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

add or update featured media

Browse Source
This commit is contained in:
Valentin CZERYBA 2023-04-03 23:45:48 +02:00
parent c9b1264153
commit 501876dac2
1 changed files with 47 additions and 7 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

54
insert_wordpress.py
View File

@ -20,8 +20,46 @@ class WPimport:
articlebody = soup.find_all("div", class_="articlebody") articlebody = soup.find_all("div", class_="articlebody")
if len(articlebody) > 0: if len(articlebody) > 0:
self._addOrUpdatePost(soup) self._addOrUpdatePost(soup)
else:
self._addOrUpdateFeaturedMedia(soup)
# Private method # Private method
## Add or update featured media
## Get or update featured image
def _addOrUpdateFeaturedMedia(self, soup):
item_div = soup.find_all("div", {"data-edittype": "post"})
for i in item_div:
h2 = i.find_all("h2")[0].text
params = {"search":h2, "type":"post"}
page = requests.get("http://{0}/wp-json/wp/v2/search".format(self._wordpress), auth=self._basic, params=params)
if page.status_code == 200:
result = page.json()
if len(result) > 0:
if h2 == result[0]["title"]:
img = i.find_all("img")
if len(img) > 0:
img_src = img[0].get("src")
page = requests.get(img_src)
if page.status_code == 200:
name_img = img_src.replace("_q", "")
name_img = name_img.split("/")[len(name_img.split("/"))-1]
params = {"search": name_img}
page = requests.get("http://{0}/wp-json/wp/v2/media".format(self._wordpress), auth=self._basic, params=params)
if page.status_code == 200:
res = page.json()
if len(res) > 0:
id_media = res[0]["id"]
headers = {'Content-Type': 'application/json', 'Accept':'application/json'}
data = {"featured_media": id_media}
r = requests.post("http://{0}/wp-json/wp/v2/posts/{1}".format(self._wordpress, result[0]["id"]), auth=self._basic, headers=headers, data=json.dumps(data))
if r.status_code == 200:
print("Ajout media featured : {0}".format(r.json()["title"]["raw"]))
else:
print("Aucun media trouvé")
## Association image to post ## Association image to post
@ -93,17 +131,19 @@ class WPimport:
new_img["old_src"]=href_img new_img["old_src"]=href_img
new_img["old_href"]=href_a new_img["old_href"]=href_a
page_img = requests.get(href_img) page_img = requests.get(href_img)
img_break = False
if page_img.status_code == 404: if page_img.status_code == 404:
href_img = href_a href_img = href_a
img_break = True page_img = requests.get(href_a)
page = requests.get(href_a) if page_img.status_code == 200:
if page.status_code == 200: media=self._addOrUpdateMedia(href_img, page_img)
media=self._addOrUpdateMedia(href_img, page)
new_img["id"]=media["id"] new_img["id"]=media["id"]
new_img["new_src"]=media["rendered"] new_img["new_src"]=media["rendered"]
new_img["break"]=img_break
list_img.append(new_img) list_img.append(new_img)
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)
comment_post = [] comment_post = []
for i in comment: for i in comment:
@ -156,7 +196,7 @@ class WPimport:
bodyhtml = bodyhtml.replace(i["old_src"], o.path) bodyhtml = bodyhtml.replace(i["old_src"], o.path)
hour = articledate[0].text hour = articledate[0].text
time = dateheader[0].text.split(" ") time = dateheader[0].text.split(" ")
data = {"title":title, "content":bodyhtml, "status":"publish", "date": "{0}-{1}-{2}T{3}:00".format(time[2],month[time[1]],time[0], hour), "tags": listelement["tags"], "categories": listelement["categories"], "featured_media":list_img[0]["id"]} data = {"title":title, "content":bodyhtml, "status":"publish", "date": "{0}-{1}-{2}T{3}:00".format(time[2],month[time[1]],time[0], hour), "tags": listelement["tags"], "categories": listelement["categories"]}
params = {"search":author} params = {"search":author}
page = requests.get("http://{0}/wp-json/wp/v2/users".format(self._wordpress), auth=self._basic, params=params) page = requests.get("http://{0}/wp-json/wp/v2/users".format(self._wordpress), auth=self._basic, params=params)
if page.status_code == 200: if page.status_code == 200: