Input: sim: environment of the fluid pre_velocity: numpy array, [1,128,128,2] c1: numpy array, [nt,nx,nx] c2: num

Function: get initial state

Gets target data and returns a function that computes guidance loss

Use explicit loss for guided inference in diffusion. J is the loss here, not Jacobian. Arguments: loss_fn: callable, calculates the l

Function: Calculate vague velocity Input: xs: random x-position for turn ys: random y-position for turn Output: v

Get the string of the current local time.

Function: Generate velocitys or control gif saved at gif_dir Input: velocitys: numpy array, [256,128,128,2]

Function: Generate velocitys or control gif saved at gif_dir Input: velocitys: numpy array, [256,128,128,2]

Calculates the gradient of a scalar field from finite differences. The gradient vectors are in reverse order, lowest dimension first. :param tens

Function: initialize fluid field Output: sim: FluidSimulation Object

Function: Intialize density field Input: xs: x-postion list ys: y-postion list Output: array: numpy array density

Function: Initialize velocity field Input: vx, vy: float velocity-x, velocity-y Output: init_op_velocity: StaggeredGrid v

Function: Write loop

Arguments: amp_compensate: Gaussian in the time domain decreases average amp, so we need to compensate it t: number of t

Make directory using filename if the directory does not exist

Multiplies the target so that its total content matches the source. :param target: a tensor :param source: a tensor or number :return: no