2024-08-07 10:25:48 +00:00
|
|
|
|
import asyncio
|
2024-01-25 19:41:27 +00:00
|
|
|
|
import json
|
|
|
|
|
import os
|
2022-11-24 17:19:43 +00:00
|
|
|
|
import time
|
2023-12-17 20:30:20 +00:00
|
|
|
|
from datetime import datetime, timedelta, timezone
|
2024-01-25 19:41:27 +00:00
|
|
|
|
from typing import Dict
|
2024-01-23 13:04:38 +00:00
|
|
|
|
|
2024-08-07 10:37:08 +00:00
|
|
|
|
# ga
|
2024-01-28 13:26:40 +00:00
|
|
|
|
from google.analytics.data_v1beta import BetaAnalyticsDataClient
|
2024-08-09 06:37:06 +00:00
|
|
|
|
from google.analytics.data_v1beta.types import (
|
|
|
|
|
DateRange,
|
|
|
|
|
Dimension,
|
|
|
|
|
Metric,
|
|
|
|
|
RunReportRequest,
|
|
|
|
|
)
|
2022-11-21 22:23:16 +00:00
|
|
|
|
|
2024-01-23 13:04:38 +00:00
|
|
|
|
from orm.author import Author
|
|
|
|
|
from orm.shout import Shout, ShoutAuthor, ShoutTopic
|
|
|
|
|
from orm.topic import Topic
|
|
|
|
|
from services.db import local_session
|
2024-08-07 10:37:50 +00:00
|
|
|
|
from utils.logger import root_logger as logger
|
2024-01-23 13:04:38 +00:00
|
|
|
|
|
2024-08-07 10:37:08 +00:00
|
|
|
|
GOOGLE_KEYFILE_PATH = os.environ.get("GOOGLE_KEYFILE_PATH", "/dump/google-service.json")
|
|
|
|
|
GOOGLE_PROPERTY_ID = os.environ.get("GOOGLE_PROPERTY_ID", "")
|
2024-04-17 15:32:23 +00:00
|
|
|
|
VIEWS_FILEPATH = "/dump/views.json"
|
2024-01-28 12:54:38 +00:00
|
|
|
|
|
2024-01-13 12:44:56 +00:00
|
|
|
|
|
2022-11-18 17:54:37 +00:00
|
|
|
|
class ViewedStorage:
|
2024-08-07 10:37:08 +00:00
|
|
|
|
lock = asyncio.Lock()
|
2024-10-31 16:06:58 +00:00
|
|
|
|
views_by_shout_slug = {}
|
|
|
|
|
views_by_shout_id = {}
|
2024-01-22 18:20:17 +00:00
|
|
|
|
shouts_by_topic = {}
|
|
|
|
|
shouts_by_author = {}
|
2024-08-07 10:37:08 +00:00
|
|
|
|
views = None
|
2024-01-23 13:04:38 +00:00
|
|
|
|
period = 60 * 60 # каждый час
|
2024-01-28 13:26:40 +00:00
|
|
|
|
analytics_client: BetaAnalyticsDataClient | None = None
|
2024-08-07 10:37:08 +00:00
|
|
|
|
auth_result = None
|
2022-11-22 07:29:54 +00:00
|
|
|
|
disabled = False
|
2024-04-26 22:51:45 +00:00
|
|
|
|
start_date = datetime.now().strftime("%Y-%m-%d")
|
2022-11-18 17:54:37 +00:00
|
|
|
|
|
2022-11-20 07:48:40 +00:00
|
|
|
|
@staticmethod
|
2024-08-07 10:25:48 +00:00
|
|
|
|
async def init():
|
2024-08-07 10:37:08 +00:00
|
|
|
|
"""Подключение к клиенту Google Analytics с использованием аутентификации"""
|
2022-11-22 07:29:54 +00:00
|
|
|
|
self = ViewedStorage
|
2024-08-07 10:37:08 +00:00
|
|
|
|
async with self.lock:
|
|
|
|
|
# Загрузка предварительно подсчитанных просмотров из файла JSON
|
|
|
|
|
self.load_precounted_views()
|
|
|
|
|
|
|
|
|
|
os.environ.setdefault("GOOGLE_APPLICATION_CREDENTIALS", GOOGLE_KEYFILE_PATH)
|
|
|
|
|
if GOOGLE_KEYFILE_PATH and os.path.isfile(GOOGLE_KEYFILE_PATH):
|
|
|
|
|
# Using a default constructor instructs the client to use the credentials
|
|
|
|
|
# specified in GOOGLE_APPLICATION_CREDENTIALS environment variable.
|
|
|
|
|
self.analytics_client = BetaAnalyticsDataClient()
|
2024-10-23 08:29:44 +00:00
|
|
|
|
logger.info(" * Google Analytics credentials accepted")
|
2024-08-07 10:37:08 +00:00
|
|
|
|
|
|
|
|
|
# Запуск фоновой задачи
|
|
|
|
|
_task = asyncio.create_task(self.worker())
|
|
|
|
|
else:
|
2024-10-15 08:12:09 +00:00
|
|
|
|
logger.warning(" * please, add Google Analytics credentials file")
|
2024-08-07 10:37:08 +00:00
|
|
|
|
self.disabled = True
|
2022-11-20 07:48:40 +00:00
|
|
|
|
|
2024-01-22 16:17:39 +00:00
|
|
|
|
@staticmethod
|
|
|
|
|
def load_precounted_views():
|
2024-08-07 10:37:08 +00:00
|
|
|
|
"""Загрузка предварительно подсчитанных просмотров из файла JSON"""
|
2024-01-22 16:17:39 +00:00
|
|
|
|
self = ViewedStorage
|
2024-10-15 08:12:09 +00:00
|
|
|
|
viewfile_path = VIEWS_FILEPATH
|
|
|
|
|
if not os.path.exists(viewfile_path):
|
|
|
|
|
viewfile_path = os.path.join(os.path.curdir, "views.json")
|
|
|
|
|
if not os.path.exists(viewfile_path):
|
|
|
|
|
logger.warning(" * views.json not found")
|
|
|
|
|
return
|
2024-10-15 16:52:12 +00:00
|
|
|
|
|
2024-10-15 08:12:09 +00:00
|
|
|
|
logger.info(f" * loading views from {viewfile_path}")
|
2024-01-22 16:17:39 +00:00
|
|
|
|
try:
|
2024-10-15 08:12:09 +00:00
|
|
|
|
start_date_int = os.path.getmtime(viewfile_path)
|
|
|
|
|
start_date_str = datetime.fromtimestamp(start_date_int).strftime("%Y-%m-%d")
|
|
|
|
|
self.start_date = start_date_str
|
|
|
|
|
now_date = datetime.now().strftime("%Y-%m-%d")
|
|
|
|
|
|
|
|
|
|
if now_date == self.start_date:
|
|
|
|
|
logger.info(" * views data is up to date!")
|
2024-03-12 12:57:46 +00:00
|
|
|
|
else:
|
2024-10-15 08:12:09 +00:00
|
|
|
|
logger.warn(f" * {viewfile_path} is too old: {self.start_date}")
|
|
|
|
|
|
|
|
|
|
with open(viewfile_path, "r") as file:
|
|
|
|
|
precounted_views = json.load(file)
|
2024-10-31 16:06:58 +00:00
|
|
|
|
self.views_by_shout_slug.update(precounted_views)
|
2024-10-15 08:12:09 +00:00
|
|
|
|
logger.info(f" * {len(precounted_views)} shouts with views was loaded.")
|
|
|
|
|
|
2024-10-31 16:06:58 +00:00
|
|
|
|
# get shout_id by slug
|
|
|
|
|
with local_session() as session:
|
|
|
|
|
for slug, views_count in self.views_by_shout_slug.items():
|
|
|
|
|
shout_id = session.query(Shout.id).filter(Shout.slug == slug).scalar()
|
|
|
|
|
if isinstance(shout_id, int):
|
|
|
|
|
self.views_by_shout_id.update({shout_id: views_count})
|
|
|
|
|
|
2024-01-22 16:17:39 +00:00
|
|
|
|
except Exception as e:
|
2024-10-15 08:12:09 +00:00
|
|
|
|
logger.error(f"precounted views loading error: {e}")
|
2024-01-22 16:17:39 +00:00
|
|
|
|
|
2024-08-07 10:37:08 +00:00
|
|
|
|
# noinspection PyTypeChecker
|
2022-11-21 22:23:16 +00:00
|
|
|
|
@staticmethod
|
2024-08-07 10:25:48 +00:00
|
|
|
|
async def update_pages():
|
2024-08-07 10:37:08 +00:00
|
|
|
|
"""Запрос всех страниц от Google Analytics, отсортированных по количеству просмотров"""
|
2022-11-21 22:23:16 +00:00
|
|
|
|
self = ViewedStorage
|
2024-10-15 08:12:09 +00:00
|
|
|
|
logger.info(" ⎧ views update from Google Analytics ---")
|
2024-01-28 09:03:41 +00:00
|
|
|
|
if not self.disabled:
|
2024-01-23 13:04:38 +00:00
|
|
|
|
try:
|
|
|
|
|
start = time.time()
|
2024-08-07 10:37:08 +00:00
|
|
|
|
async with self.lock:
|
|
|
|
|
if self.analytics_client:
|
|
|
|
|
request = RunReportRequest(
|
|
|
|
|
property=f"properties/{GOOGLE_PROPERTY_ID}",
|
|
|
|
|
dimensions=[Dimension(name="pagePath")],
|
|
|
|
|
metrics=[Metric(name="screenPageViews")],
|
|
|
|
|
date_ranges=[DateRange(start_date=self.start_date, end_date="today")],
|
|
|
|
|
)
|
|
|
|
|
response = self.analytics_client.run_report(request)
|
|
|
|
|
if response and isinstance(response.rows, list):
|
|
|
|
|
slugs = set()
|
|
|
|
|
for row in response.rows:
|
|
|
|
|
print(
|
|
|
|
|
row.dimension_values[0].value,
|
|
|
|
|
row.metric_values[0].value,
|
|
|
|
|
)
|
|
|
|
|
# Извлечение путей страниц из ответа Google Analytics
|
|
|
|
|
if isinstance(row.dimension_values, list):
|
|
|
|
|
page_path = row.dimension_values[0].value
|
|
|
|
|
slug = page_path.split("discours.io/")[-1]
|
|
|
|
|
views_count = int(row.metric_values[0].value)
|
|
|
|
|
|
|
|
|
|
# Обновление данных в хранилище
|
|
|
|
|
self.views_by_shout[slug] = self.views_by_shout.get(slug, 0)
|
|
|
|
|
self.views_by_shout[slug] += views_count
|
|
|
|
|
self.update_topics(slug)
|
|
|
|
|
|
|
|
|
|
# Запись путей страниц для логирования
|
|
|
|
|
slugs.add(slug)
|
|
|
|
|
|
2024-10-15 08:12:09 +00:00
|
|
|
|
logger.info(f" ⎪ collected pages: {len(slugs)} ")
|
2024-08-07 10:37:08 +00:00
|
|
|
|
|
|
|
|
|
end = time.time()
|
2024-10-15 08:12:09 +00:00
|
|
|
|
logger.info(" ⎪ views update time: %fs " % (end - start))
|
2024-01-28 09:03:41 +00:00
|
|
|
|
except Exception as error:
|
|
|
|
|
logger.error(error)
|
2024-06-04 09:55:12 +00:00
|
|
|
|
self.disabled = True
|
2022-11-18 17:54:37 +00:00
|
|
|
|
|
2022-11-19 11:35:34 +00:00
|
|
|
|
@staticmethod
|
2024-10-31 16:06:58 +00:00
|
|
|
|
def get_shout(shout_slug="", shout_id=0) -> int:
|
|
|
|
|
"""Получение метрики просмотров shout по slug или id."""
|
2022-11-19 11:35:34 +00:00
|
|
|
|
self = ViewedStorage
|
2024-10-31 16:06:58 +00:00
|
|
|
|
return self.views_by_shout_slug.get(shout_slug, self.views_by_shout_id.get(shout_id, 0))
|
2023-11-03 10:10:22 +00:00
|
|
|
|
|
|
|
|
|
@staticmethod
|
2024-08-07 10:15:58 +00:00
|
|
|
|
def get_shout_media(shout_slug) -> Dict[str, int]:
|
2024-08-07 10:30:41 +00:00
|
|
|
|
"""Получение метрики воспроизведения shout по slug."""
|
2023-11-03 10:10:22 +00:00
|
|
|
|
self = ViewedStorage
|
2024-08-07 10:30:41 +00:00
|
|
|
|
return self.views_by_shout.get(shout_slug, 0)
|
2022-11-19 11:35:34 +00:00
|
|
|
|
|
2022-11-21 05:18:50 +00:00
|
|
|
|
@staticmethod
|
2024-08-07 10:15:58 +00:00
|
|
|
|
def get_topic(topic_slug) -> int:
|
2024-08-07 10:30:41 +00:00
|
|
|
|
"""Получение суммарного значения просмотров темы."""
|
2022-11-21 05:18:50 +00:00
|
|
|
|
self = ViewedStorage
|
2024-08-07 10:30:41 +00:00
|
|
|
|
return sum(self.views_by_shout.get(shout_slug, 0) for shout_slug in self.shouts_by_topic.get(topic_slug, []))
|
2022-11-21 05:18:50 +00:00
|
|
|
|
|
2024-01-22 15:42:45 +00:00
|
|
|
|
@staticmethod
|
2024-08-07 10:15:58 +00:00
|
|
|
|
def get_author(author_slug) -> int:
|
2024-08-07 10:30:41 +00:00
|
|
|
|
"""Получение суммарного значения просмотров автора."""
