main_sfda.py

#/usr/bin/env python3.6
import math
import re
import argparse
import warnings
warnings.filterwarnings("ignore")
from pathlib import Path
from operator import itemgetter
from shutil import copytree, rmtree
import typing
from typing import Any, Callable, List, Tuple
import matplotlib.pyplot as plt
import torch
import numpy as np
import pandas as pd
import torch.nn.functional as F
from torch import Tensor
from torch.utils.data import DataLoader
from dice3d import dice3d
from networks import weights_init
from dataloader import get_loaders
from utils import map_, save_dict_to_file
from utils import dice_coef, dice_batch, save_images,save_images_p,save_be_images, tqdm_, save_images_ent
from utils import probs2one_hot, probs2class, mask_resize, resize, haussdorf
from utils import exp_lr_scheduler
import datetime
from itertools import cycle
import os
from time import sleep
from bounds import CheckBounds
import matplotlib.pyplot as plt
from itertools import chain
import platform



def setup(args, n_class, dtype) -> Tuple[Any, Any, Any, List[Callable], List[float],List[Callable], List[float], Callable]:
    print(">>> Setting up")
    cpu: bool = args.cpu or not torch.cuda.is_available()
    if cpu:
        print("WARNING CUDA NOT AVAILABLE")
    device = torch.device("cpu") if cpu else torch.device("cuda")
    n_epoch = args.n_epoch
    if args.model_weights:
        if cpu:
            net = torch.load(args.model_weights, map_location='cpu')
        else:
            net = torch.load(args.model_weights)
    else:
        net_class = getattr(__import__('networks'), args.network)
        net = net_class(1, n_class).type(dtype).to(device)
        net.apply(weights_init)
    net.to(device)
    if args.saveim:
        print("WARNING SAVING MASKS at each epc")

    optimizer = torch.optim.Adam(net.parameters(), lr=args.l_rate, betas=(0.9, 0.999),weight_decay=args.weight_decay)
    if args.adamw:
        optimizer = torch.optim.AdamW(net.parameters(), lr=args.l_rate, betas=(0.9, 0.999))

    print(args.target_losses)
    losses = eval(args.target_losses)
    loss_fns: List[Callable] = []
    for loss_name, loss_params, _, bounds_params, fn, _ in losses:
        loss_class = getattr(__import__('losses'), loss_name)
        loss_fns.append(loss_class(**loss_params, dtype=dtype, fn=fn))
        print("bounds_params", bounds_params)
        if bounds_params!=None:
            bool_predexist = CheckBounds(**bounds_params)
            print(bool_predexist,"size predictor")
            if not bool_predexist:
                n_epoch = 0

    loss_weights = map_(itemgetter(5), losses)

    if args.scheduler:
        scheduler = getattr(__import__('scheduler'), args.scheduler)(**eval(args.scheduler_params))
    else:
        scheduler = ''

    return net, optimizer, device, loss_fns, loss_weights, scheduler, n_epoch


def do_epoch(args, mode: str, net: Any, device: Any, epc: int,
             loss_fns: List[Callable], loss_weights: List[float],
              new_w:int, C: int, metric_axis:List[int], savedir: str = "",
             optimizer: Any = None, target_loader: Any = None, best_dice3d_val:Any=None):

    assert mode in ["train", "val"]
    L: int = len(loss_fns)
    indices = torch.tensor(metric_axis,device=device)
    if mode == "train":
        net.train()
        desc = f">> Training   ({epc})"
    elif mode == "val":
        net.eval()
        desc = f">> Validation ({epc})"

    total_it_t, total_images_t = len(target_loader), len(target_loader.dataset)
    total_iteration = total_it_t
    total_images = total_images_t

    if args.debug:
        total_iteration = 10
    pho=1
    dtype = eval(args.dtype)

