RL_dev_components_check.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "b0842814",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<style>\n",
       ".container{width:90%}\n",
       "<style>\n"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%html\n",
    "<style>\n",
    ".container{width:90%}\n",
    "<style>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "0f5514a8",
   "metadata": {},
   "outputs": [],
   "source": [
    "import RL_samp\n",
    "from RL_samp.header import *\n",
    "from RL_samp.utils import *\n",
    "from RL_samp.replay_buffer import *\n",
    "from RL_samp.models import poly_net, val_net\n",
    "from RL_samp.reconstructors import sigpy_solver\n",
    "from RL_samp.policies import DQN\n",
    "from RL_samp.trainers import DeepQL_trainer, AC1_ET_trainer\n",
    "\n",
    "from importlib import reload\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "import torch.nn.functional as Func\n",
    "\n",
    "def rolling_mean(x,window):\n",
    "    window = int(window)\n",
    "\n",
    "    # Stephen: for large data, the above gets a bit slow, so we can do this:\n",
    "#   y = np.convolve(x, np.ones(window)/window, mode='valid')\n",
    "#   return y\n",
    "    # or https://stackoverflow.com/a/27681394\n",
    "    cumsum = np.cumsum(np.insert(x, 0, 0))\n",
    "    return (cumsum[window:] - cumsum[:-window]) / float(window)\n",
    "from importlib import reload\n",
    "import os\n",
    "heg = 192\n",
    "wid = 144"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3f969222",
   "metadata": {},
   "source": [
    "## data loader"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "c7e177fb",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "current file: fs_0045_3T.pt\n",
      "Dimension of the current data file: t_ubd 25, slice_ubd 10, rep_ubd 1\n"
     ]
    }
   ],
   "source": [
    "datapath = '/home/ec2-user/SageMaker/data/OCMR_fully_sampled_images/'\n",
    "ncfiles = list([])\n",
    "for file in os.listdir(datapath):\n",
    "    if file.endswith(\".pt\") and file.startswith('fs'):\n",
    "        ncfiles.append(file)\n",
    "loader = ocmrLoader(ncfiles,datapath=datapath,batch_size=1)\n",
    "loader.reset()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "bbec43cf",
   "metadata": {},
   "outputs": [],
   "source": [
    "torch.load(datapath + ncfiles[0]).shape"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "83aa10f4",
   "metadata": {},
   "source": [
    "## reconstructor check"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "e4fe9f98",
   "metadata": {},
   "outputs": [],
   "source": [
    "source, target = loader.load()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "721c57fd",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure()\n",
    "plt.imshow(target[0,0,:,:],origin='lower')\n",
    "plt.colorbar()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "a0b06eaa",
   "metadata": {},
   "outputs": [],
   "source": [
    "mask = mask_prob(target, fix=8, other=16, fft_input=False, roll=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "9cd52c8c",
   "metadata": {},
   "outputs": [],
   "source": [
    "def kplot(y,roll=False,log=False,cmap=None,flip=True,img_name=None):\n",
    "    '''\n",
    "    This function plots the reconstructed image.\n",
    "    -- Options:\n",
    "    roll: whether to move low frequencies in the middle of the Image\n",
    "    log : whether to show the image in log scale\n",
    "    cmap: color map choice for the plot\n",
    "    '''\n",
    "    if isinstance(y,torch.Tensor):\n",
    "        y = y.numpy()\n",
    "    yshape = y.shape\n",
    "    if len(yshape)==3: # one complex-valued image\n",
    "        if yshape[-1]==1 and np.iscomplex(y).any(): # complex number input\n",
    "            ynew = np.zeros((yshape[0],yshape[1],2))\n",
    "            ynew[:,:,0] = np.real(y[:,:,0])\n",
    "            ynew[:,:,1] = np.imag(y[:,:,0])\n",
    "            y           = ynew\n",
    "            yshape      = y.shape\n",
    "        if log:\n",
    "            im1 = np.log( np.absolute(y[:,:,0]) )\n",
    "            if yshape[-1] == 2:\n",
    "                im2 = np.log( np.absolute(y[:,:,1]) )\n",
    "        else:\n",
    "            im1 = y[:,:,0]\n",
    "            if yshape[-1] == 2:\n",
