Batch querying every 10 min and processing new messages

crawler/crawler/main.py

@@ -2,10 +2,13 @@ from datetime import date
 from datetime import datetime
 from datetime import timedelta
 from getpass import getpass
+from time import sleep
 import email
 import json
 import os
 
+from dateutil import parser
+from dateutil.tz import tzutc
 from imbox import Imbox
 import requests
 
@@ -23,6 +26,7 @@ class MailCrawler(object):
         self.imap_pass = os.environ['IMAP_PASS']
 
     def get_parsers(self):
+        """Retrieves a list of parser hosts"""
         if self.parser_hosts is None:
             self.parser_hosts = []
             parser_format = 'PARSER_{}'
@@ -36,6 +40,7 @@ class MailCrawler(object):
         return self.parser_hosts
 
     def parse_message(self, message):
+        """Parses tokens from an email message"""
         text = self.get_email_text(message)
         if not text:
             print('No email text returned')
@@ -57,6 +62,7 @@ class MailCrawler(object):
         return results
 
     def get_server(self):
+        """Returns an active IMAP server"""
         return Imbox(
             self.imap_url,
             username=self.imap_user,
@@ -65,13 +71,15 @@ class MailCrawler(object):
         )
 
     def get_email_text(self, message):
+        """Retrieves the text body of an email message"""
         body = message.body.get('plain') or message.body.get('html')
         if not body:
             return None
         # Concat all known body content together since it doesn't really matter
         return ''.join([text for text in body if isinstance(text, str)])
 
-    def index_message(self, message):
+    def index_token(self, message):
+        """Sends a token from the parser to the indexer"""
         response = requests.post(
             self.indexer_host+'/token',
             json=message,
@@ -79,24 +87,61 @@ class MailCrawler(object):
         response.raise_for_status()
         return response.json()
 
+    def process_message(self, message):
+        """Process a single email message"""
+        for result in self.parse_message(message):
+            result.update({
+                'subject': message.subject,
+            })
+            print('Parsed result: ', result)
+            print('Indexed result: ', self.index_token(result))
+
+
+    def process_messages(self, server, since_date, last_message=0):
+        for uid, message in server.messages(date__gt=since_date):
+            uid = int(uid)
+            if uid <= last_message:
+                print('DDB Already seen message with uid {}. Skipping'.format(uid))
+                continue
+
+            print(
+                'Processing message uid {} message_id {} '
+                'with subject "{}"'.format(
+                    uid, message.message_id, message.subject
+                )
+            )
+            self.process_message(message)
+
+            # Update since_date
+            message_date = parser.parse(message.date)
+            print('DDB Processed message. Message date: {} Old date: {}'.format(
+                message_date, since_date
+            ))
+            since_date = max(since_date, message_date)
+            print('DDB Since date is now ', since_date)
+            last_message = max(uid, last_message)
+
+        return since_date, last_message
+
+
     def run(self):
         print('Starting crawler')
-
+        # TODO: Put server into some kind of context manager and property
         with self.get_server() as server:
-            since_date = datetime.now() - timedelta(days=30)
+            # TODO: parameterize startup date, maybe relative
-            for uid, message in server.messages(date__gt=since_date):
+            since_date = datetime.now(tzutc()) - timedelta(days=1)
-                print(
+            last_message = 0
-                    'Processing message uid {} message_id {} '
+            while True:
-                    'with subject "{}"'.format(
+                print('Lets process')
-                        uid, message.message_id, message.subject
+                since_date, last_message = self.process_messages(
-                    )
+                    server,
+                    since_date,
+                    last_message=last_message
                 )
-                for result in self.parse_message(message):
+                print('DDB Processed all. New since_date', since_date)
-                    result.update({
+                # TODO: parameterize sleep
-                        'subject': message.subject,
+                # Sleep for 10 min
-                    })
+                sleep(10 * 60)
-                    print('Parsed result: ', result)
-                    print('Indexed result: ', self.index_message(result))
 
 
 if __name__ == '__main__':

crawler/requirements.txt

@@ -1,2 +1,3 @@
+python-dateutil
 imbox
 requests