func_calculate_monthly_net_inc.py 22.1 KB
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515
# -*- coding: utf-8 -*-
"""
Created on Mon Sep 25 14:34:45 2017

calculate monthly net increase values for daily-url data
1 if a video is released after 00:00 of the first day of each month, the
accumulated values is used as monthly_net_inc values directly.
2 if a video is released before 00:00 of the first day of each month, the
current accumulated values will be substracted by the accumulated values at
the day before the first day of each month.
2.1 if a video is released before the first day of a month, but it's historical
data is not aviable

migrate to be a function on Oct 03 2017

Nov 18 2017
Add multi-threads.
Auto distribute data by release_time into threads.

Modify release_time range to be from 60 days before the passed-in date.
The reason for doing this is, data from earlier time have tiny chance to found
historical data for net increase values calculation. This lead to more search time,
which slows down the whole calculation process.

Jul 16 2018
Refactoring.
Breaking change:
    Function cal_monthly_net_inc's argument search_body changed to
be a search_body_logic_part, which is part of a complete search
body, not a json body that can be used directly in search.

@author: hanye
"""

import json
import copy
import datetime
import logging
import threading
from elasticsearch import Elasticsearch
from elasticsearch.helpers import scan
import _thread
from func_calculate_toutiao_video_id import calculate_toutiao_video_id
from func_calculate_newTudou_video_id import calculate_newTudou_video_id


hosts = '192.168.17.11'
port = 9200
es_read = Elasticsearch(hosts=hosts, port=port)
es_write = Elasticsearch(hosts=hosts, port=port)

index_sv = 'short-video-production'
doc_type_daily_url = 'daily-url'

def find_first_day_of_the_month(fetch_year, fetch_month, fetch_day):
    if fetch_day > 1:
        first_day_of_the_month_T = datetime.datetime(year=fetch_year,
                                                     month=fetch_month,
                                                     day=1)
    elif fetch_day == 1:
        if fetch_month > 1:
            first_day_of_the_month_T = datetime.datetime(year=fetch_year,
                                                         month=fetch_month-1,
                                                         day=1)
        elif fetch_month == 1:
            first_day_of_the_month_T = datetime.datetime(year=fetch_year-1,
                                                         month=12,
                                                         day=1)
    return first_day_of_the_month_T


def form_bulk_body(line_data, doc_id):
    # 3.1 update line_data dict
    # added update timestamp when calculate monthly_net_inc values on Oct 16 2017
    line_data.update({'timestamp': int(datetime.datetime.now().timestamp()*1e3)})
    # 3.2 form json body for each line
    action_json_line_str = '{"index": {"_id":"' + doc_id + '"}}'
    line_data.pop('_id', None)
    data_json_line_str = json.dumps(line_data, ensure_ascii=False)
    line_bulk_body = (action_json_line_str + '\n'
                      + data_json_line_str + '\n')
    return line_bulk_body


def bulk_write(thread_id, logger,
               data_dict_Lst, doc_type_target,
               index=index_sv):
    bulk_body = ''
    for data_dict in data_dict_Lst:
        doc_id = data_dict['_id']
        line_bulk_body = form_bulk_body(data_dict, doc_id)
        bulk_body += line_bulk_body
    t_s = datetime.datetime.now()
    es_write.bulk(body=bulk_body, index=index,
                  doc_type=doc_type_target,
                  request_timeout=100)
    logger.info('[ %d ] update %s with %d docs.'
                % (thread_id, doc_type_target, len(data_dict_Lst)))
    t_e = datetime.datetime.now()
    t_delta = t_e-t_s
    logger.info('[ %d ] bulk write for one data batch time spent %s'
                % (thread_id, t_delta))
    # 3.5 clear data_raw_collector after bulk write
    data_dict_Lst.clear()


def form_search_body_hist_uncomp(platform, fetch_time_upper_boundary_ts,
                                 fetch_time_lower_boundary_ts,
                                 url):
    if platform in ['toutiao', 'new_tudou']:
        videoID = cal_vid_from_url(platform, url)
        search_body_hist_uncomp = {
            "query": {
                "bool": {
                    "must": [
                        {"range": {"fetch_time": {
                            "gte": fetch_time_lower_boundary_ts,
                            "lt": fetch_time_upper_boundary_ts,}}},
                        {"match_phrase": {"url": videoID}},
                        {"exists": {"field": "play_count"}}
                    ],

                }
            },
            "sort": [
                {"fetch_time": {"order": "asc"}}
            ]
        }
    else:
        search_body_hist_uncomp = {
            "query": {
                "bool": {
                    "filter": [
                        {"range": {"fetch_time": {
                            "gte": fetch_time_lower_boundary_ts,
                            "lt": fetch_time_upper_boundary_ts,}}},
                        {"term": {"url.keyword": url}}
                    ],
                    "must": [
                        {"exists": {"field": "play_count"}}
                    ],
                }
            },
            "sort": [
                {"fetch_time": {"order": "asc"}}
            ]
        }
    return search_body_hist_uncomp


def cal_vid_from_url(platform, url):
    if platform == 'toutiao':
        videoID = calculate_toutiao_video_id(url)
    elif platform == 'new_tudou':
        videoID = calculate_newTudou_video_id(url)
    else:
        videoID = url
    return videoID


def get_key_field_values(logger, thread_id, line_data_dict):
    try:
        platform = line_data_dict['platform']
        line_fetch_time_ts = line_data_dict['fetch_time']
        url = line_data_dict['url']
        play_count_acc = line_data_dict['play_count']
        return (platform, line_fetch_time_ts, url, play_count_acc)
    except KeyError:
        logger.info('[%s] lack of field "platform" or "fetch_time" or "url", '
                    'function find_historical returns.'
                    % thread_id)
        return None

def update_MNI_value_for_new_released(logger, thread_id,
                                      line_data_dict):
    key_fields = get_key_field_values(logger, thread_id, line_data_dict)
    if key_fields is None:
        return None
    else:
        play_count_acc = key_fields[3]
        monthly_net_inc_play_count = play_count_acc
        monthly_cal_base = 'accumulated_values'
        if 'comment_count' in line_data_dict:
            comment_count_acc = line_data_dict['comment_count']
            monthly_net_inc_comment_count = comment_count_acc
        else:
            monthly_net_inc_comment_count = 0
        if 'favorite_count' in line_data_dict:
            favorite_count_acc = line_data_dict['favorite_count']
            monthly_net_inc_favorite_count = favorite_count_acc
        else:
            monthly_net_inc_favorite_count = 0
        line_data_dict.update({'monthly_cal_base': monthly_cal_base,
                               'monthly_net_inc_play_count': monthly_net_inc_play_count,
                               'monthly_net_inc_comment_count': monthly_net_inc_comment_count,
                               'monthly_net_inc_favorite_count': monthly_net_inc_favorite_count})
        return line_data_dict


def find_historical(logger, stdout_lock, thread_id,
                    first_day_of_the_month_T, line_data_dict,
                    f_err_log):
    key_fields = get_key_field_values(logger, thread_id, line_data_dict)
    if key_fields is None:
        return None
    else:
        platform = key_fields[0]
        line_fetch_time_ts = key_fields[1]
        url = key_fields[2]
        play_count_acc = key_fields[3]

    zero_oclock_on_first_day_T = datetime.datetime(first_day_of_the_month_T.year,
                                                   first_day_of_the_month_T.month,
                                                   first_day_of_the_month_T.day)
    twenty_four_oclock_on_first_day_T = (zero_oclock_on_first_day_T
                                         + datetime.timedelta(seconds=24*3600))
    twenty_four_oclock_on_first_day_ts = int(twenty_four_oclock_on_first_day_T.timestamp()*1e3)
    fetch_time_lower_boundary_ts = int(zero_oclock_on_first_day_T.timestamp()*1e3)
    fetch_time_upper_boundary_ts = line_fetch_time_ts

    search_body_hist_uncomp = form_search_body_hist_uncomp(
        platform, fetch_time_upper_boundary_ts, fetch_time_lower_boundary_ts, url)

