[naver] improve extraction

[yt-dlp.git] / youtube_dl / extractor / naver.py

# coding: utf-8
from __future__ import unicode_literals

import re

from .common import InfoExtractor
from ..utils import (
    clean_html,
    dict_get,
    ExtractorError,
    get_element_by_class,
    int_or_none,
    try_get,
    update_url_query,
)


class NaverBaseIE(InfoExtractor):
    _CAPTION_EXT_RE = r'\.(?:ttml|vtt)'

    def _extract_video_info(self, video_id, vid, key):
        video_data = self._download_json(
            'http://play.rmcnmv.naver.com/vod/play/v2.0/' + vid,
            video_id, query={
                'key': key,
            })
        meta = video_data['meta']
        title = meta['subject']
        formats = []
        get_list = lambda x: try_get(video_data, lambda y: y[x + 's']['list'], list) or []

        def extract_formats(streams, stream_type, query={}):
            for stream in streams:
                stream_url = stream.get('source')
                if not stream_url:
                    continue
                stream_url = update_url_query(stream_url, query)
                encoding_option = stream.get('encodingOption', {})
                bitrate = stream.get('bitrate', {})
                formats.append({
                    'format_id': '%s_%s' % (stream.get('type') or stream_type, dict_get(encoding_option, ('name', 'id'))),
                    'url': stream_url,
                    'width': int_or_none(encoding_option.get('width')),
                    'height': int_or_none(encoding_option.get('height')),
                    'vbr': int_or_none(bitrate.get('video')),
                    'abr': int_or_none(bitrate.get('audio')),
                    'filesize': int_or_none(stream.get('size')),
                    'protocol': 'm3u8_native' if stream_type == 'HLS' else None,
                })

        extract_formats(get_list('video'), 'H264')
        for stream_set in video_data.get('streams', []):
            query = {}
            for param in stream_set.get('keys', []):
                query[param['name']] = param['value']
            stream_type = stream_set.get('type')
            videos = stream_set.get('videos')
            if videos:
                extract_formats(videos, stream_type, query)
            elif stream_type == 'HLS':
                stream_url = stream_set.get('source')
                if not stream_url:
                    continue
                formats.extend(self._extract_m3u8_formats(
                    update_url_query(stream_url, query), video_id,
                    'mp4', 'm3u8_native', m3u8_id=stream_type, fatal=False))
        self._sort_formats(formats)

        replace_ext = lambda x, y: re.sub(self._CAPTION_EXT_RE, '.' + y, x)

        def get_subs(caption_url):
            if re.search(self._CAPTION_EXT_RE, caption_url):
                return [{
                    'url': replace_ext(caption_url, 'ttml'),
                }, {
                    'url': replace_ext(caption_url, 'vtt'),
                }]
            else:
                return [{'url': caption_url}]

        automatic_captions = {}
        subtitles = {}
        for caption in get_list('caption'):
            caption_url = caption.get('source')
            if not caption_url:
                continue
            sub_dict = automatic_captions if caption.get('type') == 'auto' else subtitles
            sub_dict.setdefault(dict_get(caption, ('locale', 'language')), []).extend(get_subs(caption_url))

        user = meta.get('user', {})

        return {
            'id': video_id,
            'title': title,
            'formats': formats,
            'subtitles': subtitles,
            'automatic_captions': automatic_captions,
            'thumbnail': try_get(meta, lambda x: x['cover']['source']),
            'view_count': int_or_none(meta.get('count')),
            'uploader_id': user.get('id'),
            'uploader': user.get('name'),
            'uploader_url': user.get('url'),
        }


class NaverIE(NaverBaseIE):
    _VALID_URL = r'https?://(?:m\.)?tv(?:cast)?\.naver\.com/(?:v|embed)/(?P<id>\d+)'
    _GEO_BYPASS = False
    _TESTS = [{
        'url': 'http://tv.naver.com/v/81652',
        'info_dict': {
            'id': '81652',
            'ext': 'mp4',
            'title': '[9월 모의고사 해설강의][수학_김상희] 수학 A형 16~20번',
            'description': '메가스터디 수학 김상희 선생님이 9월 모의고사 수학A형 16번에서 20번까지 해설강의를 공개합니다.',
            'upload_date': '20130903',
            'uploader': '메가스터디, 합격불변의 법칙',
            'uploader_id': 'megastudy',
        },
    }, {
        'url': 'http://tv.naver.com/v/395837',
        'md5': '8a38e35354d26a17f73f4e90094febd3',
        'info_dict': {
            'id': '395837',
            'ext': 'mp4',
            'title': '9년이 지나도 아픈 기억, 전효성의 아버지',
            'description': 'md5:eb6aca9d457b922e43860a2a2b1984d3',
            'upload_date': '20150519',
            'uploader': '4가지쇼 시즌2',
            'uploader_id': 'wrappinguser29',
        },
        'skip': 'Georestricted',
    }, {
        'url': 'http://tvcast.naver.com/v/81652',
        'only_matching': True,
    }]

    def _real_extract(self, url):
        video_id = self._match_id(url)
        content = self._download_json(
            'https://tv.naver.com/api/contents/json/v/' + video_id,
            video_id, headers=self.geo_verification_headers())
        player_json = content.get('playerJson') or {}

        vid = player_json.get('videoId')
        in_key = player_json.get('inKey')

        if not vid or not in_key:
            player_auth = player_json.get('playerAuth')
            if player_auth == 'notCountry':
                self.raise_geo_restricted(countries=['KR'])
            elif player_auth == 'notLogin':
                self.raise_login_required()
            raise ExtractorError('couldn\'t extract vid and key')
        info = self._extract_video_info(video_id, vid, in_key)

        clip_info_html = content.get('clipInfoHtml')
        if clip_info_html:
            info['description'] = clean_html(get_element_by_class('desc', clip_info_html))
            upload_date = self._search_regex(
                r'<span[^>]+class="date".*?(\d{4}\.\d{2}\.\d{2})',
                clip_info_html, 'upload date', fatal=False)
            if upload_date:
                info['upload_date'] = upload_date.replace('.', '')

        return info