Spaces:
Sleeping
Sleeping
raphael825
commited on
Commit
·
60c2359
1
Parent(s):
3c69fe4
Change to test the json showing
Browse files
app.py
CHANGED
@@ -1,12 +1,57 @@
|
|
1 |
import json
|
2 |
import gradio as gr
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
3 |
|
4 |
|
5 |
-
|
6 |
-
|
|
|
|
|
7 |
test = json.load(read_file)
|
8 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
9 |
|
10 |
|
11 |
-
iface = gr.Interface(fn=
|
12 |
iface.launch()
|
|
|
1 |
import json
|
2 |
import gradio as gr
|
3 |
+
import Model
|
4 |
+
from pytube import YouTube
|
5 |
+
import whisper
|
6 |
+
import time
|
7 |
+
import pickle
|
8 |
+
from langchain.text_splitter import RecursiveCharacterTextSplitter
|
9 |
+
from langchain.schema.document import Document
|
10 |
|
11 |
|
12 |
+
# # ==
|
13 |
+
|
14 |
+
def get_title(link):
|
15 |
+
with open("book.json", "r", encoding='utf-8') as read_file:
|
16 |
test = json.load(read_file)
|
17 |
+
title = test[0]['item'][0]['title']
|
18 |
+
return link + title
|
19 |
+
|
20 |
+
|
21 |
+
def greet(link):
|
22 |
+
model_test = Model.get_title(link)
|
23 |
+
model_link = Model.greet(link)
|
24 |
+
result = model_test + model_link
|
25 |
+
return result
|
26 |
+
|
27 |
+
|
28 |
+
def youtube_text(link):
|
29 |
+
yt = YouTube(link)
|
30 |
+
yt.streams.filter(only_audio=True).first().download \
|
31 |
+
(output_path=".", filename="test.mp3")
|
32 |
+
|
33 |
+
start = time.time()
|
34 |
+
model = whisper.load_model("small")
|
35 |
+
result = model.transcribe("test.mp3")
|
36 |
+
end = time.time()
|
37 |
+
|
38 |
+
print(result["text"])
|
39 |
+
print(f"{end - start:.2f}sec")
|
40 |
+
|
41 |
+
text_splitter = RecursiveCharacterTextSplitter(
|
42 |
+
chunk_size=2000,
|
43 |
+
chunk_overlap=50,
|
44 |
+
length_function=len, )
|
45 |
+
|
46 |
+
docs = [Document(page_content=x) for x in text_splitter.split_text(result["text"])]
|
47 |
+
|
48 |
+
split_docs = text_splitter.split_documents(docs)
|
49 |
+
|
50 |
+
with open("split_example_small.pkl", "wb") as f:
|
51 |
+
pickle.dump(split_docs, f)
|
52 |
+
|
53 |
+
return docs[0]
|
54 |
|
55 |
|
56 |
+
iface = gr.Interface(fn=youtube_text, inputs='text', outputs='text')
|
57 |
iface.launch()
|