v4l3n71n/web_scrap
1
0
Fork 0
Code Issues Pull Requests Projects Releases 10 Wiki Activity

loop replace

Browse Source
This commit is contained in:
Valentin CZERYBA 2023-05-23 11:22:37 +02:00
parent 0fc6e78a18
commit d3ec7d147d
1 changed files with 6 additions and 8 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

14
lib/WPImport.py
View File

@ -506,6 +506,8 @@ class WPimport:
self._logger.debug("{0} : {1}".format(page.content)) self._logger.debug("{0} : {1}".format(page.content))
if title[len(title)-1] == " ": if title[len(title)-1] == " ":
title = title[:-1] title = title[:-1]
if title[0] == " ":
title = title[1:]
page_is_exist = False page_is_exist = False
for index in range(1,10): for index in range(1,10):
@ -531,14 +533,10 @@ class WPimport:
if len(i["title"]["rendered"]) == len(title): if len(i["title"]["rendered"]) == len(title):
title_rendered = i["title"]["rendered"] title_rendered = i["title"]["rendered"]
else: else:
title_rendered = i["title"]["rendered"].replace('’', "'") title_rendered = i["title"]["rendered"]
title_rendered = title_rendered.replace('–', '-') list_replace = {'’': "'", '–': '-', '…': '...', '« ': '"', ' »': '"', '« ': '"', ' »': '"', '’': "'", '"‘': "'"}
title_rendered = title_rendered.replace('…', '...') for old, new in list_replace.items():
title_rendered = title_rendered.replace('« ', '"') title_rendered = title_rendered.replace(old, new)
title_rendered = title_rendered.replace(' »', '"')
title_rendered = title_rendered.replace('« ', '"')
title_rendered = title_rendered.replace(' »', '"')
title_rendered = title_rendered.replace('’', "'")
self._logger.debug("{0} : Search title posts for |{2}| : |{1}|".format(self._name, title_rendered, title)) self._logger.debug("{0} : Search title posts for |{2}| : |{1}|".format(self._name, title_rendered, title))
self._logger.debug("{0} : SIze of title : {1} - {2}".format(self._name, len(title), len(title_rendered))) self._logger.debug("{0} : SIze of title : {1} - {2}".format(self._name, len(title), len(title_rendered)))
if title_rendered == title: if title_rendered == title: