profile for draftjs exporter

profiling.py

# coding: utf-8

import cgi
import cProfile
import re
import urllib
import urlparse
from pstats import Stats

from draftjs_exporter.dom import DOM
from draftjs_exporter.html import HTML

E = DOM.create_element


def url_fix(s):
    if not s:
        return ''
    if isinstance(s, unicode):
        s = s.encode('utf8', 'ignore')
    scheme, netloc, path, qs, anchor = urlparse.urlsplit(s)
    path = urllib.quote(path, '/%')
    qs = urllib.quote_plus(qs, ':&=')
    return urlparse.urlunsplit((scheme, netloc, path, qs, anchor))


flash_html = ('<object '
              'data="//imgcache.qq.com/tencentvideo_v1/player/TencentPlayer.swf?vid=z03172fgqfr&amp;autoplay=1&amp;tpid=0" '
              'height="300" '
              'type="application/x-shockwave-flash" '
              'width="400">'
              '<param name="movie" value="//imgcache.qq.com/tencentvideo_v1/player/TencentPlayer.swf?vid=z03172fgqfr&amp;autoplay=1&amp;tpid=0"></param>'
              '<param name="wmode" value="transparent"></param>'
              '<param name="allowFullScreen" value="true"></param>'
              '<param name="allowScriptAccess" value="always"></param>'
              '<param name="flashvars" value="autoplay=0&amp;adplay=1"></param>'
              '</object>')


def get_video_info(url):
    return {
        'url': 'http://youtube.com/v/1',
        'flash_html': flash_html,
        'cover': 'b.jpg',
        'title': 'title',
        'support_https': True,
        'site_name': 'youku',
        'site_name_cn': 'youtube',
        'html5_video_url': '',
        'mp4_video_url': ''
    }


def get_subject_info(url):
    return {
        'type': 'book',
        'cover': 'http://a.com/book/1/cover.jpg',
        'title': 'good subject',
        'url': 'http://a.com/good_book/1',
        'summary': 'very good',
        'rating': {'value': '9.5', 'max': '10.0', 'reason': ''}
    }


class URLDecorator(object):
    """
    replace url in plain text to actual html link
    """
    SEARCH_RE = re.compile(r'(http://|https://|www\.)([a-zA-Z0-9\.\-%/\?&_=\+#:~!,\'\*\^$]+)')

    def __repr__(self):
        return '<URLDecorator>'

    def replace(self, match, block_type):
        u_href = match.group(1) + match.group(2)

        if block_type in ('code-block',):
            return u_href

        text = cgi.escape(u_href, quote=True)
        if u_href.startswith("www"):
            u_href = "http://" + u_href

        props = {'href': u_href, 'rel': 'nofollow'}

        return DOM.create_element('a', props, text)


class BrDecorator(object):
    """
    replace \n to <br/>
    """
    SEARCH_RE = re.compile('\n')

    def __repr__(self):
        return '<BrDecorator>'

    def replace(self, match, block_type):
        if block_type in ('code-block',):
            return '\n'

        return DOM.create_element('br')


class LinkEntity(object):

    def render(self, props):
        data = props.get('data', {})
        url = url_fix(data.get('url', ''))
        attrs = {'rel': 'nofollow', 'href': url}
        return E('a', attrs, props['children'])


class VideoEntity(object):

    def render(self, props):
        data = props.get('data', {})
        video_url = data.get('url')
        caption = data.get('caption')
        CAPTION_MAX = 35

        info = get_video_info(video_url)
        caption = caption or info['title'].decode('utf8')
        if info['flash_html'] and info['support_https']:
            flash_elem = DOM.parse_html(info['flash_html'])
            return E('div', {'class': 'video-wrapper'},
                     E('div', {'class': 'video-player'},
                       flash_elem,
                       E('div', {'class': 'video-title'},
                           caption[:CAPTION_MAX]
                         ) if caption else None
                       )
                     )
        else:
            return E('div', {'class': 'video-card'},
                     E('a', {'class': 'video-link', 'href': url_fix(video_url), 'target': '_blank'},
                       (caption or video_url)[:CAPTION_MAX]
                       )
                     )


class SeparatorEntity(object):

    def render(self, props):
        return E('div', {'class': 'separator'}, E('hr'))


class ImageEntity(object):

    def render(self, props):
        CAPTION_MAX = 140
        data = props.get('data', {})
        position = data.get('float', 'center')
        caption = data.get('caption', '')

        children = [E('div', {'class': 'image-wrapper'},
                      E('img', {'src': url_fix(data.get('src', ''))})
                      )]

        if caption:
            children.append(
                E('div', {'class': 'image-caption-wrapper'},
                  E('div', {'class': 'image-caption'},
                    caption[:CAPTION_MAX])
                  )
            )
        return E('div', {'class': 'image-container image-float-%s' % position},
                 *children)


class SubjectEntity(object):

    def render(self, props):
        data = props.get('data', {})
        url = data.get('url')
        fact_url = ''
        TITLE_MAX = 35
        ABSTRACT_MAX = 200

