# -*- coding: utf-8 -*-

'''

Created on Apr 29, 2013

@author: ymh

'''

from iri_tweet.models import (User, EntityType, adapt_json, MediaType, Media, 

    EntityMedia, Hashtag, EntityHashtag, EntityUser, EntityUrl, Url, Entity, Tweet, 

    TweetSource, TweetLog)

from iri_tweet.utils import (ObjectsBuffer, adapt_fields, fields_adapter, 

    ObjectBufferProxy, get_oauth_token, clean_keys)

from sqlalchemy.orm import joinedload

import anyjson

import logging

import twitter

import twitter_text

class TwitterProcessorException(Exception):

    pass

class TwitterProcessor(object):

    def __init__(self, json_dict, json_txt, source_id, session, consumer_token, access_token=None, token_filename=None, user_query_twitter=False, logger=None):

        if json_dict is None and json_txt is None:

            raise TwitterProcessorException("No json")

        if json_dict is None:

            self.json_dict = anyjson.deserialize(json_txt)

        else:

            self.json_dict = json_dict

        if not json_txt:

            self.json_txt = anyjson.serialize(json_dict)

        else:

            self.json_txt = json_txt

        if "id" not in self.json_dict:

            raise TwitterProcessorException("No id in json")

        self.source_id = source_id

        self.session = session

        self.consumer_key = consumer_token[0]

        self.consumer_secret = consumer_token[1]

        self.token_filename = token_filename

        self.access_token = access_token

        self.obj_buffer = ObjectsBuffer()

        self.user_query_twitter = user_query_twitter

        if not logger:

            self.logger = logging.getLogger(__name__)

        else:

            self.logger = logger

    def process(self):

        if self.source_id is None:

            tweet_source = self.obj_buffer.add_object(TweetSource, None, {'original_json':self.json_txt}, True)

            self.source_id = tweet_source.id

        self.process_source()

        self.obj_buffer.persists(self.session)

    def process_source(self):

        raise NotImplementedError()

    def log_info(self):

        return "Process tweet %s" %  repr(self.__class__)    

class TwitterProcessorStatus(TwitterProcessor):

    def __init__(self, json_dict, json_txt, source_id, session, consumer_token, access_token=None, token_filename=None, user_query_twitter=False, logger=None):

        TwitterProcessor.__init__(self, json_dict, json_txt, source_id, session, consumer_token, access_token=access_token, token_filename=token_filename, user_query_twitter=user_query_twitter, logger=logger)

    def __get_user(self, user_dict, do_merge):

        self.logger.debug("Get user : " + repr(user_dict)) #@UndefinedVariable

        user_dict = adapt_fields(user_dict, fields_adapter["stream"]["user"])

        user_id = user_dict.get("id",None)    

        user_name = user_dict.get("screen_name", user_dict.get("name", None))

        if user_id is None and user_name is None:

            return None

        user = None

        if user_id:

            user = self.obj_buffer.get(User, id=user_id)

        else:

            user = self.obj_buffer.get(User, screen_name=user_name)

        #to do update user id needed            

        if user is not None:

            user_created_at = None

            if user.args is not None:

                user_created_at = user.args.get('created_at', None)

            if user_created_at is None and user_dict.get('created_at', None) is not None and do_merge:

                if user.args is None:

                    user.args = user_dict

                else:

                    user.args.update(user_dict)

            return user

        #todo : add methpds to objectbuffer to get buffer user

        user_obj = None

        if user_id:

            user_obj = self.session.query(User).filter(User.id == user_id).first()

        else:

            user_obj = self.session.query(User).filter(User.screen_name.ilike(user_name)).first()

        #todo update user if needed

        if user_obj is not None:            

            if user_obj.created_at is not None or user_dict.get('created_at', None) is None or not do_merge :

                user = ObjectBufferProxy(User, None, None, False, user_obj)

            else:

                user = self.obj_buffer.add_object(User, None, user_dict, True, user_obj)

            return user

        user_created_at = user_dict.get("created_at", None)

        if user_created_at is None and self.user_query_twitter:

            if self.access_token is not None:

                acess_token_key, access_token_secret = self.access_token

            else:

                acess_token_key, access_token_secret = get_oauth_token(consumer_key=self.consumer_key, consumer_secret=self.consumer_secret, token_file_path=self.token_filename)