    all_dices: Tensor = torch.zeros((total_images, C), dtype=dtype, device=device)
    all_sizes: Tensor = torch.zeros((total_images, C), dtype=dtype, device=device)
    all_sizes: Tensor = torch.zeros((total_images, C), dtype=dtype, device=device)
    all_gt_sizes: Tensor = torch.zeros((total_images, C), dtype=dtype, device=device)
    all_sizes2: Tensor = torch.zeros((total_images, C), dtype=dtype, device=device)
    all_inter_card: Tensor = torch.zeros((total_images, C), dtype=dtype, device=device)
    all_card_gt: Tensor = torch.zeros((total_images, C), dtype=dtype, device=device)
    all_card_pred: Tensor = torch.zeros((total_images, C), dtype=dtype, device=device)
    all_gt = []
    all_pred = []
    if args.do_hd: 
        #all_gt: Tensor = torch.zeros((total_images, 256, 256), dtype=dtype)
        all_gt: Tensor = torch.zeros((total_images, 384, 384), dtype=dtype)
        #all_pred: Tensor = torch.zeros((total_images, 256, 256), dtype=dtype)
        all_pred: Tensor = torch.zeros((total_images, 384, 384), dtype=dtype)
    loss_log: Tensor = torch.zeros((total_images), dtype=dtype, device=device)
    loss_cons: Tensor = torch.zeros((total_images), dtype=dtype, device=device)
    loss_se: Tensor = torch.zeros((total_images), dtype=dtype, device=device)
    loss_tot: Tensor = torch.zeros((total_images), dtype=dtype, device=device)
    posim_log: Tensor = torch.zeros((total_images), dtype=dtype, device=device)
    haussdorf_log: Tensor = torch.zeros((total_images, C), dtype=dtype, device=device)
    all_grp: Tensor = torch.zeros((total_images, C), dtype=dtype, device=device)
    all_pnames = np.zeros([total_images]).astype('U256') 
    dice_3d_log: Tensor = torch.zeros((1, C), dtype=dtype, device=device)
    dice_3d_sd_log: Tensor = torch.zeros((1, C), dtype=dtype, device=device)
    hd95_3d_log: Tensor = torch.zeros((1, C), dtype=dtype, device=device)
    hd95_3d_sd_log: Tensor = torch.zeros((1, C), dtype=dtype, device=device)
    tq_iter = tqdm_(enumerate(target_loader), total=total_iteration, desc=desc)
    done: int = 0
    n_warmup = args.n_warmup
    mult_lw = [pho ** (epc - n_warmup + 1)] * len(loss_weights)
    mult_lw[0] = 1
    loss_weights = [a * b for a, b in zip(loss_weights, mult_lw)]
    losses_vec, source_vec, target_vec, baseline_target_vec = [], [], [], []
    pen_count = 0
    with warnings.catch_warnings():
        warnings.simplefilter("ignore")
        count_losses = 0
        for j, target_data in tq_iter:
            target_data[1:] = [e.to(device) for e in target_data[1:]]  # Move all tensors to device
            filenames_target, target_image, target_gt = target_data[:3]
            labels = target_data[3:3+L]
            bounds = target_data[3+L:]
            filenames_target = [f.split('.nii')[0] for f in filenames_target]
            assert len(labels) == len(bounds), len(bounds)
            B = len(target_image)
            # Reset gradients
            if optimizer:
                #adjust_learning_rate(optimizer, 1, args.l_rate, args.power)
                optimizer.zero_grad()
            # Forward
            with torch.set_grad_enabled(mode == "train"):
                pred_logits: Tensor = net(target_image)
                pred_probs: Tensor = F.softmax(pred_logits, dim=1)
                predicted_mask: Tensor = probs2one_hot(pred_probs)  # Used only for dice computation
            assert len(bounds) == len(loss_fns) == len(loss_weights)
            if epc < n_warmup:
                loss_weights = [0]*len(loss_weights)
            loss: Tensor = torch.zeros(1, requires_grad=True).to(device)
            loss_vec = []
            loss_kw = []
            for loss_fn,label, w, bound in zip(loss_fns,labels, loss_weights, bounds):
                if w > 0:
