Example #1
0
    def run(self, dao: Dao, h1):
        h1('twitter_friends_sum取得')
        tbl_tfs = dao.table('twitter_friends_sum')
        tfs_list = tbl_tfs.full_scan()

        d = DateTimeUtil.now() - timedelta(days=15)
        two_week_ago = d.strftime('%Y/%m/%d_00:00:00')

        h1('集計開始')
        for tfs in tfs_list:
            for cd in tfs['rank']:
                ds = tfs['rank'][cd]['ds']
                n_ds = []
                for r in ds:
                    if r['d'] >= two_week_ago:
                        n_ds.append(r)
                if n_ds:
                    tfs['rank'][cd]['ds'] = n_ds
                else:
                    print("tfs['rank'][cd] = 空. cd={}".format(cd))
                    del tfs['rank'][cd]
            if tfs['rank']:
                tbl_tfs.put_item(Item=tfs)
            else:
                print('空。 uid={}'.format(tfs['uid']))
                tbl_tfs.delete_item_silent({"S": tfs['uid']})
        h1('終了')
Example #2
0
    def run(self, dao, req):

        Log.info('twitter_friendsを全件取得。')
        friends = self.dao.table('twitter_friends').find({})

        Log.info('インフルエンサー毎に過去15日のツイートを集計する。')
        for friend in friends:
            cur_repo = self.dao.table('stock_report_pre').find({'tweets.user_id': friend['id_str']})
            month_ago = DateTimeUtil.now() - timedelta(days=15)
            user_tweets = []
            for r in cur_repo:
                u_tws = [t for t in r['tweets'] if t['user_id'] == friend['id_str'] and t['created_at'] >= month_ago]
                if u_tws:
                    ccode_group_list = [t for t in user_tweets if t['ccode'] == r['ccode']]
                    ccode_group = {'name': '', 'last_update_date': month_ago, 'tweet': []}
                    if ccode_group_list:
                        ccode_group = ccode_group_list[0]
                    else:
                        ccode_group['ccode'] = r['ccode']
                        brand = self.dao.table('stock_brands').find_one({'ccode': r['ccode']})
                        if brand:
                            ccode_group['name'] = brand['name']
                        user_tweets.append(ccode_group)

                    ccode_group['last_update_date'] = max(
                        ccode_group['last_update_date'], max([t['created_at'] for t in u_tws]))
                    ccode_group['tweet'].extend(u_tws)

            user_tweets = sorted(user_tweets, key=lambda x: x['last_update_date'], reverse=True)

            ret = []
            for r in user_tweets:
                r['is_market_time'] = DateTimeUtil.is_market_time(r['last_update_date'])
                str_day = r['last_update_date'].strftime('%Y/%m/%d')
                dat = [d for d in ret if d['str_day'] == str_day]
                if dat:
                    dat[0]['tweet'].append(r)
                else:
                    buf = {'str_day': str_day, 'day': r['last_update_date'], 'tweet': [r]}
                    ret.append(buf)

            friend['tweet_summary'] = ret
            dao.table('twitter_friends').update_one({'id_str': friend['id_str']}, {'$set': friend})
            Log.debug('登録: {}', friend['screen_name'])

        Log.info('終了')
Example #3
0
    def run(self, dao: Dao, h1):
        h1('stock_report取得')
        repos = dao.table('stock_report').full_scan()
        d = DateTimeUtil.now() - timedelta(days=15)
        two_week_ago = d.strftime('%Y/%m/%d_00:00:00')
        friend_tweet = {}
        ret_list = []
        h1('ツイート集計開始')
        for repo in repos:
            cd = repo['ccode']
            tw_ids = repo['tweets']
            tweets = dao.table("tweet").find_batch(tw_ids)
            tweets = [t for t in tweets if t['created_at'] > two_week_ago]
            tweets = sorted(tweets,
                            key=lambda r: r['created_at'],
                            reverse=True)

            for t in tweets:
                u_id = t['user_id']
                d = t['created_at']
                ob = {'nm': repo['name'], 'ds': [{'d': d, 't': t['id_str']}]}
                if u_id in friend_tweet:
                    if cd in friend_tweet[t['user_id']]:
                        friend_tweet[u_id][cd]['ds'].append({
                            'd': d,
                            't': t['id_str']
                        })
                    else:
                        friend_tweet[u_id][cd] = ob
                else:
                    friend_tweet[t['user_id']] = {}
                    friend_tweet[t['user_id']][cd] = ob

