calibre/recipes/nytimesbook.recipe

#!/usr/bin/env python
# vim:fileencoding=utf-8
# License: GPLv3 Copyright: 2015, Kovid Goyal <kovid at kovidgoyal.net>

import json

import mechanize

from calibre.web.feeds.news import BasicNewsRecipe
from polyglot.urllib import urlencode

use_wayback_machine = False
# This is an Apollo persisted query hash which you can get
# from looking at the XHR requests made by: https://www.nytimes.com/section/todayspaper
# or by https://www.nytimes.com/section/world
persistedQuery = '1f99120a11e94dd62a9474f68ee1255537ee3cf7eac20a0377819edb2fa1fef7'


def absolutize(url):
    if url.startswith('/'):
        url = 'https://www.nytimes.com' + url
    return url


class NewYorkTimesBookReview(BasicNewsRecipe):
    title = u'New York Times Book Review'
    language = 'en_US'
    description = 'The New York Times Sunday Book Review'
    __author__ = 'Kovid Goyal'

    no_stylesheets = True
    no_javascript = True
    ignore_duplicate_articles = {'title', 'url'}
    encoding = 'utf-8'

    extra_css = '''
        .byl, .time { font-size:small; color:#202020; }
        .cap { font-size:small; text-align:center; }
        .cred { font-style:italic; font-size:small; }
        em, blockquote { color: #202020; }
        .sc { font-variant: small-caps; }
        .lbl { font-size:small; color:#404040; }
        img { display:block; margin:0 auto; }
    '''

    articles_are_obfuscated = use_wayback_machine

    if use_wayback_machine:
        def get_obfuscated_article(self, url):
            from calibre.ptempfile import PersistentTemporaryFile
            with PersistentTemporaryFile() as tf:
                tf.write(self.get_nyt_page(url))
            return tf.name

    @property
    def nyt_parser(self):
        ans = getattr(self, '_nyt_parser', None)
        if ans is None:
            from calibre.live import load_module
            self._nyt_parser = ans = load_module('calibre.web.site_parsers.nytimes')
        return ans

    def get_nyt_page(self, url, skip_wayback=False):
        if use_wayback_machine and not skip_wayback:
            from calibre import browser
            return self.nyt_parser.download_url(url, browser())
        return self.index_to_soup(url, raw=True)

    def preprocess_raw_html(self, raw_html, url):
        return self.nyt_parser.extract_html(self.index_to_soup(raw_html), url)

    recipe_specific_options = {
        'res': {
            'short': (
                'For hi-res images, select a resolution from the following\noptions: '
                'popup, jumbo, mobileMasterAt3x, superJumbo'
            ),
            'long': (
                'This is useful for non e-ink devices, and for a lower file size\nthan '
                'the default, use mediumThreeByTwo440, mediumThreeByTwo225, articleInline.'
            ),
        },
        'comp': {
            'short': 'Compress News Images?',
            'long': 'enter yes',
            'default': 'no'
        }
    }

    def __init__(self, *args, **kwargs):
        BasicNewsRecipe.__init__(self, *args, **kwargs)
        c = self.recipe_specific_options.get('comp')
        if c and isinstance(c, str):
            if c.lower() == 'yes':
                self.compress_news_images = True

    def read_nyt_metadata(self):
        soup = self.index_to_soup('https://www.nytimes.com/pages/books/review/index.html')
        script = soup.findAll('script', text=lambda x: x and 'window.__preloadedData' in x)[0]
        script = type(u'')(script)
        json_data = script[script.find('{'):script.rfind(';')].strip().rstrip(';')  # }}
        self.nytimes_graphql_config = json.loads(self.nyt_parser.clean_js_json(json_data))['config']
        return soup

    def nyt_graphql_query(self, qid, operationName='CollectionsQuery'):
        query = {
            'operationName': operationName,
            'variables': json.dumps({
                'id': qid,
                'first': 10,
                'exclusionMode': 'HIGHLIGHTS_AND_EMBEDDED',
                'isFetchMore':False,
                'isTranslatable':False,
                'isEspanol':False,
                'highlightsListUri':'nyt://per/personalized-list/__null__',
                'highlightsListFirst':0,
                'hasHighlightsList':False
            }, separators=',:'),
            'extensions': json.dumps({
                'persistedQuery': {
                    'version':1,
                    'sha256Hash': persistedQuery,
                },
            }, separators=',:')
        }
        url = self.nytimes_graphql_config['gqlUrlClient'] + '?' + urlencode(query)
        br = self.browser
        # br.set_debug_http(True)
        headers = dict(self.nytimes_graphql_config['gqlRequestHeaders'])
        headers['Accept'] = 'application/json'
        req = mechanize.Request(url, headers=headers)
        raw = br.open(req).read()
        # open('/t/raw.json', 'wb').write(raw)
        return json.loads(raw)

    def parse_index(self):
        # return [('Articles', [{'url': 'https://www.nytimes.com/2022/09/08/books/review/karen-armstrong-by-the-book-interview.html', 'title':'test'}])]
        self.read_nyt_metadata()
        query_id = '/section/books/review'
        data = self.nyt_graphql_query(query_id)
        return parse_toc(data, self.log)

    def get_browser(self, *args, **kwargs):
        kwargs['user_agent'] = 'User-Agent: Mozilla/5.0 (compatible; archive.org_bot; Wayback Machine Live Record; +http://archive.org/details/archive.org_bot)'
        br = BasicNewsRecipe.get_browser(self, *args, **kwargs)
        return br

    def preprocess_html(self, soup):
        w = self.recipe_specific_options.get('res')
        if w and isinstance(w, str):
            res = '-' + w
            for img in soup.findAll('img', attrs={'src':True}):
                if '-article' in img['src']:
                    ext = img['src'].split('?')[0].split('.')[-1]
                    img['src'] = img['src'].rsplit('-article', 1)[0] + res + '.' + ext
        for c in soup.findAll('div', attrs={'class':'cap'}):
            for p in c.findAll(['p', 'div']):
                p.name = 'span'
        return soup


def asset_to_article(asset):
    title = asset['headline']['default']
    return {'title': title, 'url': asset['url'], 'description': asset['summary']}


def parse_toc(data, log=print):
    containers = data['data']['legacyCollection']['groupings'][0]['containers']
    articles = []
    log('Book reviews')
    for cont in containers:
        if cont['__typename'] != 'LegacyCollectionContainer':
            continue
        for rel in cont['relations']:
            if rel.get('__typename') == 'LegacyCollectionRelation':
                asset = rel['asset']
                if asset['__typename'] == 'Article':
                    articles.append(asset_to_article(asset))
                    log(' ', articles[-1]['title'] + ':', articles[-1]['url'])
    feeds = [('Book reviews', articles)]
    articles = []
    log('Books of the Times')
    try:
        containers = data['data']['legacyCollection']['collectionsPage']
        if containers.get('embeddedCollections'):
            containers = containers['embeddedCollections']
        else:
            containers = [containers]
    except Exception as e:
        log('Failed to parse web section', 'Books of the Times', 'with error:', e)
        return articles
    for cont in containers:
        for s in cont['stream']['edges']:
            asset = s['node']
            if asset['__typename'] == 'Article':
                articles.append(asset_to_article(asset))
                log(' ', articles[-1]['title'] + ':', articles[-1]['url'])
    if articles:
        feeds.append(('Book of the Times', articles))
    return feeds


if __name__ == '__main__':
    import sys
    data = json.loads(open(sys.argv[-1], 'rb').read())
    parse_toc(data)