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

add author

Browse Source
This commit is contained in:
Valentin CZERYBA
2023-08-23 22:48:38 +02:00
parent 50bf31d334
commit d88ae7ed44
2 changed files with 16 additions and 16 deletions
Download Patch File Download Diff File Expand all files Collapse all files

13
lib/WPImport.py
View File

@@ -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, no_image=False, tmp="/tmp/import_export_canablog"):
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, tmp="/tmp/import_export_canablog", author=""):
self._name = name
self._basic = basic
self._wordpress = wordpress
@@ -29,6 +29,7 @@ class WPimport:
self._no_update = no_update
self._no_image = no_image
self._tmp = tmp
self._author = author
# Destructor
def __del__(self):
@@ -131,10 +132,10 @@ class WPimport:
new_img = {}
href_img = i.get("src")
href_img_o = urlparse(href_img)
new_img["old_src"]=href_img
try:
if len(href_img_o.netloc) > 0:
img_ok = False
href_img = href_img.replace("_q", "_o")
page_img = self._request.get(href_img)
if page_img.status_code == 200:
img_ok = True
@@ -158,14 +159,12 @@ class WPimport:
content_html = ""
if len(list_img) > 0:
content_html = "<!-- wp:paragraph --><p>{0}</p><!-- /wp:paragraph -->\n\n".format(albumdesc)
content_html = "<!-- wp:paragraph -->\n{0}\n<!-- /wp:paragraph -->\n\n".format(albumdesc)
content_html = content_html + "<figure class='wp-block-gallery has-nested-images columns-default is-cropped'><!-- wp:image {\"id\":id-image,'sizeSlug':'large','linkDestination':'none'} --><figure class='wp-block-image size-large'><img src='src-image' alt='' class=wp-image-id-image/></figure><!-- /wp:image --> \n\n".replace("id-image", str(list_img[0]["id"])).replace("src-image", list_img[0]["new_src"])
for i in range(1, len(list_img)):
content_html = content_html + "<!-- wp:image {\"id\":id-image,\"sizeSlug\":\"large\",\"linkDestination\":\"none\"} --><figure class=\"wp-block-image size-large\"><img src='src-image' alt=\"\" class='wp-image-id-image'/></figure><!-- /wp:image -->\n\n".replace("id-image", str(list_img[0]["id"])).replace("src-image", list_img[0]["new_src"])
content_html = content_html + "<!-- /wp:gallery -->"
for i in range(0, len(list_img)):
content_html = content_html + "<!-- wp:image {\"id\":id-image,\"sizeSlug\":\"large\",\"linkDestination\":\"none\"} --><figure class=\"wp-block-image size-large\"><img src=\"src-image\" alt=\"\" class=\"wp-image-id-image\"/></figure><!-- /wp:image -->\n\n".replace("id-image", str(list_img[i]["id"])).replace("src-image", list_img[i]["new_src"])
self._logger.info("{0} : content html : {1}".format(self._name, content_html))
if len(content_html) > 0: