How to use re() to extract data from javascript variable using scrapy?

Question

My items.py file goes like this:

from scrapy.item import Item, Field

class SpiItem(Item):
    title = Field()
    lat = Field()
    lng = Field()
    add = Field()

and the spider is:

import scrapy
import re

from spi.items import SpiItem

class HdfcSpider(scrapy.Spider):
    name = "hdfc"
    allowed_domains = ["hdfc.com"]
    start_urls = ["http://hdfc.com/branch-locator"]

    def parse(self,response):
        addresses = response.xpath('//script')
        for sel in addresses:
            item = SpiItem()
            item['title'] = sel.xpath('//script[@type="text/javascript"][1]').re('(?<="title":).+(?=")')
            item['lat'] = sel.xpath('//script[@type="text/javascript"][1]').re('(?<="latitude":).+(?=")')
            item['lng'] = sel.xpath('//script[@type="text/javascript"][1]').re('(?<="longitude":).+(?=")')
            item['add'] = sel.xpath('//script[@type="text/javascript"][1]').re('(?<="html":).+(?=")')
            yield item

The whole javascript code, on viewing page source, is written inside: //html/body/table/tbody/tr[348]/td[2].

Why is my code not working? I want to extract just the four fields mentioned in the items file.

Answer 1

Instead of extracting field by field using regular expressions, extract the complete locations object, load it via json.loads() and extract the desired data from the Python dictionary you'll get:

def parse(self,response):
    pattern = re.compile(r"var locations= ({.*?});", re.MULTILINE | re.DOTALL)
    locations = response.xpath('//script[contains(., "var locations")]/text()').re(pattern)[0]
    locations = json.loads(locations)
    for title, data in locations.iteritems():
        print title