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

replace print by logger

Browse Source
This commit is contained in:
Valentin CZERYBA 2023-04-08 12:17:43 +02:00
parent 9ab33c169e
commit d58ead52b2
1 changed files with 42 additions and 13 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

55
insert_wordpress.py
View File

@ -7,9 +7,10 @@ import requests, os, argparse, logging, re, json
class WPimport: class WPimport:
# Constructor # Constructor
def __init__(self, basic, wordpress): def __init__(self, basic, wordpress, logger):
self._basic = basic self._basic = basic
self._wordpress = wordpress self._wordpress = wordpress
self._logger = logger
# Public method # Public method
@ -23,7 +24,7 @@ class WPimport:
def fromFile(self, files): def fromFile(self, files):
for file in files: for file in files:
if os.path.exists(file): if os.path.exists(file):
print(file) logger.info("Fichier en cours de traitement : {0}".format(file))
with open(file, 'r') as f: with open(file, 'r') as f:
content = f.read() content = f.read()
soup = BeautifulSoup(content, 'html.parser') soup = BeautifulSoup(content, 'html.parser')
@ -84,9 +85,9 @@ class WPimport:
data = {"featured_media": id_media} 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)) 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: if r.status_code == 200:
print("Ajout media featured : {0}".format(r.json()["title"]["raw"])) logger.info("Ajout media featured : {0}".format(r.json()["title"]["raw"]))
else: else:
print("Aucun media trouvé pour {0}".format(h2)) logger.info("Aucun media trouvé pour {0}".format(h2))
## Association image to post ## Association image to post
@ -95,7 +96,7 @@ class WPimport:
data = {"post": post_id} data = {"post": post_id}
r = requests.post("http://{0}/wp-json/wp/v2/media/{1}".format(self._wordpress, i["id"]), auth=self._basic, data=data) r = requests.post("http://{0}/wp-json/wp/v2/media/{1}".format(self._wordpress, i["id"]), auth=self._basic, data=data)
if r.status_code == 200: if r.status_code == 200:
print("Association d'une image à l'article {0}".format(title)) logger.info("Association d'une image à l'article {0}".format(title))
## Add or update img ## Add or update img
@ -111,7 +112,7 @@ class WPimport:
params = {"force":1} params = {"force":1}
r = requests.delete("http://{0}/wp-json/wp/v2/media/{1}".format(self._wordpress, res[0]["id"]), auth=self._basic, params=params) r = requests.delete("http://{0}/wp-json/wp/v2/media/{1}".format(self._wordpress, res[0]["id"]), auth=self._basic, params=params)
if r.status_code == 200: if r.status_code == 200:
print("Image supprimé {0}".format(img_name)) logger.info("Image supprimé {0}".format(img_name))
data = page.content data = page.content
img_type = "image/png" img_type = "image/png"
if img_name.split(".")[1] == "jpg" or img_name.split(".")[1] == "jpeg": if img_name.split(".")[1] == "jpg" or img_name.split(".")[1] == "jpeg":
@ -119,7 +120,7 @@ class WPimport:
headers={ 'Content-Type': img_type,'Content-Disposition' : 'attachment; filename={0}'.format(img_name)} headers={ 'Content-Type': img_type,'Content-Disposition' : 'attachment; filename={0}'.format(img_name)}
r = requests.post("http://{0}/wp-json/wp/v2/media".format(self._wordpress), auth=self._basic, headers=headers, data=data) r = requests.post("http://{0}/wp-json/wp/v2/media".format(self._wordpress), auth=self._basic, headers=headers, data=data)
if r.status_code == 201: if r.status_code == 201:
print("Ajout d'image {0}".format(img_name)) logger.info("Ajout d'image {0}".format(img_name))
res = r.json() res = r.json()
media["id"] = res["id"] media["id"] = res["id"]
media["rendered"] = res["guid"]["rendered"] media["rendered"] = res["guid"]["rendered"]
@ -241,12 +242,12 @@ class WPimport:
if len(result) == 0: if len(result) == 0:
page_exist = False page_exist = False
else: else:
print("La page {0} existe deja et mis à jour".format(title)) logger.info("La page {0} existe deja et mis à jour".format(title))