    search_resp = es_read.search(index=index_sv,
                                 doc_type=doc_type_daily_url,
                                 body=search_body_hist_uncomp,
                                 size=10,
                                 request_timeout=100)
    hist_data_raw_collector = search_resp['hits']['hits']
    search_resp_total = search_resp['hits']['total']
    if search_resp_total >= 1:
        earliest_hist_data = hist_data_raw_collector[0]['_source']
        hist_data_id = hist_data_raw_collector[0]['_id']

        earliest_hist_data_fetch_time = earliest_hist_data['fetch_time']
        play_count_acc_history = int(earliest_hist_data['play_count'])
        monthly_net_inc_play_count = play_count_acc - play_count_acc_history
        monthly_cal_base_fetch_time = earliest_hist_data_fetch_time

        if earliest_hist_data_fetch_time <= twenty_four_oclock_on_first_day_ts:
            monthly_cal_base = 'historical_complete'
        else:
            monthly_cal_base = 'historical_uncomplete'

        if 'comment_count' in earliest_hist_data and 'comment_count' in line_data_dict:
            comment_count_acc = line_data_dict['comment_count']
            comment_count_acc_history = int(earliest_hist_data['comment_count'])
            monthly_net_inc_comment_count = comment_count_acc - comment_count_acc_history
        else:
            monthly_net_inc_comment_count = 0

        if 'favorite_count' in earliest_hist_data and 'favorite_count' in line_data_dict:
            favorite_count_acc = line_data_dict['favorite_count']
            favorite_count_acc_history = int(earliest_hist_data['favorite_count'])
            monthly_net_inc_favorite_count = favorite_count_acc - favorite_count_acc_history
        else:
            monthly_net_inc_favorite_count = 0

        line_data_dict.update({'monthly_cal_base': monthly_cal_base,
                               'monthly_net_inc_play_count': monthly_net_inc_play_count,
                               'monthly_cal_base_fetch_time': monthly_cal_base_fetch_time,
                               'MNI_hist_data_id': hist_data_id,
                               'monthly_net_inc_comment_count': monthly_net_inc_comment_count,
                               'monthly_net_inc_favorite_count': monthly_net_inc_favorite_count})
    # if there are no hits, leave the monthly_net_inc fields empty, and
    # tag monthly_cal_base as historical_data_absent
    else:
        with stdout_lock:
            print('[', thread_id, ']', url, 'found no historical data in present survey date range',
                  datetime.datetime.fromtimestamp(fetch_time_lower_boundary_ts/1e3),
                  'to', datetime.datetime.fromtimestamp(fetch_time_upper_boundary_ts/1e3),
                  file=f_err_log)
        monthly_cal_base = 'historical_data_absent'
        line_data_dict.update({'monthly_cal_base': monthly_cal_base,})
    return line_data_dict


# 1.2 define sub-function to run in thread
def cal_monthly_net_inc_and_write_es_in_thread(
        thread_id, logger, es_read,
        index, doc_type_target,
        search_body_in_thread,
        first_day_of_the_month_T,
        stdout_lock,
        f_err_log):
    logger.info('[ %d ] search_body_in_thread %s'
                % (thread_id, str(search_body_in_thread)))
    find_hits_total_in_threads = es_read.search(index=index,
                                                doc_type=doc_type_target,
                                                body=search_body_in_thread,
                                                size=0,
                                                request_timeout=100)
    hits_total_in_threads = find_hits_total_in_threads['hits']['total']
    logger.info('[ %d ] hits_total_in_threads %d'
                % (thread_id, hits_total_in_threads))
    scan_response = scan(client=es_read,
                         query=search_body_in_thread,
                         index=index,
                         doc_type=doc_type_target,
                         scroll='5m',
                         size=1000,
                         request_timeout=100)