                    if eval(args.target_losses)[0][0]=="EntKLProp": 
                        loss_1, loss_cons_prior,est_prop =  loss_fn(pred_probs, label, bound)
                        loss = loss_1 + loss_cons_prior 
                    else:
                        loss =  loss_fn(pred_probs, label, bound)
                        loss = w*loss
                        loss_1 = loss
                    loss_kw.append(loss_1.detach())
           # Backward
            if optimizer:
                loss.backward()
                optimizer.step()
            # Compute and log metrics
            dices, inter_card, card_gt, card_pred = dice_coef(predicted_mask.detach(), target_gt.detach())
            assert dices.shape == (B, C), (dices.shape, B, C)
            sm_slice = slice(done, done + B)  # Values only for current batch
            all_dices[sm_slice, ...] = dices
            if eval(args.target_losses)[0][0] in ["EntKLProp","WeightedEntKLProp","EntKLProp2","CEKLProp2"]:
                all_sizes[sm_slice, ...] = torch.round(est_prop.detach()*target_image.shape[2]*target_image.shape[3])
            all_sizes2[sm_slice, ...] = torch.sum(predicted_mask,dim=(2,3)) 
            all_gt_sizes[sm_slice, ...] = torch.sum(target_gt,dim=(2,3)) 
            # # for 3D dice
            if 'slice' in args.grp_regex:
                all_grp[sm_slice, ...] = torch.FloatTensor([int(re.split('_',re.split('slice',x)[1])[0]) for x in filenames_target]).unsqueeze(1).repeat(1,C)
                #all_grp[sm_slice, ...] = torch.FloatTensor([int(re.split('_',re.split('Subj',x)[1])[0]) for x in filenames_target]).unsqueeze(1).repeat(1,C)
            elif 'Case' in args.grp_regex:
                all_grp[sm_slice, ...] = torch.FloatTensor([int(re.split('_',re.split('Case',x)[1])[0]) for x in filenames_target]).unsqueeze(1).repeat(1,C)
            else:
                all_grp[sm_slice, ...] = int(re.split('_', filenames_target[0])[1]) * torch.ones([1, C])
            all_pnames[sm_slice] = filenames_target
            all_inter_card[sm_slice, ...] = inter_card
            all_card_gt[sm_slice, ...] = card_gt
            all_card_pred[sm_slice, ...] = card_pred
            if args.do_hd:
                all_pred[sm_slice, ...] = probs2class(predicted_mask[:,:,:,:]).cpu().detach()
                all_gt[sm_slice, ...] = probs2class(target_gt).detach()
            loss_se[sm_slice] = loss_kw[0]
            if len(loss_kw)>1:
            	loss_cons[sm_slice] = loss_kw[1]
            	loss_tot[sm_slice] = loss_kw[1]+loss_kw[0]
            else:
            	loss_cons[sm_slice] = 0
            	loss_tot[sm_slice] = loss_kw[0]
            # Save images
            if savedir and args.saveim and mode =="val":
                with warnings.catch_warnings():
                    warnings.filterwarnings("ignore", category=UserWarning)
                    warnings.simplefilter("ignore") 
                    predicted_class: Tensor = probs2class(pred_probs)
                    save_images(predicted_class, filenames_target, savedir, mode, epc, False)
                    if args.entmap:
                        ent_map = torch.einsum("bcwh,bcwh->bwh", [-pred_probs, (pred_probs+1e-10).log()])
                        save_images_ent(ent_map, filenames_target, savedir,'ent_map', epc)
            # Logging
            big_slice = slice(0, done + B)  # Value for current and previous batches
            stat_dict = {**{f"DSC{n}": all_dices[big_slice, n].mean() for n in metric_axis},
                         **{f"SZ{n}": all_sizes[big_slice, n].mean() for n in metric_axis},
                         **({f"DSC_source{n}": all_dices_s[big_slice, n].mean() for n in metric_axis}
                           if args.source_metrics else {})}

