def scrape_entity_page(self, url):
entity_root = html_parsing.parse_tree(url).getroot()
name = html_parsing.tostring(entity_root.xpath('.//div[@class="title-desc-inner"]//h1')[0])
content_p_elems = entity_root.xpath(".//div[@class='content']//div[not(@class='image-caption')]/p")
description = '\n\n'.join(html_parsing.tostring(p) for p in content_p_elems)
photo_urls = entity_root.xpath(".//div[@class='content']//img/@data-src")
return data.Entity(name=name, description=description, photo_urls=photo_urls)
def placemark_to_entity(self, placemark_elem):
pm = placemark_elem
name_elem = self.xpath(pm, 'ns:name')
name = tostring(name_elem[0]) if name_elem is not None else None
description_elem = self.xpath(pm, 'ns:description')
description_html = tostring(description_elem[0]) if description_elem else None
description = self.html_str_to_text(description_html) if description_html else None
latlng = self.parse_latlng(self.xpath(pm, 'ns:Point'))
return data.Entity(name=name, description=description, latlng=latlng)
def get_opening_hours(self):
hours_nodes = self.root.xpath('.//div[@class="place-resume"]//table[@class="hours-open"]//tr')
texts = []
for node in hours_nodes:
day = tostring(node.xpath('.//td')[0])
times = tostring(node.xpath('.//td')[1])
texts.append('%s\t%s' % (day, times))
source_text = '\n'.join(texts)
return data.OpeningHours(source_text=source_text)
def get_opening_hours(self):
timeframes = self.root.xpath('.//div[@class="venueDetail"]//div[@class="allHours"]//ul[@class="timeframes"]//li[@class="timeframe"]')
timeframes_text = []
for t in timeframes:
text = '%s\t%s' % (tostring(t.xpath('.//span[@class="timeframeDays"]')[0]),
tostring(t.xpath('.//span[@class="timeframeHours"]')[0]))
timeframes_text.append(text)
source_text = '\n'.join(timeframes_text)
return data.OpeningHours(source_text=source_text)
def parse_latlng(self):
lat_elem = self.root.find('.//span[@class="geo-default"]//span[@class="latitude"]')
lng_elem = self.root.find('.//span[@class="geo-default"]//span[@class="longitude"]')
if lat_elem is not None:
return utils.latlng_to_decimal(tostring(lat_elem), tostring(lng_elem))
geo_elem = self.root.find('.//span[@class="geo-default"]//span[@class="geo"]')
if geo_elem is not None:
lat, lng = tostring(geo_elem).split(';')
return {
'lat': float(lat.strip()),
'lng': float(lng.strip())
}
return None
def get_address(self):
city, country = self.get_city_and_country()
if '/hotels/' in self.url:
street_and_city_node = self.root.xpath('.//span[contains(@class, "lodging__subtitle--address")]')[0]
street_and_city = tostring(street_and_city_node, True)
return '%s %s' % (street_and_city, country)
else:
street_node = self.root.find('.//dl[@class="info-list"]//dd[@class="copy--meta"]//strong')
if street_node is not None:
street = tostring(street_node, True)
return '%s %s %s' % (street, city, country)
else:
google_place = self.lookup_google_place()
return google_place.address if google_place else None
def get_address(self):
street_node = self.root.find('.//li[@class="address"]//span[@itemprop="streetAddress"]')
locality_node = self.root.find('.//li[@class="address"]//span[@itemprop="addressLocality"]')
postal_node = self.root.find('.//li[@class="address"]//span[@itemprop="postalCode"]')
if street_node is not None and locality_node is not None:
street = tostring(street_node, True).replace(',', '')
locality = tostring(locality_node, True).replace(',', '')
if postal_node is not None:
postal_code = tostring(postal_node, True).replace(',', '')
return '%s %s %s' % (street, locality, postal_code)
else:
return '%s %s' % (street, locality)
else:
return self.lookup_google_place().address
def get_address(self):
infocard_cells = self.root.findall('.//table[@class="infobox vcard"]//tr')
for tr in infocard_cells:
th = tr.find('.//th')
if th is not None and th.text == 'Address':
return tostring(tr.find('.//td'))
return None
def get_location_name(self):
page_header_node = self.root.xpath('.//h1[contains(@class, "header")]')
if page_header_node:
page_header = html_parsing.tostring(page_header_node[0])
if 'Travel Guide for' in page_header:
return page_header.replace('Travel Guide for ', '')
return None
def get_sub_category(self):
category_node = self.root.xpath('.//li[contains(@class, "categoriesList")]//div[contains(@class, "categories")]')
if category_node:
categories = tostring(category_node[0]).split(',')
categories = [s.strip().lower() for s in categories]
else:
categories = []
path_root = self.get_path_root()
tc_category = self.get_category()
if tc_category == values.Category.FOOD_AND_DRINK:
