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

add parent comment WIP

Browse Source
This commit is contained in:
Valentin CZERYBA 2023-04-14 23:10:07 +02:00
parent 1600a17383
commit b54785c455
1 changed files with 22 additions and 13 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

35
lib/WPImport.py
View File

@ -223,6 +223,25 @@ class WPimport:
if page.status_code == 201: if page.status_code == 201:
self._logger.info("Commentaire ajoute pour {0}".format(title)) self._logger.info("Commentaire ajoute pour {0}".format(title))
def _getComment(self, comment):
comment_post = []
for i in comment:
comment_div = find_all("div", class_="comment_item")[0]
comment_item = comment_div.text.split("\n")
footer = comment_div.find_all("div", class_="itemfooter")
comment_author = footer[0].text.split(",")[0].replace("Posté par ", "")
comment_date = footer[0].find_all("abbr")[0].get("title")
comment_content = "<p>"
for j in range(0, len(comment_item)-2):
if len(comment_item[j]) > 0:
comment_content = comment_content + comment_item[j] + "<br />"
comment_content = comment_content + "</p>"
parent = -1
parent_class = "level-1"
comment_post.append({"author": comment_author, "date": comment_date, "content": comment_content})
return comment_post
## Add or Update post ## Add or Update post
def _addOrUpdatePost(self, soup): def _addOrUpdatePost(self, soup):
@ -250,7 +269,7 @@ class WPimport:
articleacreator = soup.find_all("span", class_="articlecreator") articleacreator = soup.find_all("span", class_="articlecreator")
dateheader = soup.find_all("div", class_="dateheader") dateheader = soup.find_all("div", class_="dateheader")
itemfooter = soup.find_all("div", class_="itemfooter") itemfooter = soup.find_all("div", class_="itemfooter")
comment = soup.find_all("div", class_="comment_item") comment = soup.find_all("li", class_="comment")
img_a = articlebody[0].find_all("a", {"target": "_blank"}) img_a = articlebody[0].find_all("a", {"target": "_blank"})
list_img = [] list_img = []
for i in img_a: for i in img_a:
@ -284,18 +303,8 @@ class WPimport:
new_img["new_src"]=media["rendered"] new_img["new_src"]=media["rendered"]
list_img.append(new_img) list_img.append(new_img)
comment_post = [] comment_post = self._getComment(comment)
for i in comment:
comment_item = i.text.split("\n")
footer = i.find_all("div", class_="itemfooter")
comment_author = footer[0].text.split(",")[0].replace("Posté par ", "")
comment_date = footer[0].find_all("abbr")[0].get("title")
comment_content = "<p>"
for j in range(0, len(comment_item)-2):
if len(comment_item[j]) > 0:
comment_content = comment_content + comment_item[j] + "<br />"
comment_content = comment_content + "</p>"
comment_post.append({"author": comment_author, "date": comment_date, "content": comment_content})
a = itemfooter[0].find_all("a", {"rel": True}) a = itemfooter[0].find_all("a", {"rel": True})
for i in a: for i in a:
rel = i.get("rel") rel = i.get("rel")