query-gmail.py

# https://developers.google.com/gmail/api/quickstart/python
# pip3 install --upgrade google-api-python-client

from __future__ import print_function
from apiclient.discovery import build
from httplib2 import Http
from oauth2client import file, client, tools
from base64 import urlsafe_b64decode as b64d

# update this to a filter which pulls all the target emails
QUERY_FILTER = 'from:[email protected]'

def build_service():
  SCOPES = 'https://www.googleapis.com/auth/gmail.readonly'
  store = file.Storage('credentials.json')
  creds = store.get()
  if not creds or creds.invalid:
    flow = client.flow_from_clientsecrets('client_secret.json', SCOPES)
    creds = tools.run_flow(flow, store)
  service = build('gmail', 'v1', http=creds.authorize(Http()))

  return service

def get_message_ids(service):
  results = service.users().messages().list(userId='me', q=QUERY_FILTER).execute()
  token = results.get('nextPageToken', None)
  messages = results.get('messages', [])
  
  print('[*] %d messages found' % len(messages))
  
  return [m['id'] for m in messages]

def get_messages(service, msg_id_list):
  print('[*] Fetching all messages')

  messages = []
  for mid in msg_id_list:
    results = service.users().messages().get(userId='me', id=mid).execute()
    payload = results.get('payload', None)
    if payload is None:
      print('[!] Unable to read message with ID: %d' % mid)
      continue
    
    messages.append(payload)

  return messages

def parse_messages(messages):
  for msg in messages:
    # this will likely need to be changed on a per-use basis -- original target had a really weird email structure
    content = msg['parts'][0]['parts'][0]['parts'][1]['body']['data'].strip()
    content = b64d(content).decode('utf-8')
    
    # do something with the decoded body

if __name__ == '__main__':
  service = build_service()
  msg_id_list = get_message_ids(service)
  messages = get_messages(service, msg_id_list)
  parse_messages(messages)