            size_dict = {**{f"SZ{n}": all_sizes[big_slice, n].mean() for n in metric_axis}}
            nice_dict = {k: f"{v:.4f}" for (k, v) in stat_dict.items()}
            done += B
            tq_iter.set_postfix(nice_dict)
    if args.dice_3d and (mode == 'val'):
        dice_3d_log, dice_3d_sd_log, asd_3d_log, asd_3d_sd_log, hd_3d_log, hd_3d_sd_log = dice3d(all_grp, all_inter_card, all_card_gt, all_card_pred,all_pred,all_gt,all_pnames,metric_axis,args.pprint,args.do_hd, best_dice3d_val)
          
    dice_2d = torch.index_select(all_dices, 1, indices).mean().cpu().numpy()
    target_vec = [ dice_3d_log, dice_3d_sd_log,hd95_3d_log,hd95_3d_sd_log,dice_2d]
    size_mean = torch.index_select(all_sizes2, 1, indices).mean(dim=0).cpu().numpy()
    size_gt_mean = torch.index_select(all_gt_sizes, 1, indices).mean(dim=0).cpu().numpy()
    mask_pos = torch.index_select(all_sizes2, 1, indices)!=0
    gt_pos = torch.index_select(all_gt_sizes, 1, indices)!=0
    size_mean_pos = torch.index_select(all_sizes2, 1, indices).sum(dim=0).cpu().numpy()/mask_pos.sum(dim=0).cpu().numpy()
    gt_size_mean_pos = torch.index_select(all_gt_sizes, 1, indices).sum(dim=0).cpu().numpy()/gt_pos.sum(dim=0).cpu().numpy()
    size_mean2 = torch.index_select(all_sizes2, 1, indices).mean(dim=0).cpu().numpy()
    losses_vec = [loss_se.mean().item(),loss_cons.mean().item(),loss_tot.mean().item(),np.int(size_mean.mean()),np.int(size_mean_pos.mean()),np.int(size_gt_mean.mean()),np.int(gt_size_mean_pos.mean())]
    if not epc%10:
        df_t = pd.DataFrame({
           "val_ids":all_pnames,
           "proposal_size":all_sizes2.cpu()})
        df_t.to_csv(Path(savedir,mode+str(epc)+"sizes.csv"), float_format="%.4f", index_label="epoch")
    return losses_vec, target_vec,source_vec

def run(args: argparse.Namespace) -> None:
    d = vars(args)
    d['time'] = str(datetime.datetime.now())
    d['server']=platform.node()
    save_dict_to_file(d,args.workdir)
    temperature: float = 0.1
    n_class: int = args.n_class
    metric_axis: List = args.metric_axis
    lr: float = args.l_rate
    dtype = eval(args.dtype)
    savedir: str = args.workdir
    n_epoch: int = args.n_epoch
    net, optimizer, device, loss_fns, loss_weights, scheduler, n_epoch = setup(args, n_class, dtype)
    shuffle = True
    print(args.target_folders)
    target_loader, target_loader_val = get_loaders(args, args.target_dataset,args.target_folders,
                                           args.batch_size, n_class,
                                           args.debug, args.in_memory, dtype, shuffle, "target", args.val_target_folders)

    print("metric axis",metric_axis)
    best_dice_pos: Tensor = np.zeros(1)
    best_dice: Tensor = np.zeros(1)
    best_hd3d_dice: Tensor = np.zeros(1)
    best_3d_dice: Tensor = 0 
    best_2d_dice: Tensor = 0 
    print("Results saved in ", savedir)
    print(">>> Starting the training")
    for i in range(n_epoch):