post_id = result[0]["id"] post_id = result[0]["id"]
page = requests.post("http://{0}/wp-json/wp/v2/posts/{1}".format(self._wordpress, post_id), auth=self._basic, headers=headers, data=json.dumps(data)) page = requests.post("http://{0}/wp-json/wp/v2/posts/{1}".format(self._wordpress, post_id), auth=self._basic, headers=headers, data=json.dumps(data))
if page.status_code == 200: if page.status_code == 200:
result = page.json() result = page.json()
print("Article mis à jour : {0}".format(result["title"]["raw"])) logger.info("Article mis à jour : {0}".format(result["title"]["raw"]))
self._linkImgPost(result["title"]["raw"], list_img, result["id"]) self._linkImgPost(result["title"]["raw"], list_img, result["id"])
@ -255,12 +256,12 @@ class WPimport:
page = requests.post("http://{0}/wp-json/wp/v2/posts".format(self._wordpress), auth=self._basic, headers=headers, data=json.dumps(data)) page = requests.post("http://{0}/wp-json/wp/v2/posts".format(self._wordpress), auth=self._basic, headers=headers, data=json.dumps(data))
if page.status_code == 201: if page.status_code == 201:
result = page.json() result = page.json()
print("Article ajoute : {0}".format(result["title"]["raw"])) logger.info("Article ajoute : {0}".format(result["title"]["raw"]))
for i in comment_post: for i in comment_post:
data = {"post": result["id"], "content": i["content"], "date": i["date"], "author_name": i["author"]} data = {"post": result["id"], "content": i["content"], "date": i["date"], "author_name": i["author"]}
page = requests.post("http://{0}/wp-json/wp/v2/comments".format(self._wordpress), auth=self._basic, data=data) page = requests.post("http://{0}/wp-json/wp/v2/comments".format(self._wordpress), auth=self._basic, data=data)
if page.status_code == 201: if page.status_code == 201:
print("Commentaire ajoute pour {0}".format(result["title"]["raw"])) logger.info("Commentaire ajoute pour {0}".format(result["title"]["raw"]))
self._linkImgPost(result["title"]["raw"], list_img, result["id"]) self._linkImgPost(result["title"]["raw"], list_img, result["id"])
@ -274,14 +275,42 @@ if __name__ == '__main__':
parser.add_argument("--file", help="HTML file", default="") parser.add_argument("--file", help="HTML file", default="")
parser.add_argument("--directory", help="HTML directory", default="") parser.add_argument("--directory", help="HTML directory", default="")
parser.add_argument("--wordpress", help="URL Wordpress", required=True) parser.add_argument("--wordpress", help="URL Wordpress", required=True)
parser.add_argument("--debug", help="Verbosity", action="store_true")
parser.add_argument("--logfile", help="Log file", default="")
parser.add_argument("--quiet", help="No console output", action="store_true")
args = parser.parse_args() args = parser.parse_args()
logger = logging.getLogger('insert wordpress')
formatter = logging.Formatter('%(asctime)s - %(name)s - %(levelname)s - %(message)s')
if args.quiet is False:
ch = logging.StreamHandler()
if args.debug is True:
logger.setLevel(logging.DEBUG)
ch.setLevel(logging.DEBUG)
else:
logger.setLevel(logging.INFO)
ch.setLevel(logging.INFO)
ch.setFormatter(formatter)
logger.addHandler(ch)
if len(args.logfile) > 0:
fileHandler = logging.FileHandler(args.logfile)
if args.debug is True:
fileHandler.setLevel(logging.DEBUG)
else:
fileHandler.setLevel(logging.INFO)
fileHandler.setFormatter(formatter)
logger.addHandler(fileHandler)
password = getpass() password = getpass()
if len(password) == 0: if len(password) == 0:
print("No password error !!! ") logger.error("No password error !!! ")
exit(1) exit(1)
basic = HTTPBasicAuth(args.user, password) basic = HTTPBasicAuth(args.user, password)
importWp = WPimport(basic, args.wordpress) importWp = WPimport(basic, args.wordpress, logger)
if len(args.file) > 0: if len(args.file) > 0:
importWp.fromFile(args.file.split(",")) importWp.fromFile(args.file.split(","))
exit(0) exit(0)