|
2024-01-22 15:42:45 +00:00
|
|
|
|
self = ViewedStorage
|
2024-08-07 10:30:41 +00:00
|
|
|
|
return sum(self.views_by_shout.get(shout_slug, 0) for shout_slug in self.shouts_by_author.get(author_slug, []))
|
2024-01-22 15:42:45 +00:00
|
|
|
|
|
2022-11-22 13:58:55 +00:00
|
|
|
|
@staticmethod
|
2023-11-22 18:23:15 +00:00
|
|
|
|
def update_topics(shout_slug):
|
2024-08-07 10:37:08 +00:00
|
|
|
|
"""Обновление счетчиков темы по slug shout"""
|
2022-11-22 13:58:55 +00:00
|
|
|
|
self = ViewedStorage
|
2023-11-22 18:23:15 +00:00
|
|
|
|
with local_session() as session:
|
2024-01-23 13:04:38 +00:00
|
|
|
|
# Определение вспомогательной функции для избежания повторения кода
|
2024-01-22 18:20:17 +00:00
|
|
|
|
def update_groups(dictionary, key, value):
|
|
|
|
|
dictionary[key] = list(set(dictionary.get(key, []) + [value]))
|
|
|
|
|
|
2024-01-23 13:04:38 +00:00
|
|
|
|
# Обновление тем и авторов с использованием вспомогательной функции
|
2024-01-25 19:41:27 +00:00
|
|
|
|
for [_shout_topic, topic] in (
|
2024-05-30 04:12:00 +00:00
|
|
|
|
session.query(ShoutTopic, Topic).join(Topic).join(Shout).where(Shout.slug == shout_slug).all()
|
2024-01-25 19:41:27 +00:00
|
|
|
|
):
|
2024-01-22 18:20:17 +00:00
|
|
|
|
update_groups(self.shouts_by_topic, topic.slug, shout_slug)
|
|
|
|
|
|
2024-01-25 19:41:27 +00:00
|
|
|
|
for [_shout_topic, author] in (
|
2024-05-30 04:12:00 +00:00
|
|
|
|
session.query(ShoutAuthor, Author).join(Author).join(Shout).where(Shout.slug == shout_slug).all()
|
2024-01-25 19:41:27 +00:00
|
|
|
|
):
|
2024-01-22 18:20:17 +00:00
|
|
|
|
update_groups(self.shouts_by_author, author.slug, shout_slug)
|
2024-01-22 15:42:45 +00:00
|
|
|
|
|
2023-11-03 10:10:22 +00:00
|
|
|
|
@staticmethod
|
2024-08-07 10:25:48 +00:00
|
|
|
|
async def worker():
|
2024-08-07 10:37:08 +00:00
|
|
|
|
"""Асинхронная задача обновления"""
|
2022-11-21 22:23:16 +00:00
|
|
|
|
failed = 0
|
2022-11-22 07:29:54 +00:00
|
|
|
|
self = ViewedStorage
|
|
|
|
|
if self.disabled:
|
|
|
|
|
return
|
2023-10-05 22:45:32 +00:00
|
|
|
|
|
2024-11-02 01:28:16 +00:00
|
|
|
|
while True:
|
|
|
|
|
try:
|
|
|
|
|
await self.update_pages()
|
|
|
|
|
failed = 0
|
|
|
|
|
except Exception as exc:
|
|
|
|
|
failed += 1
|
|
|
|
|
logger.debug(exc)
|
|
|
|
|
logger.info(" - update failed #%d, wait 10 secs" % failed)
|
|
|
|
|
if failed > 3:
|
|
|
|
|
logger.info(" - views update failed, not trying anymore")
|
|
|
|
|
break
|
|
|
|
|
if failed == 0:
|
|
|
|
|
when = datetime.now(timezone.utc) + timedelta(seconds=self.period)
|
|
|
|
|
t = format(when.astimezone().isoformat())
|
|
|
|
|
logger.info(" ⎩ next update: %s" % (t.split("T")[0] + " " + t.split("T")[1].split(".")[0]))
|
|
|
|
|
await asyncio.sleep(self.period)
|
|
|
|
|
else:
|
|
|
|
|
await asyncio.sleep(10)
|
|
|
|
|
logger.info(" - try to update views again")
|