    data_raw_collector = []
    line_counter = 0
    for line in scan_response:
        line_counter += 1
        raw_line = line
        line_doc_id = raw_line['_id']
        line_data = raw_line['_source']
        release_time = int(line_data['release_time'])
        release_time_T = datetime.datetime.fromtimestamp(int(release_time/1e3))
        if release_time_T >= first_day_of_the_month_T:
            line_data = update_MNI_value_for_new_released(logger, thread_id,
                                                          line_data)
        else:
            line_data = find_historical(logger, stdout_lock, thread_id,
                                        first_day_of_the_month_T,
                                        line_data,
                                        f_err_log)
        line_data['_id'] = line_doc_id
        data_raw_collector.append(line_data)
        # update monthly_net_inc values every 1000 docs
        if line_counter%1000 == 0 or line_counter == hits_total_in_threads:
            logger.info('[ %d ] processing %d / %d'
                        % (thread_id, line_counter, hits_total_in_threads))
            bulk_write(thread_id, logger, data_raw_collector,
                       doc_type_target, index=index_sv)
            data_raw_collector.clear()
    if data_raw_collector != []:
        logger.info('[ %d ] processing %d / %d'
                    % (thread_id, line_counter, hits_total_in_threads))
        bulk_write(thread_id, logger, data_raw_collector,
                   doc_type_target, index=index_sv)
        data_raw_collector.clear()

    logger.info('[ %d ] Thread exits. %s' % (thread_id, datetime.datetime.now()))


def divid_by_release_time(logger, threads_num, average_data_num,
                          data_distr_by_release_time_Lst):
    # 1.1 By total hits and data distribution by release_time
    # find proper release_time range for multi-threads
    release_time_range_Lst = []
    # find the end side of each range segment
    data_counter_collector = 0
    distr_idx = 0
    for distr_by_releaseT in data_distr_by_release_time_Lst:
        data_num_each_day = distr_by_releaseT['doc_count']
        data_counter_collector += data_num_each_day
        if (data_counter_collector > average_data_num*0.9
                or distr_idx == len(data_distr_by_release_time_Lst)-1):
            release_time_range_dict = {'start':None, 'end':None,
                                       'end_idx': distr_idx,
                                       'data_num': data_counter_collector}
            release_time_range_Lst.append(release_time_range_dict)
            data_counter_collector = 0
        distr_idx += 1

    # fillup the start timestamp and the end timestamp
    start_side_cache = data_distr_by_release_time_Lst[0]['key']
    for range_seg in release_time_range_Lst:
        if range_seg['end_idx']+1 > len(data_distr_by_release_time_Lst)-1:
            range_seg['end'] = int(data_distr_by_release_time_Lst[range_seg['end_idx']]['key']
                                   + 24 * 3600 * 1e3)
        else:
            range_seg['end'] = data_distr_by_release_time_Lst[range_seg['end_idx']+1]['key']
        range_seg['start'] = start_side_cache
        start_side_cache = range_seg['end']

    # in case the splitted range segments are longer or shorter than threads_num
    if len(release_time_range_Lst) != threads_num:
        threads_num = len(release_time_range_Lst)
    logger.info('Actual threads_num %d' % threads_num)
    logger.info('release_time_range_Lst:\n %s' % str(release_time_range_Lst))
    return (threads_num, release_time_range_Lst)


def cal_monthly_net_inc(fetch_year, fetch_month, fetch_day,
                        fetch_hour=0,
                        fetch_time_seg_hours=24,
                        doc_type_target='daily-url',
                        search_body_logic_part=None,
                        threads_num=5,
                        logger_name='calculate_monthly_net_inc'):
    date_passed_in = datetime.datetime(year=fetch_year, month=fetch_month,
                                       day=fetch_day, hour=fetch_hour)
    first_day_of_the_month_T = find_first_day_of_the_month(fetch_year, fetch_month, fetch_day)

    logger = logging.getLogger('%s.func' % logger_name)
    log_path = ('/home/hanye/project_data/Python/Projects/proj-short-videos/'
                'write-data-into-es/log/')
    logger.info('************ log starts')
    logger.info('calculate monthly net_inc values')
    err_log_filename = ('calculate_monthly_net_inc_for_'
                        + str(fetch_year) + '_'
                        + str(fetch_month) + '_'
                        + str(fetch_day) + '_error_log'+ '_'
                        + str(datetime.datetime.now())[0:10])
    f_err_log = open(log_path+err_log_filename, 'a', encoding='gb18030')

    t_s_cal = datetime.datetime.now()

