Create app.py
Browse files
app.py
ADDED
@@ -0,0 +1,113 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import gradio as gr
|
2 |
+
import requests
|
3 |
+
import json
|
4 |
+
import os
|
5 |
+
|
6 |
+
def send_req(title, audio_file_path):
|
7 |
+
url = "http://https://dev-phonic-api.vuihoc.vn/api/v3/get_score_from_file"
|
8 |
+
|
9 |
+
payload = {'title': title}
|
10 |
+
files=[
|
11 |
+
('audio',('temp.wav',open(audio_file_path,'rb'),'audio/wav'))
|
12 |
+
]
|
13 |
+
headers = {
|
14 |
+
'accept': 'application/json',
|
15 |
+
'Authorization': 'Bearer eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.eyJpZCI6ODg2MSwiZGV2aWNlSWQiOjg4NjEsImlhdCI6MTcxMTk2MTY2OSwiZXhwIjoxNzE3MTQ1NjY5fQ.8Wtzyx3mKVh7K9_GNzseWdK1NH-hycYdNh1uFsoVvEg'
|
16 |
+
}
|
17 |
+
|
18 |
+
response = requests.request("POST", url, headers=headers, data=payload, files=files)
|
19 |
+
|
20 |
+
return response.json()
|
21 |
+
|
22 |
+
def create_html_output(real_transcript, is_letter_correct_all_words):
|
23 |
+
html_output = "<center><h1>"
|
24 |
+
for i, char in enumerate(real_transcript):
|
25 |
+
if is_letter_correct_all_words[i] == '1':
|
26 |
+
html_output += f"<span style='color:green;'>{char}</span>"
|
27 |
+
else:
|
28 |
+
html_output += f"<span style='color:red;'>{char}</span>"
|
29 |
+
html_output += "</h1></center>"
|
30 |
+
return html_output
|
31 |
+
|
32 |
+
def create_html_output_ipa(word_score_list):
|
33 |
+
html_output = "<center><h1>"
|
34 |
+
for word_score in word_score_list:
|
35 |
+
for phone_score in word_score["phone_score_list"]:
|
36 |
+
if phone_score["quality_score"] == 100:
|
37 |
+
html_output += f"<span style='color:green;'>{phone_score['phone_ipa']}</span>"
|
38 |
+
else:
|
39 |
+
html_output += f"<span style='color:red;'>{phone_score['phone_ipa']}</span>"
|
40 |
+
html_output += "<span style='color:red;'> </span>"
|
41 |
+
html_output += "</h1></center>"
|
42 |
+
return html_output
|
43 |
+
|
44 |
+
def download_audio_file(url, filename=None):
|
45 |
+
"""
|
46 |
+
Tải xuống tệp âm thanh từ một URL và lưu nó vào đĩa.
|
47 |
+
|
48 |
+
Args:
|
49 |
+
url (str): URL của tệp âm thanh cần tải xuống.
|
50 |
+
filename (str, optional): Tên tệp để lưu (mặc định là tên tệp từ URL).
|
51 |
+
"""
|
52 |
+
|
53 |
+
|
54 |
+
response = requests.get(url)
|
55 |
+
response.raise_for_status() # Kiểm tra lỗi HTTP
|
56 |
+
|
57 |
+
# Nếu không cung cấp tên tệp, sử dụng tên tệp từ URL
|
58 |
+
if not filename:
|
59 |
+
filename = url.split("/")[-1]
|
60 |
+
|
61 |
+
with open(filename, "wb") as f:
|
62 |
+
f.write(response.content)
|
63 |
+
|
64 |
+
|
65 |
+
def pa_check(microphone, file_upload, reference_text):
|
66 |
+
if (microphone is not None) and (file_upload is not None):
|
67 |
+
warn_output = (
|
68 |
+
"WARNING: You've uploaded an audio file and used the microphone. "
|
69 |
+
"The recorded file from the microphone will be used and the uploaded audio will be discarded.\n"
|
70 |
+
)
|
71 |
+
|
72 |
+
elif (microphone is None) and (file_upload is None):
|
73 |
+
return "ERROR: You have to either use the microphone or upload an audio file"
|
74 |
+
|
75 |
+
file = microphone if microphone is not None else file_upload
|
76 |
+
|
77 |
+
result = send_req(reference_text, file)
|
78 |
+
try:
|
79 |
+
|
80 |
+
html_output = create_html_output(result["data"]["real_transcripts"], result["data"]["is_letter_correct_all_words"])
|
81 |
+
html_output_ipa = create_html_output_ipa(result["data"]["word_score_list"])
|
82 |
+
except Exception as e:
|
83 |
+
print(e)
|
84 |
+
print(result["data"]["real_transcripts"])
|
85 |
+
print(result["data"]["is_letter_correct_all_words"])
|
86 |
+
html_output = "ERROR: Something went wrong with the server response. Please try again later."
|
87 |
+
|
88 |
+
|
89 |
+
return json.dumps(result, indent=4, ensure_ascii=False), html_output, html_output_ipa
|
90 |
+
|
91 |
+
|
92 |
+
demo = gr.Interface(
|
93 |
+
fn=pa_check,
|
94 |
+
inputs=[
|
95 |
+
# gr.Textbox(label="Url audio", type="text", placeholder="Download audio form url"),
|
96 |
+
gr.Audio(sources="microphone", type="filepath"),
|
97 |
+
gr.Audio(sources="upload", type="filepath"),
|
98 |
+
gr.Textbox(label="Reference text", type="text", placeholder="How are you?|What is your name?"),
|
99 |
+
],
|
100 |
+
outputs=[
|
101 |
+
gr.Textbox(label="Output"),
|
102 |
+
"html",
|
103 |
+
"html"
|
104 |
+
],
|
105 |
+
theme="huggingface",
|
106 |
+
title="Pronunciation Assessment",
|
107 |
+
allow_flagging="never",
|
108 |
+
examples=[[None, None, "How are you?|What is your name?"], [None, None, "It's at quarter past nine.|It's at nine fifteen."]]
|
109 |
+
)
|
110 |
+
|
111 |
+
demo.launch(auth=(os.environ['username'], os.environ['password']))
|
112 |
+
|
113 |
+
|