Skip to content

Instantly share code, notes, and snippets.

@LitvinenkoD89
Created July 27, 2017 14:35
Show Gist options
  • Save LitvinenkoD89/94bda5d621bb53facdfb3ac8641b4ca3 to your computer and use it in GitHub Desktop.
Save LitvinenkoD89/94bda5d621bb53facdfb3ac8641b4ca3 to your computer and use it in GitHub Desktop.
henribendel
# HENRI BENDEL {source_name: 'henribendel', batch_number: 6, request_id: 6000000075, request_name: 'HENRI BENDEL'}
scraper_service.scrape do |browser, scraper, init_vars|
easy_seeder = Library.lib('EasySeeder')
easy_extractor = Library.lib('EasyExtractor')
fetcher_agent = Library.lib('FetcherAgent')
easy_seeder.seed(source_name: init_vars[:source_name]) do
loop_postal_codes "us", radius: 10 do |zip, lat, long|
queue_url "http://www.henribendel.com/on/demandware.store/Sites-HB-Site/default/Stores-FindStores", {
method: :post,
body: {
dwfrm_storelocator_countryCode:'US',
dwfrm_storelocator_distanceUnit:'mi',
dwfrm_storelocator_postalCode:zip,
dwfrm_storelocator_findbyzip:'Search'
}
}
end
end
easy_extractor.extract(
source_name: init_vars[:source_name],
scraper: scraper,
batch_number: init_vars[:batch_number],
request_id: init_vars[:request_id],
request_name: init_vars[:request_name],
) do
find_pages page_format: :html, page_type: 'search' do |url, parser_page, page|
page.search('.storeinformation').each do |store|
name = store.children[1].text
tmp = store.children[3].inner_html.gsub("\t", '').gsub("\r", '').gsub("\n", '').split('<br>').map(&:strip)
address_container = tmp.join('<br>')
country = tmp.pop
adr = tmp.pop
street1,street2 = tmp
city, state_zip = adr.split(', ')
state, zip_code = state_zip.split(' ')
# tmp = el.parent.next.next.text.split('.push(new Array(')
# lat_lang_container = tmp[1]
# latitude, longitude = lat_lang_container.split(", ")
doc_id = store_doc({
# store_id: id,
brand: 'HENRI BENDEL',
type: "HENRI BENDEL",
# property_id: id,
name: '', # required
address_1: street1,
address_2: street2,
city: city,
state: state,
zipcode: zip_code,
country: '',
# lat: latitude,
# long: longitude,
address_container_html: address_container,
map_link: store.parent.parent.at('.googlemap').attr('name'),
# location_url: location_url,
flags: {
}
})
end
end
end
end
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment