我正在尝试为case类编写一个简单的宏注释,它为随播对象添加了一个方法。问题是新方法必须考虑带注释的案例类的类型参数。
这是我需要传递的测试
package my.macros
import org.scalatest._
class DefaultApplyTest extends FlatSpec with Matchers {
@defaultApply case class Generic[A, B](a: A, b: B)
it should "define defaultApply method in companion object" in {
assert(Generic.defaultApply("str", 1) == Generic("str", 1))
}
}
以下是我为完成此操作而编写的代码
package my.macros
import scala.reflect.macros._
import scala.language.experimental.macros
import scala.annotation.StaticAnnotation
class defaultApply extends StaticAnnotation {
def macroTransform(annottees: Any*): Any = macro DefaultApply.impl
}
object DefaultApply {
def impl(c: blackbox.Context)(annottees: c.Expr[Any]*): c.Expr[Any] = {
import c.universe._
def defaultApplyCompanion(classDecl: ClassDef) = {
val (name, typeParams, valueParams) = try {
val q"case class ${name: TypeName}[..${typeParams: Seq[TypeDef]}](..${valueParams: Seq[ValDef]}) extends ..$bases { ..$body }" = classDecl
(name, typeParams, valueParams)
} catch {
case e: MatchError =>
c.warning(c.enclosingPosition, e.toString)
c.abort(c.enclosingPosition, "Annotation is only supported on case class")
}
val applyDef = q"""${name.toTermName}.apply[..$typeParams]"""
c.warning(c.enclosingPosition, showRaw(applyDef))
q"""
object ${name.toTermName} {
def defaultApply: (..${valueParams.map(_.tpt)}) => $name[..$typeParams] = $applyDef
}
"""
}
def modifiedDeclaration(classDecl: ClassDef) = {
val compDecl = defaultApplyCompanion(classDecl)
c.Expr(q"""
$classDecl
$compDecl
""")
}
annottees.map(_.tree) match {
case (classDecl: ClassDef) :: Nil => modifiedDeclaration(classDecl)
case _ => c.abort(c.enclosingPosition, "Invalid annottee")
}
}
}
我理解的问题是,当我尝试lift结果语法树中的类型参数列表时,它们不会被识别为与原始树相同的类型参数。
所以我关注的是宏的这部分
val applyDef = q"""${name.toTermName}.apply[..$typeParams]"""
c.warning(c.enclosingPosition, showRaw(applyDef))
原始语法树以
的形式发出TypeApply(Select(Ident(TermName("Generic")), TermName("apply")), List(TypeDef(Modifiers(PARAM), TypeName("A"), List(), TypeBoundsTree(EmptyTree, EmptyTree)), TypeDef(Modifiers(PARAM), TypeName("B"), List(), TypeBoundsTree(EmptyTree, EmptyTree))))
但编译器对此
不满意type arguments [<notype>,<notype>] do not conform to method apply's type parameter bounds [A,B]
最终用例是生成可缓存类型类的实例,该类触及超过1k行代码。非参数化版本已经有效,这只是锦上添花。我不明白scalac的东西,但我想。花在阅读本文上的时间非常感谢。
我正在使用Scala 2.11.8和macro paradise 2.1.0
答案 0 :(得分:0)
问题似乎是您在类型参数的位置使用类型参数。这似乎有效(我还必须将类型参数添加到In [153]: df = df.groupby('A', as_index=False)['B1'].agg(['sum','count']).reset_index()
In [154]: df = pd.merge(df3, df, on='A', how='left').rename(columns={'B1':'Sum'}).fillna('')
In [155]: df
Out[155]:
A B sum count
0 Erick Yes
1 John Yes 18 3
2 David No
3 Mandy Yes 29 2
4 Jared No
5 Lisa No
方法声明中):
defaultApply