chinhon's picture
Update app.py
e631c98
import gradio as gr
import nltk
import numpy as np
import re
import warnings
from nltk.tokenize import sent_tokenize
from transformers import (
MarianTokenizer,
MarianMTModel,
)
nltk.download('punkt')
#define function for text cleaning
def clean_text(text):
text = text.encode("ascii", errors="ignore").decode(
"ascii"
) # remove non-ascii, Chinese characters
text = re.sub(r"\n", " ", text)
text = re.sub(r"\n\n", " ", text)
text = re.sub(r"\t", " ", text)
text = re.sub(r"http\S+", "", text)
text = re.sub(r"ADVERTISEMENT", " ", text)
text = re.sub(
r"Download our app or subscribe to our Telegram channel for the latest updates on the coronavirus outbreak: https://cna.asia/telegram",
" ",
text,
)
text = re.sub(
r"Download our app or subscribe to our Telegram channel for the latest updates on the COVID-19 outbreak: https://cna.asia/telegram",
" ",
text,
)
text = text.strip(" ")
text = re.sub(
" +", " ", text
).strip() # get rid of multiple spaces and replace with a single
return text
# define function for translation
modchoice = "Helsinki-NLP/opus-mt-en-zh"
def translate(text):
input_text = clean_text(text)
tokenizer = MarianTokenizer.from_pretrained(modchoice)
model = MarianMTModel.from_pretrained(modchoice)
if input_text is None or text == "":
return ("Error",)
translated = model.generate(
**tokenizer.prepare_seq2seq_batch(
sent_tokenize(input_text),
truncation=True,
padding="longest",
return_tensors="pt"
)
)
tgt_text = [tokenizer.decode(t, skip_special_tokens=True) for t in translated]
return " ".join(tgt_text)
gradio_ui = gr.Interface(
fn=translate,
title="English-to-Chinese translation",
description="Translate English text into Chinese using MarianMT's opus-mt-en-zh model.",
inputs=gr.inputs.Textbox(
lines=20, label="Paste English text here"
),
outputs=gr.outputs.Textbox(label="Chinese translation"),
theme="huggingface",
)
gradio_ui.launch(enable_queue=True)