personal-data/personal_data/fetchers/psnprofiles.py

import dataclasses
import datetime
import logging
import re
from collections.abc import Iterator

import bs4
import requests_util

import personal_data.html_util
from personal_data import secrets
from personal_data.data import DeduplicateMode, Scraper

from .. import parse_util

logger = logging.getLogger(__name__)

URL_API_ROOT = 'https://psnprofiles.com/'
URL_PROFILE = URL_API_ROOT + '{psn_id}'
URL_USER_GAME_TROPHIES = URL_API_ROOT + 'trophies/{game_id}/{psn_id}'
URL_GAMES_OVERVIEW = URL_API_ROOT + '{psn_id}'


PSN_PROFILES_DEFAULT_TIMEZONE = datetime.UTC


def game_psnprofiles_id_from_url(relative_url: str) -> int:
    m = re.match(r'/(?:trophy|trophies)/(\d+)\-(?:[\w-]+)(/[\w-]*)?', relative_url)
    result = m.group(1)
    return int(result)


MAX_NUMBER_GAMES_TO_PARSE = 10000


@dataclasses.dataclass(frozen=True)
class PsnProfiles(Scraper):
    """Downloads all trophies for the given user.

    Individual game pages are cached for a period between 1 to 30 days,
    depending upon how recently you played them.
    """

    dataset_name = 'games_played'
    deduplicate_mode = DeduplicateMode.BY_ALL_COLUMNS

    @staticmethod
    def requires_cfscrape() -> bool:
        return True

    def scrape(self):
        self._setup_cache()
        games_rows = list(self._scrape_games_overview())
        games_ids = {row['psnprofiles.game_id']: row['game.name'] for row in games_rows}

        logger.info('Found %d games from overview', len(games_rows))

        for idx, (game_id, game_name) in enumerate(reversed(games_ids.items())):
            cache_duration = datetime.timedelta(days=min(idx + 1, 30))
            yield from self._scrape_game_trophies(game_id, game_name, cache_duration)
            del game_id
            if idx >= MAX_NUMBER_GAMES_TO_PARSE:
                break
        logger.info('Found all trophies for playstation games')

    def _setup_cache(self):
        requests_util.setup_limiter(
            self.session,
            URL_API_ROOT,
            per_minute=5,
        )

    def _scrape_games_overview(self) -> Iterator[dict]:
        for page_num in range(1, 1000):
            logger.info('Getting Overview (page %d)', page_num)
            url = URL_GAMES_OVERVIEW.format(psn_id=secrets.PLAYSTATION_PSN_ID)
            response = self.session.get(url, params={'page': page_num})
            if 'page' not in response.url:
                msg = "Configuration error? psnprofiles.com made an redirection. This is possibly because your profile name wasn't exactly as expected. Please check it"
                raise RuntimeError(msg)
            response.raise_for_status()
            soup = bs4.BeautifulSoup(response.text, 'lxml')
            soup = personal_data.html_util.normalize_soup_slightly(soup, classes=False)
            games_on_page = list(self._iterate_games_from_games_table(soup))
            yield from games_on_page
            if len(games_on_page) == 0:
                return

    def _iterate_games_from_games_table(self, soup) -> Iterator[dict]:
        # Games table
        table_rows = soup.find(id='gamesTable').find_all('tr')
        assert len(table_rows) > 0, url

        if title := table_rows[0].h2:
            if title.get_text().strip() == 'No games found':
                return

        for row in table_rows:
            cells = row.find_all('td')

            # Check for pagination
            if re.match(
                r'show \d+ more games',
                cells[0].get_text().strip(),
                re.IGNORECASE,
            ):
                break

            game_name = cells[1].find(class_='title').get_text()
            psnprofiles_id = game_psnprofiles_id_from_url(cells[0].find('a')['href'])
            game_icon = cells[0].find('img')['src']

            game_name = row.select_one('.title').get_text()
            game_platform = row.select_one('.platform').get_text()

            small_infos = cells[1].find_all('div')
            if len(small_infos) > 2:
                time_played_div = small_infos[2]
                time_played_div.sup.extract()
                time_played = parse_util.parse_date(
                    time_played_div.get_text(),
                )
            else:
                time_played = None

            d = {
                # Important fields
                'game.name': game_name,
                # Secondary fields
                'game.platform': game_platform,
                'game.icon': game_icon,
                'psnprofiles.game_id': psnprofiles_id,
            }
            if time_played:
                d['me.last_played_time'] = time_played
            yield d

    def _parse_game_release_date(self, soup: bs4.BeautifulSoup) -> datetime.date:
        table_rows = soup.select('table.gameInfo tr')
        for row in table_rows:
            cells = row.select('td')
            if cells[0].get_text() in {'Release', 'Releases'}:
                text = cells[1].get_text()
                dates = re.findall(r'\w+\s+\d+,\s+\d{4}', text)
                return min(parse_util.parse_date(date) for date in dates)
        assert False, 'Could not find release date'

    def _scrape_game_trophies(
        self,
        psnprofiles_id: int,
        game_name: str,
        cache_duration: datetime.timedelta,
    ) -> Iterator[dict]:
        assert isinstance(psnprofiles_id, int), psnprofiles_id
        assert isinstance(game_name, str), game_name

        logger.info('Getting Game Trophies %s', psnprofiles_id)

        url = URL_USER_GAME_TROPHIES.format(
            psn_id=secrets.PLAYSTATION_PSN_ID,
            game_id=psnprofiles_id,
        )
        response = self.session.get(url, expire_after=cache_duration)
        response.raise_for_status()

        # Parse data
        soup = bs4.BeautifulSoup(response.content, 'lxml')

        # Normalize before parsing trophies
        soup = personal_data.html_util.normalize_soup_slightly(soup, classes=False)

        # Parse release year
        game_release_date = self._parse_game_release_date(soup)
        assert game_release_date

        # Remove redundant
        for redundant in soup.select('.wide-ad'):
            redundant.extract()
        for redundant in soup.select('div.col-xs-4'):
            redundant.extract()

        # Recent trophies.
        soup_tropies = soup.select(
            '#content.page > .row > div.col-xs div.box table.zebra tr.completed',
        )
        for row in soup_tropies:
            cells = row.find_all('td')

            trophy_name_a = cells[1].a
            if trophy_name_a is None:
                continue
            trophy_name = trophy_name_a.get_text().strip()
            trophy_name_a.extract()
            trophy_desc = cells[1].get_text().strip()

            trophy_icon = cells[0].img['src']

            if 'Missing\nTimestamp' in cells[2].get_text().strip():
                continue
            cells[2].span.span.nobr.sup.extract()
            gotten_at = parse_util.parse_time(
                cells[2].get_text(),
                timezone=PSN_PROFILES_DEFAULT_TIMEZONE,
            )

            yield {
                'game.name': game_name,
                'game.release_date': game_release_date,
                'me.last_played_time': gotten_at,
                # Trophy Data
                'trophy.name': trophy_name,
                'trophy.desc': trophy_desc,
                'trophy.icon': trophy_icon,
                # Ids
                'psnprofiles.game_id': psnprofiles_id,
            }

            del row, cells