name="attributes">Attribute of <paramref name="rawType"/>. It can be <see langword="null"/> if attributes don't exist.</param>
/// <param name="isVector">Whether this appears to be a vector type.</param>
/// <param name="itemType">
/// The corresponding <see cref="PrimitiveDataViewType"/> RawType of the type, or items of this type if vector.
/// </param>
public static void GetVectorAndItemType(string name, Type rawType, IEnumerable <Attribute> attributes, out bool isVector, out Type itemType)
{
// Determine whether this is a vector, and also determine the raw item type.
isVector = true;
if (rawType.IsArray)
{
itemType = rawType.GetElementType();
}
else if (rawType.IsGenericType && rawType.GetGenericTypeDefinition() == typeof(VBuffer <>))
{
itemType = rawType.GetGenericArguments()[0];
}
else
{
itemType = rawType;
isVector = false;
}
// The internal type of string is ReadOnlyMemory<char>. That is, string will be stored as ReadOnlyMemory<char> in IDataView.
if (itemType == typeof(string))
{
itemType = typeof(ReadOnlyMemory <char>);
}
// Check if the itemType extracted from rawType is supported by ML.NET's type system.
// It must be one of either ML.NET's pre-defined types or custom types registered by the user.
else if (!itemType.TryGetDataKind(out _) && !DataViewTypeManager.Knows(itemType, attributes))
{
throw Contracts.ExceptParam(nameof(rawType), "Could not determine an IDataView type for member {0}", name);
}
}
/// <summary>
/// Given a type and name for a variable, returns whether this appears to be a vector type,
/// and also the associated data type for this type. If a valid data type could not
/// be determined, this will throw.
/// </summary>
/// <param name="name">The name of the variable to inspect.</param>
/// <param name="rawType">The type of the variable to inspect.</param>
/// <param name="attributes">Attribute of <paramref name="rawType"/>. It can be <see langword="null"/> if attributes don't exist.</param>
/// <param name="isVector">Whether this appears to be a vector type.</param>
/// <param name="itemType">
/// The corresponding <see cref="PrimitiveDataViewType"/> RawType of the type, or items of this type if vector.
/// </param>
public static void GetVectorAndItemType(string name, Type rawType, IEnumerable <Attribute> attributes, out bool isVector, out Type itemType)
{
GetMappedType(rawType, out itemType, out isVector);
// Check if the itemType extracted from rawType is supported by ML.NET's type system.
// It must be one of either ML.NET's pre-defined types or custom types registered by the user.
if (!itemType.TryGetDataKind(out _) && !DataViewTypeManager.Knows(itemType, attributes))
{
throw Contracts.ExceptParam(nameof(rawType), "Could not determine an IDataView type and registered custom types for member {0}", name);
}
}
private static void ValidateMemberInfo(MemberInfo memberInfo, IDataView data)
{
if (!SchemaDefinition.NeedToCheckMemberInfo(memberInfo))
{
return;
}
var mappingNameAttr = memberInfo.GetCustomAttribute <ColumnNameAttribute>();
var singleName = mappingNameAttr?.Name ?? memberInfo.Name;
Type actualType = null;
bool isVector = false;
IEnumerable <Attribute> customAttributes = null;
switch (memberInfo)
{
case FieldInfo fieldInfo:
InternalSchemaDefinition.GetMappedType(fieldInfo.FieldType, out actualType, out isVector);
customAttributes = fieldInfo.GetCustomAttributes();
break;
case PropertyInfo propertyInfo:
InternalSchemaDefinition.GetMappedType(propertyInfo.PropertyType, out actualType, out isVector);
customAttributes = propertyInfo.GetCustomAttributes();
break;
default:
Contracts.Assert(false);
throw Contracts.ExceptNotSupp("Expected a FieldInfo or a PropertyInfo");
}
if (!actualType.TryGetDataKind(out _) && !DataViewTypeManager.Knows(actualType, customAttributes))
{
int colIndex;
data.Schema.TryGetColumnIndex(singleName, out colIndex);
DataViewType expectedType = data.Schema[colIndex].Type;
if (!actualType.Equals(expectedType.RawType))
{
throw Contracts.ExceptParam(nameof(actualType), $"The expected type '{expectedType.RawType}' does not match the type of the '{singleName}' member: '{actualType}'. Please change the {singleName} member to '{expectedType.RawType}'");
}
}
}
private Action <TRow> GenerateSetter(DataViewRow input, int index, InternalSchemaDefinition.Column column, Delegate poke, Delegate peek)
{
var colType = input.Schema[index].Type;
var fieldType = column.OutputType;
var genericType = fieldType;
Func <DataViewRow, int, Delegate, Delegate, Action <TRow> > del;
if (fieldType.IsArray)
{
Ch.Assert(colType is VectorDataViewType);
// VBuffer<ReadOnlyMemory<char>> -> String[]
if (fieldType.GetElementType() == typeof(string))
{
Ch.Assert(colType.GetItemType() is TextDataViewType);
return(CreateConvertingVBufferSetter <ReadOnlyMemory <char>, string>(input, index, poke, peek, x => x.ToString()));
}
// VBuffer<T> -> T[]
if (fieldType.GetElementType().IsGenericType&& fieldType.GetElementType().GetGenericTypeDefinition() == typeof(Nullable <>))
{
Ch.Assert(colType.GetItemType().RawType == Nullable.GetUnderlyingType(fieldType.GetElementType()));
}
else
{
Ch.Assert(colType.GetItemType().RawType == fieldType.GetElementType());
}
del = CreateDirectVBufferSetter <int>;
genericType = fieldType.GetElementType();
}
else if (colType is VectorDataViewType vectorType)
{
// VBuffer<T> -> VBuffer<T>
// REVIEW: Do we care about accommodating VBuffer<string> -> VBuffer<ReadOnlyMemory<char>>?
Ch.Assert(fieldType.IsGenericType);
Ch.Assert(fieldType.GetGenericTypeDefinition() == typeof(VBuffer <>));
Ch.Assert(fieldType.GetGenericArguments()[0] == vectorType.ItemType.RawType);
del = CreateVBufferToVBufferSetter <int>;
genericType = vectorType.ItemType.RawType;
}
else if (colType is PrimitiveDataViewType)
{
if (fieldType == typeof(string))
{
// ReadOnlyMemory<char> -> String
Ch.Assert(colType is TextDataViewType);
Ch.Assert(peek == null);
return(CreateConvertingActionSetter <ReadOnlyMemory <char>, string>(input, index, poke, x => x.ToString()));
}
// T -> T
if (fieldType.IsGenericType && fieldType.GetGenericTypeDefinition() == typeof(Nullable <>))
{
Ch.Assert(colType.RawType == Nullable.GetUnderlyingType(fieldType));
}
else
{
Ch.Assert(colType.RawType == fieldType);
}
del = CreateDirectSetter <int>;
}
else if (DataViewTypeManager.Knows(colType))
{
del = CreateDirectSetter <int>;
}
else
{
// REVIEW: Is this even possible?
throw Ch.ExceptNotImpl("Type '{0}' is not yet supported.", column.OutputType.FullName);
}
MethodInfo meth = del.GetMethodInfo().GetGenericMethodDefinition().MakeGenericMethod(genericType);
return((Action <TRow>)meth.Invoke(this, new object[] { input, index, poke, peek }));
}
