drewThomasson commited on
Commit
81cc7ab
·
verified ·
1 Parent(s): c13fcbb

Delete download_github_folder.py

Browse files
Files changed (1) hide show
  1. download_github_folder.py +0 -54
download_github_folder.py DELETED
@@ -1,54 +0,0 @@
1
- import os
2
- import requests
3
- from bs4 import BeautifulSoup
4
-
5
- # GitHub folder URL
6
- BASE_URL = "https://github.com/DrewThomasson/ebook2audiobook/tree/main/voices"
7
- RAW_BASE_URL = "https://raw.githubusercontent.com/DrewThomasson/ebook2audiobook/main/voices"
8
-
9
- # Directory to save downloaded files
10
- DOWNLOAD_DIR = "voices"
11
- os.makedirs(DOWNLOAD_DIR, exist_ok=True)
12
-
13
- def get_file_links(folder_url):
14
- """Get the raw file links from a GitHub folder."""
15
- response = requests.get(folder_url)
16
- if response.status_code != 200:
17
- print(f"Failed to fetch folder content: {response.status_code}")
18
- return []
19
-
20
- soup = BeautifulSoup(response.text, 'html.parser')
21
- links = []
22
- for a_tag in soup.find_all('a', class_='js-navigation-open'):
23
- href = a_tag.get('href')
24
- if href and '/blob/' in href: # File link
25
- file_path = href.split('/blob/')[-1]
26
- links.append(f"{RAW_BASE_URL}/{file_path}")
27
- return links
28
-
29
- def download_file(url, save_dir):
30
- """Download a file from a URL to the specified directory."""
31
- local_filename = os.path.join(save_dir, url.split('/')[-1])
32
- with requests.get(url, stream=True) as response:
33
- if response.status_code == 200:
34
- with open(local_filename, 'wb') as f:
35
- for chunk in response.iter_content(chunk_size=8192):
36
- f.write(chunk)
37
- print(f"Downloaded: {local_filename}")
38
- else:
39
- print(f"Failed to download {url}: {response.status_code}")
40
-
41
- def main():
42
- print("Fetching file links...")
43
- file_links = get_file_links(BASE_URL)
44
- if not file_links:
45
- print("No files found to download.")
46
- return
47
-
48
- print("Downloading files...")
49
- for file_link in file_links:
50
- download_file(file_link, DOWNLOAD_DIR)
51
-
52
- # Ensure the script runs only when executed directly
53
- if __name__ == "__main__":
54
- main()