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

distinct tags and categories

Browse Source
This commit is contained in:
Valentin CZERYBA 2023-03-27 23:51:51 +02:00
parent 3622e37942
commit 0c41dc3e65
1 changed files with 33 additions and 25 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

58
insert_wordpress.py
View File

@ -3,7 +3,7 @@ from bs4 import BeautifulSoup
from urllib.parse import urlparse
from requests.auth import HTTPBasicAuth
from getpass import getpass
import requests, os, argparse, logging
import requests, os, argparse, logging, re
if __name__ == '__main__':
@ -19,13 +19,19 @@ if __name__ == '__main__':
exit(1)
basic = HTTPBasicAuth(args.user, password)
page = requests.get("http://localhost:8080/wp-json/wp/v2/tags")
if page.status_code == 200:
tags = page.json()
print(tags)
liste = ["categories", "tags"]
elements = {}
element = {}
listelement = {}
for i in liste:
page = requests.get("http://localhost:8080/wp-json/wp/v2/{0}".format(i))
if page.status_code == 200:
elements[i] = page.json()
element[i] = []
listelement[i] = []
with open(args.file, 'r') as f:
contents = f.read()
@ -36,34 +42,36 @@ if __name__ == '__main__':
dateheader = soup.find_all("div", class_="dateheader")
itemfooter = soup.find_all("div", class_="itemfooter")
a = itemfooter[0].find_all("a", {"rel": True})
tag = []
for i in a:
rel = i.get("rel")
if rel[0] == 'tag':
tag.append(i.text)
listtag = []
for i in tag:
tag_exist = False
for j in tags:
if j["name"] == i:
tag_exist = True
listtag.append(j["id"])
if tag_exist is False:
data = {"name": i}
page = requests.post("http://localhost:8080/wp-json/wp/v2/tags", auth=basic, data=data)
if page.status_code == 201:
result = page.json()
listtag.append(result["id"])
href = i.get("href")
if re.search(r'/tag/', href):
element["tags"].append(i.text)
if re.search(r'/archives/', href):
element["categories"].append(i.text)
for i in liste:
for j in element[i]:
element_exist = False
for k in elements[i]:
if k["name"] == j:
element_exist = True
array = listelement[i].append(k["id"])
if element_exist is False:
data = {"name": j}
page = requests.post("http://localhost:8080/wp-json/wp/v2/{0}".format(i), auth=basic, data=data)
if page.status_code == 201:
result = page.json()
listelement[i].append(result["id"])
title = articletitle[0].text
body = articlebody[0]
hour = articledate[0].text
time = dateheader[0].text.split(" ")
data = {"title":title, "content":body, "status":"publish", "date": "{0}-{1}-{2}T{3}:00".format(time[2],month[time[1]],time[0], hour), "tags": listtag}
print(data)
exit(0)
data = {"title":title, "content":body, "status":"publish", "date": "{0}-{1}-{2}T{3}:00".format(time[2],month[time[1]],time[0], hour), "tags": listelement["tags"], "categories": listelement["categories"]}
page = requests.post("http://localhost:8080/wp-json/wp/v2/posts", auth=basic, data=data)
print(page.status_code)