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

menu and submenu

Browse Source
This commit is contained in:
Valentin CZERYBA 2023-07-11 00:17:24 +02:00
parent 20c4adb3cf
commit 5059a15826
1 changed files with 6 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

10
lib/WPMenu.py
View File

@ -56,6 +56,7 @@ class WPMenu:
## Get from URL ## Get from URL
def fromUrl(self, canalblog): def fromUrl(self, canalblog):
self._canalblog = canalblog
try: try:
o = urlparse(canalblog) o = urlparse(canalblog)
o = o._replace(scheme=self._protocol_canalblog) o = o._replace(scheme=self._protocol_canalblog)
@ -105,11 +106,12 @@ class WPMenu:
for anchor in ul.find_all("li"): for anchor in ul.find_all("li"):
parent = anchor.find("a").get_text().replace(" \xa0", "") parent = anchor.find("a").get_text().replace(" \xa0", "")
href = anchor.find("a").get("href")
if href == "{0}://{1}/".format(self._protocol_canalblog, self._canalblog):
parent = "home"
itemMenu = {"id":"", "type":"", "title": parent, "children":list()} itemMenu = {"id":"", "type":"", "title": parent, "children":list()}
href = anchor.find("a").get("href")
if href == self._wordpress:
parent = "home"
#menu = self._child(parent, parent) #menu = self._child(parent, parent)
if href == "#": if href == "#":
li = anchor.find("ul").find_all("li") li = anchor.find("ul").find_all("li")
@ -131,4 +133,4 @@ class WPMenu:
menu[j]["children"].append({"id":"", "title":children[i]["title"], "parent": children[i]["parent"]}) menu[j]["children"].append({"id":"", "title":children[i]["title"], "parent": children[i]["parent"]})
for i in menu: for i in menu:
self._logger.info("{0} : Menu {1}".format(self._name)) self._logger.info("{0} : Menu {1} {2}".format(self._name, i["title"], len(i["children"])))