add parameter no-image
This commit is contained in:
@@ -8,7 +8,7 @@ from requests.packages.urllib3.util.retry import Retry
|
||||
|
||||
class WPimport:
|
||||
# Constructor
|
||||
def __init__(self, name="Thread-0", basic=None, wordpress="", logger=None, parser="html.parser", ssl_wordpress=True, no_create=False, no_update=False):
|
||||
def __init__(self, name="Thread-0", basic=None, wordpress="", logger=None, parser="html.parser", ssl_wordpress=True, no_create=False, no_update=False, no_image=False):
|
||||
self._name = name
|
||||
self._basic = basic
|
||||
self._wordpress = wordpress
|
||||
@@ -26,6 +26,7 @@ class WPimport:
|
||||
self._request.mount('{0}://'.format(self._protocol), HTTPAdapter(max_retries=retries))
|
||||
self._no_create = no_create
|
||||
self._no_update = no_update
|
||||
self._no_image = no_image
|
||||
|
||||
# Destructor
|
||||
def __del__(self):
|
||||
@@ -436,48 +437,49 @@ class WPimport:
|
||||
img_a = articlebody[0].find_all("a", {"target": "_blank"})
|
||||
self._logger.debug("{0} : Number of image's link : {1}".format(self._name, len(img_a)))
|
||||
list_img = []
|
||||
for i in img_a:
|
||||
new_img = {}
|
||||
img = i.find_all("img")
|
||||
self._logger.debug("{0} : Number of image's tag : {1}".format(self._name, len(img)))
|
||||
if len(img) > 0:
|
||||
href_a = i.get("href")
|
||||
href_img = img[0].get("src")
|
||||
new_img["old_src"]=href_img
|
||||
new_img["old_href"]=href_a
|
||||
try:
|
||||
page_img = self._request.get(href_img)
|
||||
if self._no_image is False:
|
||||
for i in img_a:
|
||||
new_img = {}
|
||||
img = i.find_all("img")
|
||||
self._logger.debug("{0} : Number of image's tag : {1}".format(self._name, len(img)))
|
||||
if len(img) > 0:
|
||||
href_a = i.get("href")
|
||||
href_img = img[0].get("src")
|
||||
new_img["old_src"]=href_img
|
||||
new_img["old_href"]=href_a
|
||||
try:
|
||||
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)
|
||||
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)
|
||||
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:
|
||||
media=self._addOrUpdateMedia(href_img, page_img)
|
||||
new_img["id"]=media["id"]
|
||||
new_img["new_src"]=media["rendered"]
|
||||
list_img.append(new_img)
|
||||
if href_img != href_a:
|
||||
media=self._addOrUpdateMedia(href_a, page_img)
|
||||
if page_img.status_code == 404:
|
||||
href_img = href_a
|
||||
try:
|
||||
page_img = self._request.get(href_a)
|
||||
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)
|
||||
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:
|
||||
media=self._addOrUpdateMedia(href_img, page_img)
|
||||
new_img["id"]=media["id"]
|
||||
new_img["new_src"]=media["rendered"]
|
||||
list_img.append(new_img)
|
||||
if page_img.status_code not in [200, 404]:
|
||||
self._logger.error("{0} : Connection error with status code for get image : {1}".format(self._name, page_img.status_code))
|
||||
self._logger.debug("{0} : {1}".format(self._name, page_img.content))
|
||||
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)
|
||||
if href_img != href_a:
|
||||
media=self._addOrUpdateMedia(href_a, page_img)
|
||||
new_img["id"]=media["id"]
|
||||
new_img["new_src"]=media["rendered"]
|
||||
list_img.append(new_img)
|
||||
if page_img.status_code not in [200, 404]:
|
||||
self._logger.error("{0} : Connection error with status code for get image : {1}".format(self._name, page_img.status_code))
|
||||
self._logger.debug("{0} : {1}".format(self._name, page_img.content))
|
||||
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)
|
||||
|
||||
self._logger.debug("{0} : Number of image : {1}".format(self._name, len(list_img)))
|
||||
comment_post = self._getComment(comment)
|
||||
|
Reference in New Issue
Block a user