            #TODO pass it as argument    

            t = twitter.Twitter(auth=twitter.OAuth(acess_token_key, access_token_secret, self.consumer_key, self.consumer_secret))

            try:

                if user_id:

                    user_dict = t.users.show(user_id=user_id)

                else:

                    user_dict = t.users.show(screen_name=user_name)

            except Exception as e:

                self.logger.info("get_user : TWITTER ERROR : " + repr(e)) #@UndefinedVariable

                self.logger.info("get_user : TWITTER ERROR : " + str(e)) #@UndefinedVariable

                return None

        if "id" not in user_dict:

            return None

        #TODO filter get, wrap in proxy

        user_obj = self.session.query(User).filter(User.id == user_dict["id"]).first()

        if user_obj is not None and not do_merge:

            return ObjectBufferProxy(User, None, None, False, user_obj)

        else:        

            return self.obj_buffer.add_object(User, None, user_dict, True)        

    def __get_or_create_object(self, klass, filter_by_kwargs, filter_arg, creation_kwargs, must_flush, do_merge):

        obj_proxy = self.obj_buffer.get(klass, **filter_by_kwargs)

        if obj_proxy is None:

            query = self.session.query(klass)

            if filter_arg is not None:

                query = query.filter(filter_arg)

            else:

                query = query.filter_by(**filter_by_kwargs)

            obj_instance = query.first()

            if obj_instance is not None:

                if not do_merge:

                    obj_proxy = ObjectBufferProxy(klass, None, None, False, obj_instance)

                else:

                    obj_proxy = self.obj_buffer.add_object(klass, None, creation_kwargs, must_flush, obj_instance)

        if obj_proxy is None:

            obj_proxy = self.obj_buffer.add_object(klass, None, creation_kwargs, must_flush)

        return obj_proxy

    def __process_entity(self, ind, ind_type):

        self.logger.debug("Process_entity : " + repr(ind) + " : " + repr(ind_type)) #@UndefinedVariable

        ind = clean_keys(ind)

        entity_type = self.__get_or_create_object(EntityType, {'label':ind_type}, None, {'label':ind_type}, True, False)

        entity_dict = {

           "indice_start"   : ind["indices"][0],

           "indice_end"     : ind["indices"][1],

           "tweet_id"       : self.tweet.id,

           "entity_type_id" : entity_type.id,

           "source"         : adapt_json(ind)

        }

        def process_medias():

            media_id = ind.get('id', None)

            if media_id is None:

                return None, None

            type_str = ind.get("type", "photo")

            media_type = self.__get_or_create_object(MediaType, {'label': type_str}, None, {'label':type_str}, True, False)

            media_ind = adapt_fields(ind, fields_adapter["entities"]["medias"])

            if "type" in media_ind:

                del(media_ind["type"])

            media_ind['type_id'] = media_type.id            

            media = self.__get_or_create_object(Media, {'id':media_id}, None, media_ind, True, False)

            entity_dict['media_id'] = media.id

            return EntityMedia, entity_dict

        def process_hashtags():

            text = ind.get("text", ind.get("hashtag", None))

            if text is None:

                return None, None

            ind['text'] = text

            hashtag = self.__get_or_create_object(Hashtag, {'text':text}, Hashtag.text.ilike(text), ind, True, False)

            entity_dict['hashtag_id'] = hashtag.id

            return EntityHashtag, entity_dict             

        def process_user_mentions():

            user_mention = self.__get_user(ind, False)

            if user_mention is None:

                entity_dict['user_id'] = None

            else:

                entity_dict['user_id'] = user_mention.id

            return EntityUser, entity_dict

        def process_urls():

            url = self.__get_or_create_object(Url, {'url':ind["url"]}, None, ind, True, False)

            entity_dict['url_id'] = url.id

            return EntityUrl, entity_dict

        #{'': lambda }

        entity_klass, entity_dict =  { 

            'hashtags': process_hashtags,

            'user_mentions' : process_user_mentions,

            'urls' : process_urls,

            'media': process_medias,

            }.get(ind_type, lambda: (Entity, entity_dict))()

        self.logger.debug("Process_entity entity_dict: " + repr(entity_dict)) #@UndefinedVariable

        if entity_klass:

            self.obj_buffer.add_object(entity_klass, None, entity_dict, False)

    def __process_twitter(self):

        tweet_nb = self.session.query(Tweet).filter(Tweet.id == self.json_dict["id"]).count()

        if tweet_nb > 0:

            return

        ts_copy = adapt_fields(self.json_dict, fields_adapter["stream"]["tweet"])

