shortcode.py

# _*_ coding:utf-8 _*_
import app_config
import datetime
import logging
import requests
import shortcodes

from PIL import Image
from StringIO import StringIO
from bs4 import BeautifulSoup
from functools import partial
from jinja2 import Environment, FileSystemLoader
from pymongo import MongoClient

TWITTER_OEMBED_URL = 'https://api.twitter.com/1.1/statuses/oembed.json'
IMAGE_URL_TEMPLATE = '%s/%s'
IMAGE_TYPES = ['image', 'graphic']
SHORTCODE_DICT = {
    'tweet': {
        'show_media': 1,
        'show_thread': 0
    },
    'youtube': {
        'start_time': 0
    },
    'image': {
        'caption': '',
        'credit': 'Image credit',
        'width': '100%'
    },
    'internal_link': {
        'link_text': 'Link to post'
    },
    'graphic': {
        'caption': 'Graphic caption',
        'credit': 'Graphic credit',
        'width': '100%'
    },
    'facebook': {},
    'npr_video': {},
    'ap_live_video': {},
}

env = Environment(loader=FileSystemLoader('templates/shortcodes'))

logging.basicConfig(format=app_config.LOG_FORMAT)
logger = logging.getLogger(__name__)
logger.setLevel(app_config.LOG_LEVEL)


def _process_id(url, tag):
    """
    Extract an ID from a url (or just return the URL).
    """
    if tag == 'tweet':
        parts = url.split('/')
        return parts[5]
    else:
        return url


def _get_extra_context(id, tag):
    """
    Do some processing
    """
    extra = dict()
    if tag in IMAGE_TYPES:
        extra.update(_get_image_context(id))
    if tag == 'tweet':
        extra.update(_get_tweet_context(id))
    return extra


def _handler(context, content, pargs, kwargs, tag, defaults):
    """
    Default handler all other handlers inherit from.
    """
    if pargs:
        id = _process_id(pargs[0], tag)
        template_context = dict(url=pargs[0],
                            id=id)
        extra_context = _get_extra_context(id, tag)
        template_context.update(extra_context)
    else:
        template_context = dict()
    template_context.update(defaults)
    template_context.update(kwargs)
    template = env.get_template('%s.html' % tag)
    output = template.render(**template_context)
    return output


"""
Register handlers
"""
parser = shortcodes.Parser()
for tag, defaults in SHORTCODE_DICT.items():
    tag_handler = partial(_handler, tag=tag, defaults=defaults)
    parser.register(tag_handler, tag)


def process_shortcode(tag):
    """
    Generates html from shortcode
    """
    # Replace unicode <br>
    text = tag.get_text().replace(u'\xa0', u' ')
    try:
        return parser.parse(text)
    except shortcodes.RenderingError as e:
        logger.error('Could not render short code in: "%s"' % text)
        logger.error('cause: %s' % e.__cause__)
        return ''


def _get_image_context(id):
    """
    Download image and get/cache aspect ratio.
    """
    url = IMAGE_URL_TEMPLATE % (app_config.IMAGE_URL, id)

    client = MongoClient(app_config.MONGODB_URL)
    database = client['liveblog']
    collection = database.images
    result = collection.find_one({'_id': id})

    if not result:
        logger.info('image %s: uncached, downloading %s' % (id, url))
        response = requests.get(url)
        image = Image.open(StringIO(response.content))
        ratio = float(image.height) / float(image.width)
        collection.insert({
            '_id': id,
            'date': datetime.datetime.utcnow(),
            'ratio': ratio,
        })
    else:
        logger.info('image %s: retrieved from cache' % id)
        ratio = result['ratio']

    ratio = round(ratio * 100, 2)
    return dict(ratio=ratio, url=url)


def _get_tweet_context(id):
    """
    Try and figure out a tweet's aspect ratio har dee har.
    """
    layout = 'text'

    client = MongoClient(app_config.MONGODB_URL)
    database = client['liveblog']
    collection = database.tweets
    result = collection.find_one({'_id': id})

    if not result:
        logger.info('tweet %s: uncached, downloading' % id)
        response = requests.get(TWITTER_OEMBED_URL, params=(('id', id),))
        response.raise_for_status()
        data = response.json()
        soup = BeautifulSoup(data['html'])
        links = soup.find_all('a')
        for link in links:
            logger.info(link)
            if link.text == link.attrs['href']:
                layout = 'attached_link'
            if 'pic.twitter.com' in link.text:
                layout = 'image'

        logger.info('tweet %s: is layout `%s`' % (id, layout))

        collection.insert({
            '_id': id,
            'date': datetime.datetime.utcnow(),
            'layout': layout,
        })
    else:
        layout = result['layout']
        logger.info('tweet %s: retrieved from cache, is layout `%s`' % (id, layout))

    return dict(layout=layout)