moved update and download functionality to kontor-schema

This commit is contained in:
Thomas Peetz
2025-01-28 15:10:10 +01:00
parent c61e49720e
commit e733fa21e6
14 changed files with 118 additions and 162 deletions
+37 -2
View File
@@ -1,3 +1,8 @@
import re
import subprocess
from datetime import datetime
from pathlib import Path
import requests
from bs4 import BeautifulSoup
from sqlalchemy import Column, DateTime, Integer, String
@@ -26,9 +31,39 @@ class MediaFile(Base, BaseMixin, BaseVideoMixin):
except:
self.title = None
self.review = 1
self.last_modified_date = datetime.now()
def download_file(self, download_dir: str):
print(f"download file for {self.url}")
def download_file(self, download_dir: str, dl_tool: str):
print(f"download file for {self.url} to {download_dir}")
result = subprocess.run([dl_tool, self.url], cwd=download_dir, capture_output=True, text=True)
if result.returncode == 0:
output = result.stdout
output = re.sub(' +', ' ', output)
lines_list = output.splitlines()
file_name = self.__parse_output__(lines_list)
if file_name is None:
self.review = 1
self.should_download = 1
self.file_name = None
else:
download_file = Path(file_name)
self.should_download = 0
self.file_name = download_file.name
self.cloud_link = str(download_file.absolute())
self.last_modified_date = datetime.now()
def __parse_output__(self, lines_list):
self.file_name = None
for line in lines_list:
if 'has already been downloaded' in line:
end_len = len(' has already been downloaded')
self.file_name = line[11:-end_len]
if 'Destination' in line:
line_len = len(line)
start_len = len('[download] Destination: ')
file_len = line_len - start_len
self.file_name = line[-file_len:]
return self.file_name
class MediaArticle(Base, BaseMixin):