    "                im2 = y[:,:,1]\n",
    "        if roll:\n",
    "            im1 = np.roll(im1,tuple(n//2 for n in im1.shape[:2]), axis=(0,1))  # move the center frequency to the center of the image\n",
    "            if yshape[-1] == 2:\n",
    "                im2 = np.roll(im2,tuple(n//2 for n in im2.shape[:2]), axis=(0,1))\n",
    "        if yshape[-1] == 2:\n",
    "            fig,axs = plt.subplots(2,1,figsize=(15,10))\n",
    "            hd1 = axs[0].imshow(im1,cmap=cmap,origin='lower')\n",
    "            axs[0].set_title('Real part mag')\n",
    "            plt.colorbar(hd1,ax=axs[0])\n",
    "            hd2 = axs[1].imshow(im2,cmap=cmap,origin='lower')\n",
    "            axs[1].set_title('Imag part mag')\n",
    "            fig.colorbar(hd2,ax=axs[1])\n",
    "            plt.show()\n",
    "        else:\n",
    "            fig,axs = plt.subplots(1,1,figsize=(5,5))\n",
    "            hd1 = axs.imshow(im1,cmap=cmap,origin='lower')\n",
    "            # axs.set_title('Real part mag')\n",
    "            plt.colorbar(hd1,ax=axs)\n",
    "            plt.show()\n",
    "        if img_name is not None:\n",
    "            axs.set_title(img_name)\n",
    "    elif len(yshape)==2: # one real-valued image\n",
    "        fig,axs = plt.subplots(1,1,figsize=(10, 10))\n",
    "        if roll:\n",
    "            y = np.roll(y,tuple(n//2 for n in y.shape[:2]), axis=(0,1))  # move the center frequency to the center of the image\n",
    "        if not log:\n",
    "            try:\n",
    "                hd1 = axs.imshow(y,cmap=cmap,origin='lower')\n",
    "            except TypeError:\n",
    "                hd1 = axs.imshow(np.abs(y),cmap=cmap,origin='lower')\n",
    "        else:\n",
    "            try:\n",
    "                hd1 = axs.imshow(np.log(y),cmap=cmap,origin='lower')\n",
    "            except TypeError:\n",
    "                hd1 = axs.imshow(np.log(np.abs(y)),cmap=cmap,origin='lower')\n",
    "        if img_name is not None:\n",
    "            axs.set_title(img_name)\n",
    "        plt.colorbar(hd1,ax=axs)\n",
    "        plt.show()\n",
    "    elif len(yshape)==1: # mask\n",
    "        fig,axs = plt.subplots(1,1,figsize=(5, 5))\n",
    "        if roll:\n",
    "            y = np.roll(y,yshape[0]//2,axis=0)\n",
    "        mask    = np.reshape(y,(-1,1))\n",
    "        mask2D  = np.tile(mask,(1,mask.size))\n",
    "        if flip:\n",
    "            hd1 = axs.imshow(1-mask2D,cmap='Greys',origin='lower')\n",
    "        else:\n",
    "            hd1 = axs.imshow(mask2D,cmap='Greys',origin='lower')\n",
    "        axs.set_xticks([])\n",
    "        axs.set_xticks([], minor=True) # for minor ticks\n",
    "        axs.set_ylabel('Frequencies')\n",
    "        if not flip:\n",
    "            plt.colorbar(hd1,ax=axs)\n",
    "        if img_name is not None:\n",
    "            axs.set_title(img_name)\n",
    "        plt.rcParams.update({'font.size': 25})\n",
    "        plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "26bf1c4b",
   "metadata": {},
   "outputs": [],
   "source": [
    "def mask_prob(imgs,fix=10,other=30,roll=True,fft_input=True):\n",
    "    '''\n",
    "    input imgs: [NCHW]\n",
    "    '''\n",
    "    fix   = int(fix)\n",
    "    other = int(other)\n",
    "    [N,C,_,sampdim] = imgs.shape\n",
    "    if not fft_input:\n",
    "        y = F.fftn(imgs,dim=(2,3),norm='ortho')\n",
    "    p = torch.sum(torch.abs(y),axis=(0,1,2))/(N*C)\n",
    "    fixInds  = np.concatenate((np.arange(0,round(fix//2) ),np.arange(sampdim-1,sampdim-1-round(fix/2),-1)))\n",
    "    p[fixInds] = 0\n",
    "    p = p/p.sum() # normalize probability vector\n",
    "    addInds  = np.random.choice(np.arange(sampdim),size=other,replace=False,p=p.numpy())\n",
    "    maskInds = np.concatenate((fixInds,addInds))\n",
    "    mask     = np.zeros(sampdim)\n",
    "    mask[maskInds]= 1\n",
    "    if roll:\n",
    "        mask = np.roll(mask,shift=sampdim//2,axis=0)\n",
    "    return mask"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "a6d730a1",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "cda60336",
   "metadata": {},
   "outputs": [],
   "source": [
    "mask = mask_naiveRand(144,fix=10,other=26,roll=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "33d72517",
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "fft_recon_img = fft_observe(target, mask, return_opt='img')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "7dcf2778",
   "metadata": {},
   "outputs": [],
   "source": [
    "plt.figure()\n",
    "plt.imshow(torch.abs(fft_recon_img[0,0,:,:]),origin='lower')\n",
    "error = torch.norm(torch.abs(fft_recon_img[0,0,:,:]) - target[0,0,:,:])/torch.norm(target[0,0,:,:]) \n",
    "plt.colorbar()\n",
    "plt.title(f'Rel. L2 error: {error*100:.2f}%')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "54359f13",
   "metadata": {},
   "outputs": [],
   "source": [
    "from unet.unet_model_fbr import Unet\n",
    "unet = Unet(in_chans=2,out_chans=1,chans=64,num_pool_layers=6,drop_prob=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "d439bb7a",
   "metadata": {},
   "outputs": [],
   "source": [
    "unet_path = '/home/ec2-user/SageMaker/RLsamp/output/recon_models/unet_fbr_2_chans_64base10_budget26.pt'\n",
    "checkpoint = torch.load(unet_path)\n",
    "unet.load_state_dict(checkpoint['model_state_dict'])\n",
    "print('Unet loaded successfully from: ' + unet_path )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "0ee954aa",
   "metadata": {},
   "outputs": [],
   "source": [
    "fft_recon_img = fft_observe(target, mask, return_opt='img', abs_opt=False)\n",
    "fft_recon_img.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "990511aa",
   "metadata": {},
   "outputs": [],
   "source": [
    "unet_recon = unet(fft_recon_img).detach().cpu()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "e68c31d7",
   "metadata": {},
   "outputs": [],
   "source": [
    "plt.figure()\n",
    "plt.imshow(unet_recon[0,0,:,:],origin='lower')\n",
    "error = torch.norm(torch.abs(unet_recon[0,0,:,:]) - target[0,0,:,:])/torch.norm(target[0,0,:,:]) \n",
    "plt.colorbar()\n",
    "plt.title(f'Rel. L2 error: {error*100:.2f}%')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "b01a54d3",
   "metadata": {},
   "outputs": [],
   "source": [
    "recon_fft = torch.abs(F.ifftn(F.ifftshift(curr_obs,dim=(2,3)),dim=(2,3),norm='ortho'))\n",
    "# recon_fft = fft_observe(target, mask, return_opt='img')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "7e3c7fed",
   "metadata": {},
   "outputs": [],
   "source": [
    "plt.figure()\n",
    "recon_error = torch.norm(recon_fft - target)/torch.norm(target)\n",
    "plt.title(f'recon L2 error is {recon_error*100:.2f}%')\n",
    "plt.imshow(recon_fft[0,0,:,:],origin='lower')\n",
    "plt.colorbar()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "298a57f9",
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "recon_obs = sigpy_solver(curr_obs,max_iter=50, L=5e-3)\n",
    "plt.figure()\n",
    "recon_error = torch.norm(recon_obs - target)/torch.norm(target)\n",
    "plt.title(f'recon L2 error is {recon_error*100:.2f}%')\n",
    "\n",
    "plt.imshow(torch.abs(recon_obs[0,0,:,:]),origin='lower')\n",
    "plt.colorbar()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e50eb527",
   "metadata": {},
   "source": [
    "## Net and reconstructor preparation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "6304ae52",
   "metadata": {},
   "outputs": [],
   "source": [
    "import models\n",
    "reload(models)\n",
    "from models import poly_net"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "42dfd140",
   "metadata": {},
   "outputs": [],
   "source": [
    "net = poly_net(samp_dim=144)\n",
    "net"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "887ad8f5",
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "img = torch.rand(2,3,193,144)\n",
    "net(img).shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "b8cc6b8a",
   "metadata": {},
   "outputs": [],
   "source": [
    "plt.figure()\n",
    "plt.imshow(img[0][0],origin='lower')\n",
    "plt.colorbar()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "2933c928",
   "metadata": {},
   "outputs": [],
   "source": [
    "lowfreqmask.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "c74b7e63",
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "lowfreqmask,_,_ = mask_naiveRand(144,fix=10,other=4,roll=False)\n",
    "img_crude       = fft_observe(img,lowfreqmask)\n",
    "plt.figure()\n",
