can read email and click on the link

src/mail/mail_reader.py

@@ -0,0 +1,139 @@
+import email
+import imaplib
+import re
+from concurrent.futures import ThreadPoolExecutor
+from email.header import decode_header
+from email.message import Message
+
+from builtins import list
+
+from src import params
+from src.db.mongo_manager import MONGO_STORE_MANAGER
+from src.pojo.mail.mail_pojo import MailPojo
+from src.proxy.proxy_type import ProxyType
+from src.workers.link_validator import LinkValidator
+
+AOL_IMAP_SERVER = "imap.aol.com"
+VALIDATION_URL_SUBJECT = 'Validation de votre demande de rendez-vous'
+VALIDATION_URL_REGEX = """https:\/\/rendezvousparis.hermes.com\/client\/register\/[A-Z0-9]+\/validate.code=[A-Z0-9]+"""
+
+
+class MailReader():
+    def __init__(self, login, password):
+        self.login = login
+        self.password = password
+
+    def read_emails(self, email_number=0) -> list:
+        # create an IMAP4 class with SSL
+        mail_list = []
+        imap = imaplib.IMAP4_SSL(AOL_IMAP_SERVER)
+        # authenticate
+        imap.login(self.login, password)
+        status, messages = imap.select("INBOX")
+        # total number of emails
+        messages = int(messages[0])
+        for i in range(messages, 0, -1):
+            # fetch the email message by ID
+            res, msg = imap.fetch(str(i), "(RFC822)")
+            # res, msg = imap.fetch(str(i))
+            body = ''
+            for response in msg:
+                if isinstance(response, tuple):
+                    # parse a bytes email into a message object
+                    msg = email.message_from_bytes(response[1])
+                    # decode the email subject
+                    subject, subject_encoded = decode_header(msg["Subject"])[0]
+                    received_date = msg["Date"]
+                    if isinstance(subject, bytes):
+                        # if it's a bytes, decode to str
+                        subject = subject.decode(subject_encoded)
+                    # decode email sender
+                    from_address, subject_encoded = decode_header(msg.get("From"))[0]
+                    if isinstance(from_address, bytes):
+                        from_address = from_address.decode(subject_encoded)
+                    print("Subject:", subject)
+                    print("From:", from_address)
+                    # if the email message is multipart
+                    if msg.is_multipart():
+                        # iterate over email parts
+                        for part in msg.walk():
+                            try:
+                                # get the email body
+                                payloads = part.get_payload()
+                                if isinstance(payloads, list):
+                                    for payload in payloads:
+                                        if isinstance(payload, Message):
+                                            body = body + payload.get_payload(decode=True).decode("iso-8859-1")
+                                # print(body)
+                            except Exception as Error:
+                                print(Error)
+                    else:
+                        body = msg.get_payload(decode=True).decode()
+                        print(body)
+                    if VALIDATION_URL_SUBJECT in subject:
+                        mail = MailPojo(subject=subject, body=body, from_address=from_address)
+                        mail_list.append(mail)
+        # close the connection and logout
+        imap.close()
+        imap.logout()
+        return mail_list
+
+
+hermes_email = "no-reply@hermes.com"
+# account credentials
+# username = "appointment2022@aol.com"
+# password = "gyilpmvyyvlcaviq"
+
+
+username = "chenpeijun@aol.com"
+password = "ytifuwguknzifqyb"
+
+
+def clean(text):
+    # clean text for creating a folder
+    return "".join(c if c.isalnum() else "_" for c in text)
+
+
+def need_to_valid_url(url: str, successful_items) -> bool:
+    print("url is :" + url)
+    parts = url.split('/')
+    id = parts[5]
+    if len(id) == 6:
+        for item in successful_items:
+            if item.url_validated is not None:
+                print("id:{}, status:{} ".format(id, str(item.url_validated)))
+            if item.id == id:
+                if item.url_validated is not None:
+                    return not item.url_validated
+                else:
+                    # if url_validated is None
+                    return True
+        #         return True by default
+        return False
+    else:
+        print("id not valid:{}".format(id))
+        return False
+
+
+# check whether the url has already been clicked
+
+
+if __name__ == '__main__':
+    mail_reader = MailReader(username, password)
+    successful_items = MONGO_STORE_MANAGER.get_all_successful_items_for_day()
+    list = mail_reader.read_emails()
+    with ThreadPoolExecutor(max_workers=10) as executor:
+        for mail in list:
+            match = re.search(VALIDATION_URL_REGEX, mail.body)
+            if match:
+                url = match.group(0)
+                if need_to_valid_url(url, successful_items):
+                    url_validator = LinkValidator(url)
+                    print("need to validate url: " + url)
+                    # .start_page(params.get_proxy(ProxyType.BRIGHT_DATA))
+                    executor.submit(url_validator.start_page, params.get_proxy(ProxyType.BRIGHT_DATA))
+                else:
+                    print("do not need to click url --> {}".format(mail))
+
+    # find link from mails
+    print(list)