pathikrit

val n = 9
val s = Math.sqrt(n).toInt
type Board = IndexedSeq[IndexedSeq[Int]]
 
def solve(board: Board, cell: Int = 0): Option[Board] = (cell%n, cell/n) match {
  case (r, `n`) => Some(board)
  case (r, c) if board(r)(c) > 0 => solve(board, cell + 1)
  case (r, c) =>
    def guess(x: Int) = solve(board.updated(r, board(r).updated(c, x)), cell + 1)  
    val used = board.indices.flatMap(i => Seq(board(r)(i), board(i)(c), board(s*(r/s) + i/s)(s*(c/s) + i%s)))

noelwelsh

object Nullable {
  sealed trait NullableTag
  type Nullable[A] = A with NullableTag

  def nullAs[B]: Nullable[B] =
    null.asInstanceOf[Nullable[B]]

  implicit class ToNullable[A](val a: A) extends AnyVal {
    def `?`: Nullable[A] = a.asInstanceOf[Nullable[A]]
  }

gyglim

"""Simple example on how to log scalars and images to tensorboard without tensor ops.

License: BSD License 2.0
"""
__author__ = "Michael Gygli"

import tensorflow as tf
from StringIO import StringIO
import matplotlib.pyplot as plt
import numpy as np