# Gogobot doesn't seem to have categories like restaurant/bar/bakery
# They do have cuisine types like French though.
return values.SubCategory.RESTAURANT
elif tc_category == values.Category.ATTRACTIONS:
if contains_any(categories, ['monument', 'historic site']):
return values.SubCategory.LANDMARK
elif contains_any(categories, ['sights and museums', 'art museum']):
return values.SubCategory.MUSEUM
elif tc_category == values.SubCategory.LODGING:
if path_root.endswith('hotel'):
return values.SubCategory.HOTEL
elif path_root.endswith('vacation-rental'):
return values.SubCategory.VACATION_RENTAL
return None
def get_sub_category(self):
category_node = self.root.find('.//div[@class="primaryInfo"]//div[@class="categories"]')
category_str = tostring(category_node).lower()
parsed_category = self.get_category()
if parsed_category == values.Category.FOOD_AND_DRINK:
if 'restaurant' in category_str:
return values.SubCategory.RESTAURANT
if 'coffee' in category_str:
return values.SubCategory.COFFEE_SHOP
if 'bar' in category_str:
return values.SubCategory.BAR
if contains_any(category_str, ('ice cream', 'dessert')):
return values.SubCategory.DESSERT
if 'bakery' in category_str:
return values.SubCategory.BAKERY
if parsed_category == values.Category.LODGING:
if contains_any(category_str, ('hotel', 'motel')):
return values.SubCategory.HOTEL
if 'hostel' in category_str:
return values.SubCategory.HOSTEL
if parsed_category == values.Category.ENTERTAINMENT:
if contains_any(category_str, ('concert hall', 'jazz club', 'rock club')):
return values.SubCategory.MUSIC
if 'stadium' in category_str:
return values.SubCategory.SPORTS
return None
def get_sub_category(self):
category_node = self.root.xpath('.//div[@class="place-post"]//span[@class="date"]')[0]
category_text = tostring(category_node).split('|')[0].strip().lower()
if 'bar' in category_text:
return values.SubCategory.BAR
elif 'club' in category_text:
return values.SubCategory.NIGHTCLUB
return values.SubCategory.RESTAURANT
def get_category(self):
categories_parent = self.root.find('body//span[@class="category-str-list"]')
categories_str = tostring(categories_parent)
categories = [c.strip().lower() for c in categories_str.split(',')]
if 'hotel' in categories or 'hotels' in categories or 'bed & breakfast' in categories:
return values.Category.LODGING
else:
return values.Category.FOOD_AND_DRINK
def parse(self):
raw_entities = []
for placemark in self.xpath(self.root, './/ns:Placemark'):
raw_entities.append(self.placemark_to_entity(placemark))
entities = utils.parallelize(self.augment_entity, [(e,) for e in raw_entities])
name = tostring(self.xpath(self.root, 'ns:Document/ns:name')[0])
# TODO: Parse the latlngs into a Bounds object for the trip plan.
# Right now this is happening the javascript as a hack.
return data.TripPlan(name=name, entities=entities)
def get_opening_hours(self):
hours_nodes = self.root.xpath('.//table[contains(@class, "hours-table")]//tr')
texts = []
for node in hours_nodes:
day = tostring(node.find('th'))
times = tostring_with_breaks(node.find('td'))
texts.append('%s\t%s' % (day, times))
source_text = '\n'.join(texts)
return data.OpeningHours(source_text=source_text)
def get_raw_entities(self):
items = self.root.xpath(
".//div[@id='guides']//h3[text() = 'Top Things to Do' or text() = 'Top Things to See and Do']/following-sibling::ul//li")
entities = []
for item in items:
raw_text = html_parsing.tostring(item).strip()
name, desc = re.split(u'\s?(?:\u2013|-|:)\s?', raw_text, 1, re.UNICODE)[:2]
entities.append(data.Entity(name=name, description=desc))
return entities
def get_raw_entities(self):
entities = []
items = self.root.xpath('.//h2[@class="accordion-title" and contains(., "At a Glance")]/following-sibling::div//p')
for item in items:
num_stars = len(item.text.strip())
starred = num_stars == 3
name = item.xpath('.//strong')[0].text.strip()
temp_html = re.sub('<strong>.*</strong>', 'SPLIT_POINT', etree.tostring(item))
temp_node = html_parsing.parse_tree_from_string(temp_html.encode('utf-8'))
desc = html_parsing.tostring(temp_node).split('SPLIT_POINT')[1].strip()
entities.append(data.Entity(name=name, starred=starred, description=desc))
return entities
def get_sub_category(self):
categories_parent = self.root.find('body//span[@class="category-str-list"]')
categories_str = tostring(categories_parent)
categories = [c.strip().lower() for c in categories_str.split(',')]
if 'bed & breakfast' in categories:
return values.SubCategory.BED_AND_BREAKFAST
elif 'hotel' in categories or 'hotels' in categories:
return values.SubCategory.HOTEL
else:
for category in categories:
if 'bar' in category:
return values.SubCategory.BAR
return values.SubCategory.RESTAURANT
def get_category(self):
category_node = self.root.find('.//div[@class="primaryInfo"]//div[@class="categories"]')
category_str = tostring(category_node).lower()
if contains_any(category_str, ('restaurant', 'bar', 'ice cream', 'dessert', 'bakery', 'coffee')):
return values.Category.FOOD_AND_DRINK
if contains_any(category_str, ('hotel', 'motel', 'hostel')):
return values.Category.LODGING
if contains_any(category_str, ('monument', 'landmark')):
return values.Category.ATTRACTIONS
if contains_any(category_str, ('store', 'shop', 'boutique')):
return values.Category.SHOPPING
if contains_any(category_str, ('concert hall', 'jazz club', 'rock club', 'stadium')):
return values.Category.ENTERTAINMENT
return None
def get_sub_category(self):
url = self.url.lower()
if '/hotels/' in url:
hotel_type_node = self.root.xpath('.//span[contains(@class, "lodging__subtitle")]')[0]
hotel_type = tostring(hotel_type_node, True)
if hotel_type == 'Guesthouse':
return values.SubCategory.BED_AND_BREAKFAST
elif hotel_type == 'Hostel':
return values.SubCategory.HOSTEL
else:
return values.SubCategory.HOTEL
elif '/restaurants/' in url:
return values.SubCategory.RESTAURANT
return None
def run(self):
entity_datas = []
for p in self.getroot().findall('.//footer//div[@class="story-info"]//p'):
line_text = html_parsing.tostring(p, with_tail=False)
if self.NUMBERED_LINE_RE.match(line_text):
for child in p.iterchildren():
tag = child.tag.lower()
text = html_parsing.tostring(child, with_tail=False)
if tag == 'strong':
if self.NUMBERED_LINE_RE.match(text):
name = text.split('.')[1]
else:
name = text
current_entity = EntityData(name=name.strip().strip(string.punctuation))
entity_datas.append(current_entity)
elif tag == 'a':
current_entity.website = child.get('href')
tail = child.tail.strip().strip(string.punctuation) if child.tail else ''
if tail:
parts = tail.split(';')
current_entity.address = parts[0].strip()
if len(parts) >= 2:
current_entity.phone = parts[1].strip()
self.build_from_entity_data(entity_datas)
def get_description(self):
desc_nodes = self.root.xpath('.//div[@id="listing_main"]//div[@class="listing_description"]')
if not desc_nodes:
return None
desc_node = desc_nodes[0]
details_link = desc_node.xpath('.//a/@href')
if details_link:
url = self.absolute_url(details_link[0])
details_page_tree = html_parsing.parse_tree(url)
details_node = details_page_tree.getroot().xpath('.//div[@class="articleBody"]')[0]
if details_node.xpath('.//p'):
return html_parsing.join_element_text_using_xpaths(details_node, ['.//p'], '\n\n')
else:
return html_parsing.tostring(details_node)
elif desc_node.xpath('.//span[@class="onShow"]'):
return ''.join(desc_node.xpath('.//span[@class="onShow"]/text()')).strip()
else:
return ''.join(desc_node.xpath('text()')).strip()
def get_entity_overrides(self):
overrides = {}
current_day = 0
for node in self.root.xpath(
'.//div[@id="GUIDE_DETAIL"]//div[contains(@class, "guideOverview")]')[0].itersiblings():
if node.tag == 'h5':
current_day = int(node.text.replace('Day', '').strip())
elif node.tag == 'div':
tags = [data.Tag(text='Day %d' % current_day)]
desc = None
# Items with long descriptions on the entity page will not have 'shortDesc',
# node, they'll have an untagged <p> tagged that contains a 'more' link.
desc_nodes = node.xpath('.//p[contains(@id, "shortDesc")]')
if desc_nodes:
desc = html_parsing.tostring(desc_nodes[0])
rel_source_url = node.xpath('div[@class="guideItemInfo"]//a[@class="titleLink"]/@href')[0]
overrides[self.absolute_url(rel_source_url)] = data.Entity(tags=tags, description=desc)
return overrides
def get_trip_plan_name(self):
if self.trip_plan_name:
return self.trip_plan_name
return html_parsing.tostring(self.getroot().find('.//h1[@id="HEADING"]'))
def get_description(self):
return html_parsing.tostring(
self.root.xpath('.//div[@class="excerpt"]/p')[0])
def get_address(self):
elems = self.root.findall('body//div[@class="addresspanel"]//p[@class="address"]')
return '%s %s' % (tostring(elems[0], True), tostring(elems[1], True))
def get_description(self):
guide_text = html_parsing.tostring(self.root.find(".//div[@id='guides']")).strip()
summary_text = guide_text[:guide_text.find('Top Things to Do')].strip()
if summary_text.startswith(self.get_title()):
summary_text = summary_text[len(self.get_title()):].strip()
return summary_text
def get_trip_plan_name(self):
base_name = super(Nytimes36hours, self).get_trip_plan_name()
if base_name:
return base_name
return html_parsing.tostring(self.getroot().find('.//h1[@itemprop="headline"]'))
def get_opening_hours(self):
source_text = tostring(self.root.xpath('.//dl[@class="info-list"]//dt[contains(@class, "icon--time")]/following-sibling::dd')[0])
return data.OpeningHours(source_text=source_text)
def get_location_name(self):
return html_parsing.tostring(
self.root.xpath('.//div[contains(@class, "left-sidebar")]//h3')[0])
