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

debug level comment

Browse Source
This commit is contained in:
Valentin CZERYBA 2023-04-16 19:32:00 +02:00
parent 0e15e88f31
commit bd8ac241c1
1 changed files with 8 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

12
lib/WPImport.py
View File

@ -226,6 +226,7 @@ class WPimport:
## Check class name ## Check class name
def _hasClassName(self, tag, className): def _hasClassName(self, tag, className):
print(tag["class"])
for i in tag["class"]: for i in tag["class"]:
if i == className: if i == className:
return True return True
@ -243,8 +244,8 @@ class WPimport:
def _getComment(self, comment): def _getComment(self, comment):
comment_post = [] comment_post = []
for i in comment: for i in range(0, len(comment)):
comment_div = find_all("div", class_="comment_item")[0] comment_div = comment[i].find("div", class_="comment_item")
comment_item = comment_div.text.split("\n") comment_item = comment_div.text.split("\n")
footer = comment_div.find_all("div", class_="itemfooter") footer = comment_div.find_all("div", class_="itemfooter")
comment_author = footer[0].text.split(",")[0].replace("Posté par ", "") comment_author = footer[0].text.split(",")[0].replace("Posté par ", "")
@ -256,15 +257,17 @@ class WPimport:
comment_content = comment_content + "</p>" comment_content = comment_content + "</p>"
parent = -1 parent = -1
parent_class = "level-1" parent_class = "level-1"
if self._hasClassName(i, parent_class) is False: print(self._hasClassName(comment[i], parent_class))
if self._hasClassName(comment[i], parent_class) is False:
block = False block = False
className = self._getClassName(tag, "level-").split("-") className = self._getClassName(comment[i], "level-").split("-")
level = 1 level = 1
if len(className) > 0: if len(className) > 0:
level = int(className[1]) level = int(className[1])
for j in range(i-1, 0): for j in range(i-1, 0):
if block is False: if block is False:
levelName = "level-{0}".format(level - 1) levelName = "level-{0}".format(level - 1)
print(levelName)
if self._hasClassName(comment[j], levelName): if self._hasClassName(comment[j], levelName):
parent = j parent = j
parent_class = levelName parent_class = levelName
@ -334,6 +337,7 @@ class WPimport:
list_img.append(new_img) list_img.append(new_img)
comment_post = self._getComment(comment) comment_post = self._getComment(comment)
exit(0)
a = itemfooter[0].find_all("a", {"rel": True}) a = itemfooter[0].find_all("a", {"rel": True})
for i in a: for i in a: