File size: 3,053 Bytes
b8f95bb
 
 
 
 
 
 
84b28fd
b8f95bb
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
import gradio as gr
from transformers import pipeline
import pandas as pd
import numpy as np
import os

model_checkpoint = "penpen/novel-zh-en"
translator = pipeline("translation", model=model_checkpoint, max_time=7, num_beams=1)
default_dict = pd.read_csv("example_dictionary.csv", names=["Chinese", "English"])
examples = pd.read_csv("examples.csv", header = None)

def predict(text, df):
  translation = ""
  terms_dict = {chinese: english for chinese, english in zip(df["Chinese"].tolist(), df["English"].tolist())}
  for key in terms_dict:
    if key in text:
      masking = "MASK"*len(key)
      text = text.replace(key, "<TERM>" + masking+ "<GLOS>" + terms_dict[key] + "</GLOS>")

  split_text = text.splitlines()
  for text in split_text:
    text = text.strip()
    if text:
      if len(text) < 512:
        sentence = translator(text)[0]["translation_text"] + '\n\n'
        translation+=sentence
        print(split_text)
      else:
        for i in range(0,len(text),512):
          if i+512>len(text):
            sentence = translator(text[i:])[0]["translation_text"]
          else:
            sentence = translator(text[i:i+512])[0]["translation_text"]
          translation+=sentence
  return translation


def load_dict(file):
  df = pd.read_csv(file.name, names=["Chinese", "English"])
  return df, df

def search_dict(query, df):
  if not query:
    return df
  mask = np.column_stack([df[col].str.contains(query, na=False) for col in df])
  return df.loc[mask.any(axis=1)]

with gr.Blocks() as project:
  dict_hidden = gr.State(default_dict)
  gr.Markdown("<center><h1>Chinese Webnovel Translator</h1> A translator that is fine-tuned on Chinese Webnovels</center>")
  with gr.Tab("Translator"):
    with gr.Row():
      with gr.Column(scale=1, min_width=600):
        translate_input = gr.Textbox(label="Chinese", lines=7, max_lines = 100, placeholder="Chinese...")
        translate_button = gr.Button("Translate")
        translate_hidden = gr.State("")
      translate_output = gr.Textbox(label="English", lines=7, max_lines = 100, placeholder="English...")
    example = gr.Examples(inputs = translate_input, examples=examples[0].tolist())

  with gr.Tab("Proper Noun Dictionary"):
    with gr.Row():
      with gr.Column(scale=1, min_width=600):
        dict_example_file = gr.File(label="Example Dictionary", value = "example_dictionary.csv")
        dict_file = gr.File(interactive = True, label="Upload a custom dictionary (CSV File)")
        dict_upload_button = gr.Button("Upload")
        dict_search = gr.Textbox(label="Search Dictionary")
        dict_search_button = gr.Button("Search")

      dict_display = gr.Dataframe(value = default_dict, max_rows = 5, col_count=(2, "fixed"))
       
  translate_button.click(predict, inputs=[translate_input, dict_hidden], outputs=translate_output)
  
  dict_upload_button.click(load_dict, inputs=dict_file, outputs = [dict_hidden, dict_display])
  dict_search_button.click(search_dict, inputs=[dict_search, dict_hidden], outputs = dict_display)

project.launch(debug=True)