import project from david with uv
This commit is contained in:
commit
8056704736
11 changed files with 1530 additions and 0 deletions
55
pygea/main.py
Normal file
55
pygea/main.py
Normal file
|
|
@ -0,0 +1,55 @@
|
|||
"""Pygea main entry point"""
|
||||
import hashlib
|
||||
from pygea.pangeafeed import PangeaFeed
|
||||
from pygea.pexception import PangeaServiceException
|
||||
|
||||
def main():
|
||||
# Feeds are generated for a single, specified, domain
|
||||
domain = 'www.martinoticias.com'
|
||||
|
||||
args = {
|
||||
# tuple values:
|
||||
# [0] category name or a string representing a content query
|
||||
# [1] only the newest content desired (as configured in pygea.ini)?
|
||||
# [2] special content_type for this category only (from the approved list of types)
|
||||
'categories': [
|
||||
('Titulares',True, None),
|
||||
('Cuba', True, None),
|
||||
('América Latina', True, None),
|
||||
('Info Martí ', False, None), # YES! this category name has a space character at the end!
|
||||
('Noticiero Martí Noticias', True, None)
|
||||
],
|
||||
'default_content_type': "articles"
|
||||
}
|
||||
|
||||
# TWO OPTIONS from the args defined above:
|
||||
# 1. Generate a single feed from the defined categories
|
||||
#try:
|
||||
# pf = PangeaFeed(domain, args)
|
||||
# pf.acquire_content()
|
||||
# pf.generate_feed()
|
||||
# pf.disgorge()
|
||||
#except PangeaServiceException as error:
|
||||
# print(error)
|
||||
|
||||
# 2. Generate different feeds for each defined category
|
||||
try:
|
||||
for cat_tuple in args['categories']:
|
||||
# form new args for each category/query
|
||||
newargs = {
|
||||
'categories': [cat_tuple],
|
||||
'default_content_type': "articles"
|
||||
}
|
||||
pf = PangeaFeed(domain, newargs)
|
||||
pf.acquire_content()
|
||||
pf.generate_feed()
|
||||
# put each feed into a different sub-directory
|
||||
feed_subdir = hashlib.md5(cat_tuple[0].encode('utf-8')).hexdigest()[:7]
|
||||
pf.disgorge(feed_subdir)
|
||||
print("feed for {} output to sub-directory {}".format(cat_tuple[0], feed_subdir))
|
||||
except PangeaServiceException as error:
|
||||
print(error)
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
main()
|
||||
Loading…
Add table
Add a link
Reference in a new issue