        if url:
            info = get_subject_info(url)
            stars = []
            reason = info['rating']['reason']
            try:
                rating = float(info['rating']['value'])
            except ValueError:
                rating = 0

            for i in range(2, 11, 2):
                if i <= rating:
                    stars.append(E('span', {'class': 'rating-star1'}))
                elif 0 < i - rating <= 1:
                    stars.append(E('span', {'class': 'rating-star2'}))
                else:
                    stars.append(E('span', {'class': 'rating-star0'}))
            if reason:
                stars.append(E('span', {'class': 'rating-reason'}, reason))
            else:
                stars.append(E('span', {'class': 'rating-score'}, str(rating)))

            if info['title']:
                url = fact_url if fact_url else url_fix(info['url'])
                return E('div', {'class': 'subject-wrapper'},
                         E('a', {'href': url},
                           E('div', {'class': 'subject-cover'},
                             E('img', {'src': url_fix(info['cover'])})
                             ),
                           E('div', {'class': 'subject-info'},
                             E('div', {'class': 'subject-title'},
                               info['title'].decode('utf8')[:TITLE_MAX]
                               ),
                             E('div', {'class': 'subject-rating'}, *stars),
                             E('div', {'class': 'subject-summary'},
                               info['summary'].decode('utf8')[:ABSTRACT_MAX]
                               )
                             )
                           )
                         )
            return E('div', {'class': 'subject-link'},
                     E('a', {'href': fact_url if fact_url else url}, url))
        return E('span', {'data-error': 'no subject url'})


def run():
    config = {
        'entity_decorators': {
            'LINK': LinkEntity(),
            'IMAGE': ImageEntity(),
            'VIDEO': VideoEntity(),
            'SEPARATOR': SeparatorEntity(),
            'SUBJECT': SubjectEntity()
        },
        'composite_decorators': [
            URLDecorator(),
            BrDecorator(),
        ],
        'block_map': {
            'unstyled': {'element': 'p'},
            'blockquote': {'element': 'blockquote'},
            'atomic': {'element': ''},
            'header-one': {'element': 'h1'},
            'header-two': {'element': 'h2'},
        },
        'style_map': {
            'ITALIC': {'fontStyle': 'italic'},
            'BOLD': {'fontWeight': 'bold'},
        }
    }

    json_data = {
        "entityMap": {
            "0": {
                "type": "IMAGE",
                "mutability": "IMMUTABLE",
                "data": {
                    "id": "1111",
                    "src": "http://www.d.com/1.jpg",
                    "caption": "Image caption",
                    "float": "left"
                }
            },
            "1": {
                "type": "LINK",
                "mutability": "MUTABLE",
                "data": {
                    "url": "https://github.com/icelab/draft-js-ast-exporter/好"
                }
            },
            "2": {
                "type": "VIDEO",
                "mutability": "IMMUTABLE",
                "data": {
                    "url": "http://v.qq.com/x/cover/01asoh7vfei7w2c/z03172fgqfr.html"
                }
            },
            "3": {
                "type": "SEPARATOR",
                "mutability": "IMMUTABLE",
                "data": {}
            },
            "4": {
                "type": "IMAGE",
                "mutability": "IMMUTABLE",
                "data": {
                    "id": "1111",
                    "src": "http://www.d.com/1.jpg",
                    "caption": "Image caption",
                    "float": "left"
                }
            },
            "5": {
                "type": "IMAGE",
                "mutability": "IMMUTABLE",
                "data": {
                    "id": "1111",
                    "src": "http://www.d.com/1.jpg",
                    "caption": "Image caption",
                    "float": "left"
                }
            },
            "6": {
                "type": "IMAGE",
                "mutability": "IMMUTABLE",
                "data": {
                    "id": "1111",
                    "src": "http://www.d.com/1.jpg",
                    "caption": "Image caption",
                    "float": "left"
                }
            },
            "7": {
                "type": "IMAGE",
                "mutability": "IMMUTABLE",
                "data": {
                    "id": "1111",
                    "src": "http://www.d.com/1.jpg",
                    "caption": "Image caption",
                    "float": "left"
                }
            },
            "10": {
                "type": "SUBJECT",
                "mutability": "IMMUTABLE",
                "data": {
                    "url": "http://www.book.com/1",
                }
            },
            "11": {
                "type": "SUBJECT",
                "mutability": "IMMUTABLE",
                "data": {
                    "url": "http://www.book.com/1",
                }
            },
            "12": {
                "type": "SUBJECT",
                "mutability": "IMMUTABLE",
                "data": {
                    "url": "http://www.book.com/1",
                }
            },
            "13": {
                "type": "SUBJECT",
                "mutability": "IMMUTABLE",
                "data": {
                    "url": "http://www.book.com/1",
                }
            },
            "14": {
                "type": "SUBJECT",
                "mutability": "IMMUTABLE",
                "data": {
                    "url": "http://www.book.com/1",
                }
            },
            "15": {
                "type": "SUBJECT",
                "mutability": "IMMUTABLE",
                "data": {
                    "url": "http://www.book.com/1",
                }
            },

