How to fix a def function issue for applying-CodePudding

I'm a new student. I think the class code is correct. But the 'def update_vocab' is not applied. (Under bar is always appeared at update_vocab(q),update_vocab(a)) How can I fix this problem? Is the def update_vocab is wrong?

class sequence:
  id_to_char = {}
  char_to_id = {}

  def update_vocab(txt):
    chars = list(txt)
    for i, char in enumerate(chars):
        if char not in char_to_id:
            tmp_id = len(char_to_id)
            char_to_id[char] = tmp_id
            id_to_char[tmp_id] = char

  def load_data(file_name='addition.txt', seed=1984):
    file_path = os.path.dirname(os.path.abspath('data.txt'))   '/'   file_name
    if not os.path.exists(file_path):
        print('No file: %s' % file_name)
    return None

    questions, answers = [], []
    for line in open(file_path, 'r'):
        idx = line.find('_')enter code here
        questions.append(line[:idx])
        answers.append(line[idx:-1])

    for i in range(len(questions)):
        q,a = questions[i], answers[i]
        update_vocab(q)
        update_vocab(a)

    x = torch.zeros((len(questions), len(questions[0])), dtype=torch.int)
    t = torch.zeros((len(questions), len(answers[0])), dtype=torch.int)

    for i, sentence in enumerate(questions):
        x[i] = [char_to_id[c] for c in list(sentence)]
    for i, sentence in enumerate(answers):
        t[i] = [char_to_id[c] for c in list(sentence)]

    indices = torch.arange(len(x))
    if seed is not None:
        torch.random.seed(seed)
    torch.random.shuffle(indices)
    x = x[indices]
    t = t[indices]

    split_at = len(x) - len(x) // 10
    (x_train, x_test) = x[:split_at], x[split_at:]
    (t_train, t_test) = t[:split_at], t[split_at:]
    return (x_train, t_train), (x_test, t_test)

  def get_vocab():
    return char_to_id, id_to_char'

CodePudding user response：

please add self, in param function. check

I try your code:

import os
import torch

class sequence:
    def __init__(self):
        self.id_to_char = {}
        self.char_to_id = {}

    def update_vocab(self, txt):
        chars = list(txt)
        for i, char in enumerate(chars):
            if char not in self.char_to_id:
                tmp_id = len(self.char_to_id)
                self.char_to_id[char] = tmp_id
                self.id_to_char[tmp_id] = char

    def load_data(self, file_name='addition.txt', seed=1984):
        file_path = os.path.dirname(os.path.abspath('data.txt'))   '/'   file_name
        if not os.path.exists(file_path):
            print('No file: %s' % file_name)
        return None

        questions, answers = [], []
        for line in open(file_path, 'r'):
            idx = line.find('_')
            questions.append(line[:idx])
            answers.append(line[idx:-1])

        for i in range(len(questions)):
            q,a = questions[i], answers[i]
            update_vocab(q)
            update_vocab(a)

        x = torch.zeros((len(questions), len(questions[0])), dtype=torch.int)
        t = torch.zeros((len(questions), len(answers[0])), dtype=torch.int)

        for i, sentence in enumerate(questions):
            x[i] = [self.char_to_id[c] for c in list(sentence)]
        for i, sentence in enumerate(answers):
            t[i] = [self.char_to_id[c] for c in list(sentence)]

        indices = torch.arange(len(x))
        if seed is not None:
            torch.random.seed(seed)
        torch.random.shuffle(indices)
        x = x[indices]
        t = t[indices]

        split_at = len(x) - len(x) // 10
        (x_train, x_test) = x[:split_at], x[split_at:]
        (t_train, t_test) = t[:split_at], t[split_at:]
        return (x_train, t_train), (x_test, t_test)

    def get_vocab(self):
        return self.char_to_id, self.id_to_char

seq = sequence()
seq.update_vocab('a')
print(seq.get_vocab())

CodePudding user response：

Wrong indentation: method load_data() always return None, try:

...
if not os.path.exists(file_path): 
  print('No file: %s' % file_name) 
  return None
...