        for k in friend_tweet:
            r = {'uid': k, 'rank': friend_tweet[k]}
            ret_list.append(r)

        dao.table('twitter_friends_sum').insert(ret_list)
        print('hoge')
Example #4
0
 def day_key_now(self):
     return DateTimeUtil.strf_ymd_st(DateTimeUtil.now())
Example #5
0
    def run(self, dao: Dao, h1):
        edinet = Edinet()
        start_day = DateTimeUtil.yesterday()
        end_day = DateTimeUtil.now()
        # start_day = datetime(2018, 5, 19)
        # end_day = datetime(2018, 5, 20)

        h1('EDINETの書類検索から指定期間の有価証券報告書の検索結果を取得する。')
        search_rows = edinet.get_report_search_results(start_day, end_day)
        search_rows.reverse()

        # search_rows = [r for r in search_rows if 'スタートトゥデイ' in r['company_name']]

        h1('検索結果から有価証券報告書を取得し、銘柄毎スクレイピング開始。')
        dao_identify = dao.table('stock_identify')
        dao_brand = dao.table('stock_brands')
        not_brands = []
        for search_row in search_rows:
            data_html = edinet.get_report_html(search_row['syorui_kanri_no'])
            shorui_mei = search_row['syorui_mei']
            Log.info('html取得完了 : {} : {}'.format(search_row['company_name'],
                                                 shorui_mei))
            if '大株主' not in data_html:
                Log.warn('HTMLに大株主が無い。別資料かも。 company_name : {}'.format(
                    search_row['company_name']))
                continue

            table_list = edinet.report_html_to_split_table_list(
                data_html, shorui_mei)
            Log.info('HTML -> table分割 : {}'.format(search_row['company_name']))
            dat = {}
            for t in table_list:
                if t['title'] == '発行済株式' and 'outstanding_share' not in dat:
                    dat['outstanding_share'] = edinet.outstanding_share(
                        t, search_row)
                    Log.debug('** 発行済株式 : {}'.format(
                        search_row['company_name']))

                if t['title'] == '所有者別状況':
                    dat['holder_rate'] = edinet.holder_rate_status(
                        t, search_row)
                    Log.debug('** 所有者別状況 : {}'.format(
                        search_row['company_name']))

                if t['title'] == '大株主の状況':
                    dat['holders'] = edinet.major_shareholders(t, search_row)
                    Log.debug('** 大株主の状況 : {}'.format(
                        search_row['company_name']))

            if not dat:
                continue

            name = search_row['company_name'].replace('(株)',
                                                      '').replace(' ', '')
            name = re.sub(r'^ ', '', name)
            name = re.sub(r' $', '', name)

            identify = dao_identify.find_query({'nm': name})
            if not identify:
                name_s = zenhan.z2h(name, mode=1)
                identify = dao_identify.find_query({'nm': name_s})
                if not identify:
                    name_s = name_s.replace(' ', '')
                    identify = dao_identify.find_query({'nm': name_s})

            if identify and len(identify) == 1:
                brand = dao_brand.find_by_key(identify[0]['ccode'])
                Log.info('stock_brands証券コード取得。 ccode, name : {}, {}'.format(
                    brand['ccode'], name))
                dat['ccode'] = brand['ccode']
                dat['name'] = brand['name']
                dat['date'] = DateTimeUtil.strf_ymd_st(
                    DateTimeUtil.date_from_japanese_era(
                        search_row['submition_day'], short=True))

                if empty(dat, 'outstanding_share') or empty(
                        dat, 'holder_rate') or empty(dat, 'holders'):
                    edi = dao.table('stock_edinet').find_by_key(dat['ccode'])
                    if edi:
                        if empty(dat, 'outstanding_share') and not empty(
                                edi, 'outstanding_share'):
                            dat['outstanding_share'] = edi['outstanding_share']
                        if empty(dat, 'holder_rate') and not empty(
                                edi, 'holder_rate'):
                            dat['holder_rate'] = edi['holder_rate']
                        if empty(dat, 'holders') and not empty(edi, 'holders'):
                            dat['holders'] = edi['holders']

                h1('stock_edinetに登録')
                dao.table('stock_edinet').put_item(Item=dat)
            else:
                dao.table('err_value').put_item(
                    Item={
                        'key': 'Job013_edi_name',
                        'd': DateTimeUtil.str_now(),
                        'val': name
                    })

        h1('終了')