       if args.mode =="makeim":
            with torch.no_grad():
                 
                val_losses_vec, val_target_vec,val_source_vec                                        = do_epoch(args, "val", net, device,
                                                                                                i, loss_fns,
                                                                                               loss_weights,
                                                                                               args.resize,
                                                                                                n_class,metric_axis,
                                                                                               savedir=savedir,
                                                                                               target_loader=target_loader_val, best_dice3d_val=best_3d_dice)
                
                tra_losses_vec = val_losses_vec
                tra_target_vec = val_target_vec
                tra_source_vec = val_source_vec
       else:
            tra_losses_vec, tra_target_vec,tra_source_vec                                    = do_epoch(args, "train", net, device,
                                                                                           i, loss_fns,
                                                                                           loss_weights,
                                                                                           args.resize,
                                                                                           n_class, metric_axis,
                                                                                           savedir=savedir,
                                                                                           optimizer=optimizer,
                                                                                           target_loader=target_loader, best_dice3d_val=best_3d_dice)
       
            with torch.no_grad():
                val_losses_vec, val_target_vec,val_source_vec                                        = do_epoch(args, "val", net, device,
                                                                                               i, loss_fns,
                                                                                               loss_weights,
                                                                                               args.resize,
                                                                                               n_class,metric_axis,
                                                                                               savedir=savedir,
                                                                                               target_loader=target_loader_val, best_dice3d_val=best_3d_dice)
       current_val_target_2d_dice = val_target_vec[4]
       current_val_target_3d_dice = val_target_vec[0]
       if args.dice_3d:
           if current_val_target_3d_dice > best_3d_dice:
               best_epoch = i
               best_3d_dice = current_val_target_3d_dice
               with open(Path(savedir, "3dbestepoch.txt"), 'w') as f:
                   f.write(str(i)+','+str(best_3d_dice))
               best_folder_3d = Path(savedir, "best_epoch_3d")
               if best_folder_3d.exists():
                    rmtree(best_folder_3d)
               if args.saveim:
                    copytree(Path(savedir, f"iter{i:03d}"), Path(best_folder_3d))
           torch.save(net, Path(savedir, "best_3d.pkl"))
       if not(i % 10) :
            print("epoch",str(i),savedir,'best 3d dice',best_3d_dice)
            torch.save(net, Path(savedir, "epoch_"+str(i)+".pkl"))
       if i == n_epoch - 1:
            with open(Path(savedir, "last_epoch.txt"), 'w') as f:
                f.write(str(i))
            last_folder = Path(savedir, "last_epoch")
            if last_folder.exists():
                rmtree(last_folder)
            if args.saveim:
                copytree(Path(savedir, f"iter{i:03d}"), Path(last_folder))
            torch.save(net, Path(savedir, "last.pkl"))

        # remove images from iteration
       if args.saveim:
           rmtree(Path(savedir, f"iter{i:03d}"))

       df_t_tmp = pd.DataFrame({
            "epoch":i,
            "tra_loss_s":[tra_losses_vec[0]],
            "tra_loss_cons":[tra_losses_vec[1]],
            "tra_loss_tot":[tra_losses_vec[2]],
            "tra_size_mean":[tra_losses_vec[3]],
            "tra_size_mean_pos":[tra_losses_vec[4]],
            "val_loss_s":[val_losses_vec[0]],
            "val_loss_cons":[val_losses_vec[1]],
            "val_loss_tot":[val_losses_vec[2]],
            "val_size_mean":[val_losses_vec[3]],
            "val_size_mean_pos":[val_losses_vec[4]],
            "val_gt_size_mean":[val_losses_vec[5]],
            "val_gt_size_mean_pos":[val_losses_vec[6]],
            'tra_dice': [tra_target_vec[4]],
            'val_dice': [val_target_vec[4]],
            "val_dice_3d_sd": [val_target_vec[1]],
            "val_dice_3d": [val_target_vec[0]]})