    "plt.imshow(img_crude[0][0],origin='lower')\n",
    "plt.colorbar()\n",
    "plt.show()\n",
    "print(f'error = { (torch.norm(img_crude[0][0]-img[0][0])/torch.norm(img[0][0])).item():.4f}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "87a225f0",
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "kin = fft_observe(img,lowfreqmask,return_opt='freq')\n",
    "plt.figure()\n",
    "plt.imshow(torch.log(torch.abs(kin[0][0])),origin='lower')\n",
    "plt.colorbar()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "3cffaaa8",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "005c766f",
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "x_sigpy = sigpy_solver(kin[:,0:1,:,:],L=5e-3,solver='ADMM')\n",
    "plt.figure()\n",
    "plt.imshow(x_sigpy[0,0,:,:],origin='lower')\n",
    "plt.colorbar()\n",
    "plt.show()\n",
    "\n",
    "plt.figure()\n",
    "plt.imshow(img[0][0],origin='lower')\n",
    "plt.colorbar()\n",
    "plt.show()\n",
    "err = torch.norm(x_sigpy[0,0,:,:]-img[0][0],p='fro')/torch.norm(img[0][0],p='fro')\n",
    "print(f'error = { err.item():.4f}')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f3b61741",
   "metadata": {},
   "source": [
    "## learning with memory replay"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "9ff2a1ad",
   "metadata": {},
   "outputs": [],
   "source": [
    "dim = 100\n",
    "mask = torch.zeros(dim)\n",
    "mask[np.random.permutation(dim)[0:60]] = 1\n",
    "(mask==0).nonzero(as_tuple=True)[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "b9adaba6",
   "metadata": {},
   "outputs": [],
   "source": [
    "a = torch.randn((2,1))\n",
    "a.mean().item()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "8d2ab1a8",
   "metadata": {},
   "outputs": [],
   "source": [
    "class RL_trainer():\n",
    "    def __init__(self,dataloader,policy,memory,episodes:int=10,eps:float=1e-3,\n",
    "                 fulldim:int=144,base:int=10,budget:int=50):\n",
    "        self.dataloader = dataloader\n",
    "        self.dataloader.reset()\n",
    "        \n",
    "        self.policy   = policy\n",
    "        self.memory   = memory\n",
    "        self.episodes = episodes\n",
    "        self.epi = 0\n",
    "        self.fulldim = fulldim\n",
    "        self.base = base\n",
    "        self.budget = budget\n",
    "        self.eps = eps\n",
    "        self.training_record = {'loss':[],'grad_norm':[],'q_values_mean':[],'q_values_std':[]}\n",
    "        self.steps = 1\n",
    "    def train(self):      \n",
    "        # run training\n",
    "        while self.epi < self.episodes:\n",
    "            print(f'episode {self.epi:.3d} / {self.episodes}')\n",
    "            mask = mask_naiveRand(self.fulldim,fix=self.base,other=0,roll=False)   # one mask at a time\n",
    "            while mask.sum() < self.budget + self.base:\n",
    "                data_source, data_target = self.dataloader.load()\n",
    "#                 epsilon = _get_epsilon(steps_epsilon, self.options)\n",
    "                curr_obs = fft_observe(data_source,mask)\n",
    "                action   = self.policy.get_action(data_source, mask=mask, eps_threshold=self.eps)\n",
    "                next_obs, reward = self.policy.step(action, data_target, mask)\n",
    "                self.memory.push(curr_obs, action, next_obs, reward)\n",
    "                \n",
    "                ### compare with random policy\n",
    "                with torch.no_grad():\n",
    "                    action_rand = self.policy.get_rand_action(mask=mask)\n",
    "                    _, reward_rand = self.policy.step(action_rand, data_target, mask)\n",
    "                ########################################################################\n",
    "                update_results = self.policy.update_parameters()\n",
    "                if update_results is not None:\n",
    "                    for key in self.training_record.keys():\n",
    "                        self.training_record[key].append(update_results[key])\n",
    "                    curr_loss = update_results['loss']\n",
    "                    print(f'step: {self.steps:5d}, loss: {curr_loss:.4f}, RL reward: {reward.mean().item():.4f}, Rand reward: {reward_rand.mean().item():.4f}')\n",
    "                    torch.cuda.empty_cache()\n",
