This commit is contained in:
mdivecky 2022-07-09 20:50:33 +02:00
parent d30baa49f6
commit d53dcaeb54
2 changed files with 16 additions and 14 deletions

View file

@ -14,3 +14,7 @@ feeds:
- name: "novinky.cz"
rss_source: "https://www.novinky.cz/rss"
unique_tag: "guid"
- name: "CTK"
rss_source: "https://www.ceskenoviny.cz/sluzby/rss/zpravy.php"
unique_tag: "guid"

View file

@ -8,7 +8,6 @@ import json
import imgkit
from diff_match_patch import diff_match_patch
import difflib
from pprint import pprint
import hashlib
@ -29,21 +28,20 @@ dmp = diff_match_patch()
rc = redis.Redis(host='localhost', port=6379, db=0)
image_options = {
'width': '450',
}
article_count = 0
def write_article(article, rc):
rval = json.dumps(article['content'])
rc.set(article['rss_id'], rval)
def process_diff(diff, article):
def process_diff(old, new, rss_id):
diff = dmp.diff_main(old['title'], new['title'])
dmp.diff_cleanupSemantic(diff)
html_diff = dmp.diff_prettyHtml(diff)
filename = hashlib.md5(article['rss_id'].encode()).hexdigest() + ".jpg"
image = imgkit.from_string(html_diff, filename, options = {'width': '450'})
print(old['link'])
#print(diff)
#filename = "./img/" + hashlib.md5(rss_id.encode()).hexdigest() + ".jpg"
#image = imgkit.from_string(html_diff, filename, options = {'width': '450', 'quiet': ''})
return(True)
@ -52,10 +50,9 @@ def process_item(article, rc):
old = json.loads(rc.get(article['rss_id']))
new = article['content']
if old['title'] != new['title']:
print('Article changed. Fuck the world.')
diff = dmp.diff_main(old['title'], new['title'])
process_diff(diff, article)
#write_article(article_rc)
print('Article changed. World is fucked.')
diff = process_diff(old, new, article['rss_id'])
#write_article(article, rc)
return(True)
else:
# Article is the same. All good!
@ -66,7 +63,7 @@ def process_item(article, rc):
article_count = 0
for feed in config['feeds']:
rss_source = str(feed['rss_source'])
unique_tag = str(feed['unique_tag'])
@ -93,5 +90,6 @@ for feed in config['feeds']:
}
article_count += 1
process_item(article, rc)
print("Processed articles:")
print(article_count)