       if i == 0:
            df_t = df_t_tmp
       else:
            df_t = df_t.append(df_t_tmp)

       df_t.to_csv(Path(savedir, "_".join((args.target_folders.split("'")[1],"target", args.csv))), float_format="%.4f", index=False)

       if args.flr==False:
            exp_lr_scheduler(optimizer, i, args.lr_decay,args.lr_decay_epoch)
    print("Results saved in ", savedir, "best 3d dice",best_3d_dice)


def get_args() -> argparse.Namespace:
    parser = argparse.ArgumentParser(description='Hyperparams')
    parser.add_argument('--target_dataset', type=str, required=True)
    parser.add_argument("--workdir", type=str, required=True)
    parser.add_argument("--target_losses", type=str, required=True,
                        help="List of (loss_name, loss_params, bounds_name, bounds_params, fn, weight)")
    parser.add_argument("--target_folders", type=str, required=True,
                        help="List of (subfolder, transform, is_hot)")
    parser.add_argument("--val_target_folders", type=str, required=True,
                        help="List of (subfolder, transform, is_hot)")
    parser.add_argument("--network", type=str, required=True, help="The network to use")
    parser.add_argument("--grp_regex", type=str, required=True)
    parser.add_argument("--n_class", type=int, required=True)
    parser.add_argument("--mode", type=str, default="learn")
    parser.add_argument("--lin_aug_w", action="store_true")
    parser.add_argument("--both", action="store_true")
    parser.add_argument("--trainval", action="store_true")
    parser.add_argument("--valonly", action="store_true")
    parser.add_argument("--flr", action="store_true")
    parser.add_argument("--augment", action="store_true")
    parser.add_argument("--mix", type=bool, default=True)
    parser.add_argument("--do_hd", type=bool, default=False)
    parser.add_argument("--saveim", type=bool, default=False)
    parser.add_argument("--debug", action="store_true")
    parser.add_argument("--csv", type=str, default='metrics.csv')
    parser.add_argument("--source_metrics", action="store_true")
    parser.add_argument("--adamw", action="store_true")
    parser.add_argument("--dice_3d", action="store_true")
    parser.add_argument("--ontest", action="store_true")
    parser.add_argument("--ontrain", action="store_true")
    parser.add_argument("--best_losses", action="store_true")
    parser.add_argument("--pprint", action="store_true")
    parser.add_argument("--entmap", action="store_true")
    parser.add_argument("--model_weights", type=str, default='')
    parser.add_argument("--cpu", action='store_true')
    parser.add_argument("--in_memory", action='store_true')
    parser.add_argument("--resize", type=int, default=0)
    parser.add_argument("--pho", nargs='?', type=float, default=1,
                        help='augment')
    parser.add_argument("--n_warmup", type=int, default=0)
    parser.add_argument('--n_epoch', nargs='?', type=int, default=200,
                        help='# of the epochs')
    parser.add_argument('--l_rate', nargs='?', type=float, default=5e-4,
                        help='Learning Rate')
    parser.add_argument('--lr_decay', nargs='?', type=float, default=0.7),
    parser.add_argument('--lr_decay_epoch', nargs='?', type=float, default=20),
    parser.add_argument('--weight_decay', nargs='?', type=float, default=1e-5,
                        help='L2 regularisation of network weights')
    parser.add_argument('--batch_size', type=int, default=1)
    parser.add_argument("--dtype", type=str, default="torch.float32")
    parser.add_argument("--scheduler", type=str, default="DummyScheduler")
    parser.add_argument("--scheduler_params", type=str, default="{}")
    parser.add_argument("--power",type=float, default=0.9)
    parser.add_argument("--metric_axis",type=int, nargs='*', required=True, help="Classes to display metrics. \
        Display only the average of everything if empty")
    args = parser.parse_args()
    print(args)

    return args


if __name__ == '__main__':
    run(get_args())