    # 1 get all the data from ccr es for given fetch_time
    # from _type=daily-url
    fetch_time_start_ts = int(date_passed_in.timestamp()*1e3)
    fetch_time_end_ts = int((date_passed_in
                             + datetime.timedelta(seconds=fetch_time_seg_hours*3600)
                            ).timestamp()*1e3)

    fetch_time_start_iso = datetime.datetime.fromtimestamp(
        int(fetch_time_start_ts/1e3)).isoformat()
    fetch_time_end_iso = datetime.datetime.fromtimestamp(
        int(fetch_time_end_ts/1e3)).isoformat()
    logger.info(('fetch_time_start: %s, fetch_time_end: %s'
                 % (fetch_time_start_iso, fetch_time_end_iso)))
    # Modify release_time range to be from 90 days before the passed-in date
    # Nov 18 2017
    release_time_start_ts = int((date_passed_in - datetime.timedelta(days=60))
                                .timestamp()*1000)
    release_time_end_ts = int((date_passed_in + datetime.timedelta(days=365))
                              .timestamp()*1000)
    search_body = {
        "query": {
            "bool": {
                "filter": [
                    {"range": {"release_time": {
                        "gte": release_time_start_ts,
                        "lt": release_time_end_ts}}
                    },
                    {"range": {"fetch_time": {
                        "gte": fetch_time_start_ts,
                        "lt": fetch_time_end_ts}}
                    }
                ],
            }
        },
        "size": 2,
        "aggs": {
            "release_time_distribution": {
                "date_histogram": {
                    "field": "release_time",
                    "interval": "day",
                    "time_zone": "Asia/Shanghai"
                }
            }
        }
    }

    if search_body_logic_part is not None:
        if 'filter' in search_body_logic_part:
            search_body['query']['bool']['filter'].append(search_body_logic_part['filter'])
        else:
            search_body['query']['bool'].update(search_body_logic_part)
    else:
        pass
    search_response = es_read.search(index=index_sv,
                                     doc_type=doc_type_target,
                                     body=search_body,
                                     request_timeout=100)
    hits_total = search_response['hits']['total']
    logger.info('find total hits %d in _type %s.'
                % (hits_total, doc_type_target))
    if hits_total == 0:
        return
    else:
        pass

    # 1.1 By total hits and data distribution by release_time
    # find proper release_time range for multi-threads
    data_distr_by_release_time_Lst = search_response['aggregations']['release_time_distribution']['buckets']
    average_data_num = hits_total // threads_num
    sharding_by_release_time = divid_by_release_time(logger, threads_num,
                                                     average_data_num,
                                                     data_distr_by_release_time_Lst)
    threads_num = sharding_by_release_time[0]
    release_time_range_Lst = sharding_by_release_time[1]

    # global lock to prevent conflit on stdout when print
    # used for error log
    stdout_lock = _thread.allocate_lock()

    search_body_in_thread_Lst = []
    waitfor = []
    for i in range(0, threads_num):
        release_time_ts_start = release_time_range_Lst[i]['start']
        release_time_ts_end = release_time_range_Lst[i]['end']
        search_body_in_thread = copy.deepcopy(search_body)
        search_body_in_thread.pop('size', None)
        search_body_in_thread.pop('aggs', None)
        search_body_in_thread['query']['bool']['filter'][0]['range']['release_time']['gte'] = release_time_ts_start
        search_body_in_thread['query']['bool']['filter'][0]['range']['release_time']['lt'] = release_time_ts_end
        search_body_in_thread['query']['bool']['filter'][1]['range']['fetch_time']['gte'] = fetch_time_start_ts
        search_body_in_thread['query']['bool']['filter'][1]['range']['fetch_time']['lt'] = fetch_time_end_ts

        search_body_in_thread_Lst.append(search_body_in_thread)

        thread = threading.Thread(target=cal_monthly_net_inc_and_write_es_in_thread,
                                  args=(i, logger, es_read, index_sv,
                                        doc_type_target, search_body_in_thread_Lst[i],
                                        first_day_of_the_month_T,
                                        stdout_lock, f_err_log))
        waitfor.append(thread)
        thread.start()

    # check if every thread exists
    for thread in waitfor:
        thread.join()

    t_e_cal = datetime.datetime.now()
    t_delta_cal = t_e_cal-t_s_cal
    logger.info('for all data of date %s total time cost: %s'
                % (date_passed_in.isoformat()[:10], t_delta_cal))
    logger.info('Main thread exiting.')
    f_err_log.close()