    "                self.steps += 1\n",
    "#                 if self.steps % self.options.target_net_update_freq == 0:\n",
    "#                     self.logger.info(\"Updating target network.\")\n",
    "#                     self.target_net.load_state_dict(self.policy.state_dict())\n",
    "            self.dataloader.reset()\n",
    "            self.epi += 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "7b96adb2",
   "metadata": {},
   "outputs": [],
   "source": [
    "import dqn\n",
    "reload(dqn)\n",
    "from dqn import DQN\n",
    "\n",
    "import replay_buffer\n",
    "reload(replay_buffer)\n",
    "from replay_buffer import *"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "4c3952da",
   "metadata": {},
   "outputs": [],
   "source": [
    "memory = ReplayMemory(capacity=10,curr_obs_shape=(3,192,144),next_obs_shape=(1,192,144),batch_size=2,burn_in=2)\n",
    "model  = poly_net(samp_dim=144)\n",
    "policy = DQN(model,memory)\n",
    "trainer = RL_trainer(loader,policy,memory)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "c1bc697b",
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "trainer.train()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e2efb434",
   "metadata": {},
   "source": [
    "## view data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "97b4f1bb",
   "metadata": {},
   "outputs": [],
   "source": [
    "'''\n",
    "[kx, ky, kz, coil, phase, set, slice, rep, avg], \n",
    "frequency encoding, first phase encoding, second phase encoding, \n",
    "coil, phase (time), set (velocity encoding), slice, repetition, and number of averages\n",
    "\n",
    "better not to take average over the repetition dimension\n",
    "'''"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "e220a3fa",
   "metadata": {},
   "outputs": [],
   "source": [
    "datapath = '/mnt/shared_a/OCMR/OCMR_data_fully_sampled/'\n",
    "h5files = list([])\n",
    "for file in os.listdir(datapath):\n",
    "    if file.endswith(\".h5\"):\n",
    "        h5files.append(file)\n",
    "print(len(h5files))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "f8765121",
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "regular_files = []\n",
    "for file in h5files:\n",
    "    kData,param = read.read_ocmr(datapath+file)\n",
    "    print('filename', file, 'Dimension of kData: ', kData.shape)\n",
    "    curr_shape = kData.shape\n",
    "    if curr_shape[0]==384 and curr_shape[1]==144:\n",
    "        regular_files.append([file, curr_shape])\n",
    "#     print('Scan paramters:')\n",
    "#     pprint.pprint(param)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "83bfd6fe",
   "metadata": {},
   "outputs": [],
   "source": [
    "path = '/home/huangz78/rl_samp/useful_files.npz'\n",
    "# np.savez(path,files=regular_files)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "93946d54",
   "metadata": {
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "data = np.load(path,allow_pickle=True)\n",
    "data['files']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "acca70fa",
   "metadata": {},
   "outputs": [],
   "source": [
    "file = 'fs_0074_1_5T.h5'\n",
    "kData,param = read.read_ocmr(datapath+file)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "774b8f5b",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Image reconstruction (SoS)\n",
    "dim_kData = kData.shape\n",
    "print('Dimension of kData: ', dim_kData)\n",
    "CH  = dim_kData[3]\n",
    "SLC = dim_kData[6]\n",
    "print('SLC: ', SLC)\n",
    "kData_tmp = np.mean(kData, axis = 8); # average the k-space if average > 1, avg dimension is reduced\n",
    "\n",
    "im_coil = transform.transform_kspace_to_image(kData_tmp, [0,1]); # IFFT (2D image)\n",
    "im_sos  = np.sqrt(np.sum(np.abs(im_coil) ** 2, 3)) # Sum of Square, coil dimension is reduced\n",
    "print('Dimension of Image (with ReadOut ovesampling): ', im_sos.shape)\n",
    "RO    = im_sos.shape[0]\n",
    "\n",
    "# breakpoint()\n",
    "image = im_sos[math.floor(RO/4):math.floor(RO/4*3),:,:] # Remove RO oversampling\n",
    "print('Dimension of Image (without ReadOout ovesampling): ', image.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "d8201c7d",
   "metadata": {},
   "outputs": [],
   "source": [