        # get or create user

        user = self.__get_user(self.json_dict["user"], True)

        if user is None:

            self.logger.warning("USER not found " + repr(self.json_dict["user"])) #@UndefinedVariable

            ts_copy["user_id"] = None

        else:

            ts_copy["user_id"] = user.id

        del(ts_copy['user'])

        ts_copy["tweet_source_id"] = self.source_id

        self.tweet = self.obj_buffer.add_object(Tweet, None, ts_copy, True)

        self.__process_entities()

    def __process_entities(self):

        if "entities" in self.json_dict:

            for ind_type, entity_list in self.json_dict["entities"].iteritems():

                for ind in entity_list:

                    self.__process_entity(ind, ind_type)

        else:

            text = self.tweet.text

            extractor = twitter_text.Extractor(text)

            for ind in extractor.extract_hashtags_with_indices():

                self.__process_entity(ind, "hashtags")

            for ind in extractor.extract_urls_with_indices():

                self.__process_entity(ind, "urls")

            for ind in extractor.extract_mentioned_screen_names_with_indices():

                self.__process_entity(ind, "user_mentions")

    def process_source(self):

        status_id = self.json_dict["id"]

        log = self.session.query(TweetLog).filter(TweetLog.status_id==status_id).first()

        if(log):

            self.obj_buffer.add_object(TweetLog, log, {'status': TweetLog.TWEET_STATUS['DELETE'], 'status_id': None})

            self.session.query(TweetSource).filter(TweetSource.id==self.source_id).delete()

        else:

            self.__process_twitter()

        self.obj_buffer.add_object(TweetLog, None, {'tweet_source_id':self.source_id, 'status':TweetLog.TWEET_STATUS['OK']}, True)

    def log_info(self):

        screen_name = self.json_dict.get("user",{}).get("screen_name","")

        return u"Process Tweet from %s : %s" % (screen_name, self.json_dict.get('text',u""))

class TwitterProcessorDelete(TwitterProcessor):

    """

    {

      "delete":{

        "status":{

            "id":1234,

            "id_str":"1234",

            "user_id":3,

            "user_id_str":"3"

        }

      }

    }

    """

    def __init__(self, json_dict, json_txt, source_id, session, consumer_token, access_token=None, token_filename=None, user_query_twitter=False, logger=None):

        TwitterProcessor.__init__(self, json_dict, json_txt, source_id, session, consumer_token, access_token=access_token, token_filename=token_filename, user_query_twitter=user_query_twitter, logger=logger)

    def process(self):

        #find tweet

        tweet_id = self.json_dict.get('delete',{}).get('status',{}).get('id',None)

        if tweet_id:

            t = self.session.query(Tweet).options(joinedload(Tweet.tweet_source)).filter(Tweet.id == tweet_id).first()

            if t:

                tsource = t.tweet_source                

                self.session.delete(t)

                self.session.query(TweetLog).filter(TweetLog.tweet_source_id == tsource.id).delete()

                self.session.delete(tsource)

                self.obj_buffer.add_object(TweetLog, None, {'tweet_source_id':self.source_id, 'status':TweetLog.TWEET_STATUS['DELETE']}, True)

            else:

                self.obj_buffer.add_object(TweetLog, None, {'tweet_source_id':self.source_id, 'status_id': tweet_id,'status':TweetLog.TWEET_STATUS['DELETE_PENDING']}, True)

    def log_info(self):

        status_del = self.json_dict.get('delete', {}).get("status",{})

        return u"Process delete for %s : %s" % (status_del.get('user_id_str',u""), status_del.get('id_str',u""))

class TwitterProcessorScrubGeo(TwitterProcessor):

    """

    {

        "scrub_geo":{

        "user_id":14090452,

        "user_id_str":"14090452",

        "up_to_status_id":23260136625,

        "up_to_status_id_str":"23260136625"

      }

    }

    """

    def __init__(self, json_dict, json_txt, source_id, session, consumer_token, access_token=None, token_filename=None, user_query_twitter=False, logger=None):

        TwitterProcessor.__init__(self, json_dict, json_txt, source_id, session, consumer_token, access_token=access_token, token_filename=token_filename, user_query_twitter=user_query_twitter, logger=logger)

    def process_source(self):        

        up_to_status_id = self.json_dict.get("scrub_geo", {}).get("up_to_status_id", None)

        if not up_to_status_id:

            return

        tweets = self.session.query(Tweet).options(joinedload(Tweet.tweet_source)).filter(Tweet.id <= up_to_status_id)

        for t in tweets:

            self.obj_buffer.add_object(Tweet, t, {'geo': None})

            tsource = t.tweet_source

            tsource_dict = anyjson.serialize(tsource.original_json)

            if tsource_dict.get("geo", None):

                tsource_dict["geo"] = None

                self.obj_buffer.add_object(TweetSource, tsource, {'original_json': anyjson.serialize(tsource_dict)})

        self.obj_buffer.add_object(TweetLog, None, {'tweet_source_id':self.source_id, 'status':TweetLog.TWEET_STATUS['SCRUB_GEO']}, True)

    def log_info(self):

        return u"Process scrub geo for %s : %s" % (self.json_dict["scrub_geo"].get('user_id_str',u""), self.json_dict["scrub_geo"].get('id_str',u""))