utils.py

from __future__ import division
import csv
import numpy as np
import pdb
import os

class AverageMeter(object):
    """Computes and stores the average and current value"""

    def __init__(self):
        self.reset()

    def reset(self):
        self.val = 0
        self.avg = 0
        self.sum = 0
        self.count = 0

    def update(self, val, n=1):
        self.val = val
        self.sum += val * n
        self.count += n
        self.avg = self.sum / self.count


class Logger(object):

    def __init__(self, path, header, resume_path, begin_epoch):
        if (not os.path.exists(path)) or (resume_path==''):
            self.log_file = open(path, 'w+')
            self.logger = csv.writer(self.log_file, delimiter='\t')
            self.logger.writerow(header)
        else:
            self.log_file = open(path, 'r+')
            self.logger = csv.writer(self.log_file, delimiter='\t')
            reader = csv.reader(self.log_file, delimiter='\t')
            lines = []
            print("begin = ", begin_epoch)
            for line in reader:
                lines.append(line)
                if len(lines) == begin_epoch +1 :
                    break
            self.log_file.close()
            self.log_file = open(path, 'w')
            self.logger = csv.writer(self.log_file, delimiter='\t')
            self.logger.writerows(lines[:begin_epoch+1])
            self.log_file.flush()
            
        self.header = header


    def __del(self):
        self.log_file.close()

    def log(self, values):
        write_values = []
        for col in self.header:
            assert col in values
            write_values.append(values[col])

        self.logger.writerow(write_values)
        self.log_file.flush()

class Logger_MARS(object):

    def __init__(self, path, header, resume_path, begin_epoch):
        if resume_path == '':
            self.log_file = open(path, 'w+')
            self.logger = csv.writer(self.log_file, delimiter='\t')
            self.logger.writerow(header)
        else:
            self.log_file = open(path, 'r+')
            self.logger = csv.writer(self.log_file, delimiter='\t')
            reader = csv.reader(self.log_file, delimiter='\t')
            lines = []
            print("begin = ", begin_epoch)
            for line in reader:
                lines.append(line)
                if len(lines) == begin_epoch +1 :
                    break
            self.log_file.close()
            self.log_file = open(path, 'w')
            self.logger = csv.writer(self.log_file, delimiter='\t')
            self.logger.writerows(lines[:begin_epoch+1])
            self.log_file.flush()

        self.header = header


    def __del(self):
        self.log_file.close()

    def log(self, values):
        write_values = []
        for col in self.header:
            assert col in values
            write_values.append(values[col])

        self.logger.writerow(write_values)
        self.log_file.flush()


def load_value_file(file_path):
    with open(file_path, 'r') as input_file:
        value = float(input_file.read().rstrip('\n\r'))

    return value

def calculate_accuracy(outputs, targets):
    batch_size = targets.size(0)

    _, pred = outputs.topk(1, 1, True)
    pred = pred.t()
    correct = pred.eq(targets.view(1, -1))

    n_correct_elems = correct.float().sum().item()
    

    return n_correct_elems / batch_size
    
def calculate_accuracy5(outputs, targets):
    batch_size = targets.size(0)

    pred = outputs.topk(5, 1, True)[1].data[0].tolist()
    print("true = ", targets.view(1, -1).data[0].tolist()[0], "pred = ", pred)
    correct = targets.view(1, -1).data[0].tolist()[0] in pred
    print(correct)
    n_correct_elems = int(correct)

    return n_correct_elems / batch_size
    
def calculate_accuracy_video(output_buffer, i):
    true_value = output_buffer[: i+1,-1]
    pred_value = np.argmax(output_buffer[:i+1, :-1], axis = 1)
#    print(output_buffer[0:3,:])
    # print(true_value)
    # print(pred_value)
    # print("accuracy = ", 1*(np.equal(true_value, pred_value)).sum()/len(true_value))
    return 1*(np.equal(true_value, pred_value)).sum()/len(true_value)