Backup your Read It Later bookmarks with CouchDB

backup_bookmarks.py

#!/usr/bin/env python

import simplejson as json
import urllib
import httplib

import restkit
from restkit.resource import Resource

# ReadItLater credentials
RIL_KEY = "****"
RIL_USER = "****"
RIL_PASS = "****"

# CouchDB informations
COUCH_KEY = "****" # User
COUCH_PASS = "****" # Password
COUCH_DB = "****" # Database
COUCH_HOST = "****" # Server


def get_couch():
    cfg = {
        'user': COUCH_KEY,
        'pass': COUCH_PASS,
        'host': COUCH_HOST,
    }
    return Resource("https://%(user)s:%(pass)s@%(host)s" % cfg)

def get_since():
    r = get_couch()
    try:
        res = r.get("/%s/infos" % COUCH_DB)
    except restkit.errors.ResourceNotFound:
        return None
    return int(json.loads(res.body_string())['since'])

def get_infos_rev():
    r = get_couch()
    try:
        res = r.get("/%s/infos" % COUCH_DB)
    except restkit.errors.ResourceNotFound:
        return None
    return json.loads(res.body_string())['_rev']

def create_since(since):
    r = get_couch()
    payload = json.dumps({'since': since})
    headers = { "Content-type": "application/x-www-form-urlencoded" }
    r.put("/%s/infos" % COUCH_DB, payload, headers)

def update_since(since):
    r = get_couch()
    rev = get_infos_rev()
    payload = json.dumps({'since': since, '_rev': rev})
    headers = { "Content-type": "application/json" }
    r.put("/%s/infos" % COUCH_DB, payload, headers)


def main():
    # First grab the state from RIL
    conn = httplib.HTTPSConnection("readitlaterlist.com")

    payload = {
        'apikey': RIL_KEY,
        'username': RIL_USER,
        'password': RIL_PASS,
        'since': 0,
    }

    since = get_since()

    if since == None:
        del payload['since']
    else:
        payload['since'] = since

    params = urllib.urlencode(payload)
    
    conn.request("GET", "/v2/get?"+params)

    resp = json.loads(conn.getresponse().read())

    new_since = int(resp['since'])

    # Update the since?
    if not since == None:
        update_since(new_since)
    # Or create it
    else:
        create_since(new_since)

    if resp['list'] == []:
        return None

    for key, data in resp['list'].iteritems():
        del data['item_id']
        data['time_updated'] = int(data['time_updated'])
        data['time_added'] = int(data['time_added'])
        data['tags'] = list(data.get('tags', '').split(','))
        data['read'] = bool(data['state'])
        del data['state']

        r = get_couch()

        r.put("/%s/%s" % (COUCH_DB, key), json.dumps(data))



if __name__ == '__main__':
    main()