        },
        "blocks": [
            {
                "key": "59kd9",
                "text": "In your draft-js, exporting your content:",
                "type": "unstyled",
                "depth": 0,
                "inlineStyleRanges": [
                    {
                        "offset": 18,
                        "length": 9,
                        "style": "BOLD"
                    },
                    {
                        "offset": 33,
                        "length": 7,
                        "style": "ITALIC"
                    }
                ],
                "entityRanges": []
            },
            {
                "key": "9vgd",
                "text": "",
                "type": "atomic",
                "depth": 0,
                "inlineStyleRanges": [],
                "entityRanges": [
                    {
                        "offset": 0,
                        "length": 1,
                        "key": 0
                    }
                ]
            },
            {
                "key": "0vgd",
                "text": "",
                "type": "atomic",
                "depth": 0,
                "inlineStyleRanges": [],
                "entityRanges": [
                    {
                        "offset": 0,
                        "length": 1,
                        "key": 4
                    }
                ]
            },
            {
                "key": "8vgd",
                "text": "",
                "type": "atomic",
                "depth": 0,
                "inlineStyleRanges": [],
                "entityRanges": [
                    {
                        "offset": 0,
                        "length": 1,
                        "key": 5
                    }
                ]
            },
            {
                "key": "avgd",
                "text": "",
                "type": "atomic",
                "depth": 0,
                "inlineStyleRanges": [],
                "entityRanges": [
                    {
                        "offset": 0,
                        "length": 1,
                        "key": 6
                    }
                ]
            },
            {
                "key": "dvgd",
                "text": "",
                "type": "atomic",
                "depth": 0,
                "inlineStyleRanges": [],
                "entityRanges": [
                    {
                        "offset": 0,
                        "length": 1,
                        "key": 7
                    }
                ]
            },
            {
                "key": "kst0",
                "text": "Find the project on Github.",
                "type": "unstyled",
                "depth": 0,
                "inlineStyleRanges": [],
                "entityRanges": [
                    {
                        "offset": 20,
                        "length": 6,
                        "key": 1
                    }
                ]
            },
            {
                "key": "8987",
                "type": "atomic",
                "text": "",
                "depth": 0,
                "inlineStyleRanges": [],
                "entityRanges": [
                    {
                        "offset": 0,
                        "length": 1,
                        "key": 2
                    }
                ]
            },
            {
                "key": "abcd",
                "text": "",
                "type": "atomic",
                "depth": 0,
                "inlineStyleRanges": [],
                "entityRanges": [
                    {
                        "offset": 0,
                        "length": 1,
                        "key": 3
                    }
                ]
            },
            {
                "key": "c123",
                "text": "",
                "type": "atomic",
                "depth": 0,
                "inlineStyleRanges": [],
                "entityRanges": [
                    {
                        "offset": 0,
                        "length": 1,
                        "key": 10
                    }
                ]
            },
            {
                "key": "a123",
                "text": "",
                "type": "atomic",
                "depth": 0,
                "inlineStyleRanges": [],
                "entityRanges": [
                    {
                        "offset": 0,
                        "length": 1,
                        "key": 11
                    }
                ]
            },
            {
                "key": "1123",
                "text": "",
                "type": "atomic",
                "depth": 0,
                "inlineStyleRanges": [],
                "entityRanges": [
                    {
                        "offset": 0,
                        "length": 1,
                        "key": 12
                    }
                ]
            },
            {
                "key": "2123",
                "text": "",
                "type": "atomic",
                "depth": 0,
                "inlineStyleRanges": [],
                "entityRanges": [
                    {
                        "offset": 0,
                        "length": 1,
                        "key": 13
                    }
                ]
            },
            {
                "key": "3123",
                "text": "",
                "type": "atomic",
                "depth": 0,
                "inlineStyleRanges": [],
                "entityRanges": [
                    {
                        "offset": 0,
                        "length": 1,
                        "key": 14
                    }
                ]
            },
            {
                "key": "4123",
                "text": "",
                "type": "atomic",
                "depth": 0,
                "inlineStyleRanges": [],
                "entityRanges": [
                    {
                        "offset": 0,
                        "length": 1,
                        "key": 15
                    }
                ]
            },
            {
                "key": "ag6qs",
                "text": "<b>DraftJS</b> AST Exporter<script>miao()</script>",
                "type": "header-two",
                "depth": 0,
                "inlineStyleRanges": [],
                "entityRanges": []
            },
        ]
    }

    exporter = HTML(config)

    print "\n<<<---"
    pr = cProfile.Profile()
    pr.enable()

    html = exporter.render(json_data)
    pr.disable()
    assert html

    p = Stats(pr)
    p.strip_dirs().sort_stats('cumulative').print_stats(20)
    print "\n--->>>"
    # print DOM.pretty_print(html)


if __name__ == '__main__':
    run()