    "print('image sos shape',im_sos.shape)\n",
    "im_sos_slice = im_sos[:,:,0,0,0,0,0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "453a1d8b",
   "metadata": {},
   "outputs": [],
   "source": [
    "kslice = kData_tmp[:,:,0,0,0,0,0,0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "314d0fa0",
   "metadata": {},
   "outputs": [],
   "source": [
    "im_sos_slice[0,:10]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "d9a5b852",
   "metadata": {},
   "outputs": [],
   "source": [
    "# coil_ind = 19\n",
    "# print(im_coil.shape)\n",
    "# im_slice = im_coil[:,:,0,coil_ind,0,0,0,0]\n",
    "\n",
    "plt.figure()\n",
    "plt.imshow(image[:,:,0,0,0,0,0],origin='lower')\n",
    "# plt.imshow(np.abs(im_slice),origin='lower')\n",
    "# plt.imshow(np.log(np.abs(kslice)),origin='lower')\n",
    "plt.colorbar()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "8d2180c6",
   "metadata": {},
   "outputs": [],
   "source": [
    "image_slc.shape[2]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "a3f29500",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Show the reconstructed cine image\n",
    "from IPython.display import clear_output\n",
    "import time\n",
    "\n",
    "slc_idx = 2 # math.floor(SLC/2)\n",
    "print(slc_idx)\n",
    "image_slc = np.squeeze(image[:,:,:,:,:,slc_idx,:])/image[:,:,:,:,:,slc_idx,:].max()\n",
    "print(image_slc.shape)\n",
    "# image_slc = np.squeeze(image[:,:,:,:,:,:,slc_idx])\n",
    "for rep in range(5): # repeate the movie for 5 times\n",
    "    for frame in range(image_slc.shape[2]):\n",
    "        clear_output(wait=True)\n",
    "        plt.imshow(image_slc[:,:,frame], vmin=0, vmax=0.6*np.amax(image_slc), cmap = 'gray',origin='lower')\n",
    "        plt.axis('off')\n",
    "        plt.colorbar()\n",
    "        plt.show() \n",
    "        time.sleep(0.03)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "cc5341bc",
   "metadata": {},
   "outputs": [],
   "source": [
    "savepath = '/mnt/shared_a/OCMR/OCMR_fully_sampled_images/'\n",
    "for fileind in range(len(regular_files)):\n",
    "    filename    = regular_files[fileind][0]\n",
    "    kData,param = read.read_ocmr(datapath+file)\n",
    "    kData_tmp   = np.mean(kData, axis = 8); # average the k-space if average > 1\n",
    "\n",
    "    im_coil = transform.transform_kspace_to_image(kData_tmp, [0,1]); # IFFT (2D image)\n",
    "    im_sos  = np.sqrt(np.sum(np.abs(im_coil) ** 2, 3)) # Sum of Square\n",
    "    print('Dimension of Image (with ReadOut ovesampling): ', im_sos.shape)\n",
    "    RO    = im_sos.shape[0]\n",
    "    image = im_sos[math.floor(RO/4):math.floor(RO/4*3),:,:] # Remove RO oversampling\n",
    "    print('Dimension of Image (without ReadOout ovesampling): ', image.shape)\n",
    "    \n",
    "    for sliceind in range(image.shape[5]):\n",
    "        image[:,:,:,:,:,sliceind,:] /= image[:,:,:,:,:,sliceind,:].max()\n",
    "    \n",
    "    torch.save(torch.tensor(image) , savepath + filename[:-3] + '.pt')\n",
    "    print(f'file {savepath + filename} is successfully saved')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "d3eaea9f",
   "metadata": {},
   "outputs": [],
   "source": [
    "datapath = '/mnt/shared_a/OCMR/OCMR_fully_sampled_images/'\n",
    "file = 'fs_0074_1_5T.pt'\n",
    "data = torch.load(datapath+file)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "abe30831",
   "metadata": {},
   "outputs": [],
   "source": [
    "slc_idx = 6 # math.floor(SLC/2)\n",
    "print(slc_idx)\n",
    "# image_slc = np.squeeze(image[:,:,:,:,:,slc_idx,:])/image[:,:,:,:,:,slc_idx,:].max()\n",
    "# print(image_slc.shape)\n",
    "# image_slc = np.squeeze(image[:,:,:,:,:,:,slc_idx])\n",
    "for rep in range(3): # repeat the movie for 5 times\n",
    "    for frame in range(data.shape[3]):\n",
    "        clear_output(wait=True)\n",
    "        plt.imshow(data[:,:,0,frame,0,slc_idx,:], vmin=0, vmax=data.max(), cmap = 'gray'); plt.axis('off');\n",
    "        plt.colorbar()\n",
    "        plt.show() \n",
    "        time.sleep(0.03)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "conda_pytorch_p39",
   "language": "python",
   "name": "conda_pytorch_p39"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.9.15"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}