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

Merge pull request 'diff-img' (#19) from diff-img into master

Browse Source 
Reviewed-on: #19
This commit is contained in:
v4l3n71n 2023-06-16 22:08:50 +00:00
commit 703cc8922a
1 changed files with 29 additions and 12 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

19
lib/WPImport.py
View File

@ -16,6 +16,7 @@ class WPimport:
self._parser = parser
self._headers_json = {'Content-Type': 'application/json; charset=utf-8', 'Accept':'application/json'}
self._protocol = "https"
self._directory = "backup"
if ssl_wordpress is False:
self._protocol = "http"
self._request = requests.Session()
@ -60,6 +61,7 @@ class WPimport:
def fromDirectory(self, directory="", number_thread=1, max_thread=1):
self._directory = directory
directory = "{0}/archives".format(directory)
directories = self._getDirectories([], "{0}".format(directory))
if len(directories) > 0:
@ -445,23 +447,38 @@ class WPimport:
if len(img) > 0:
href_a = i.get("href")
href_img = img[0].get("src")
href_a_o = urlparse(href_a)
href_img_o = urlparse(href_img)
new_img["old_src"]=href_img
new_img["old_href"]=href_a
try:
if len(href_img_o.netloc) > 0:
img_ok = False
page_img = self._request.get(href_img)
if page_img.status_code == 404:
href_img = href_a
try:
page_img = self._request.get(href_a)
if page_img.status_code == 200:
img_ok = True
except ConnectionError as err:
self._logger.error("{0} : Connection error for get image : {1}".format(self._name, err))
exit(1)
except Exception as err:
self._logger.error("{0} : Exception error for get image : {1}".format(self._name, err))
exit(1)
else:
if os.path.exists("{0}/..{1}".format(self._directory, href_img)):
page_img = open("{0}/..{1}".format(self._directory, href_img), "r")
img_ok = True
else:
if os.path.exists("{0}/..{1}".format(self._directory, href_a)):
page_img = open("{0}/..{1}".format(self._directory, href_a), "r")
img_ok = True
self._logger.debug("{0} : Status code for image {1} : {2}".format(self._name, href_img, page_img.status_code))
if page_img.status_code == 200:
if img_ok is True:
media=self._addOrUpdateMedia(href_img, page_img)
new_img["id"]=media["id"]
new_img["new_src"]=media["rendered"]