python - How to crawl only two pre-defined pages, but they scrape different items? -

i'm trying scrape site driven user input. example, user gives me pid of product , name, , separate program launch spider, gather data, , return user.

however, information want product , person found in 2 links xml. if know these 2 links , pattern, how build callback parse different items?

for example, if have these 2 items defined:

class personitem(item):      name = field()      ...  class productitem(item):      pid = field()      ... 

and know links have pattern:

www.example.com/person/*<name_of_person>*/person.xml www.example.com/*<product_pid>*/product.xml 

then spider this:

class myspider(basespider):      name = "myspider"       # simulated given user      pid = "4545-fw"      person = "bob"       allowed_domains = ["http://www.example.com"]      start_urls = ['http://www.example.com/person/%s/person.xml'%person, 'http://www.example.com/%s/product.xml'%pid]       def parse(self, response):            # not sure here if scrapping person or item 

i know can define rules using rule(sgmllinkextractor()) , giving person , product each own parse callback. however, i'm not sure how apply here since think rules meant crawling deeper, whereas need scrape surface level.

if want retro-active put logic in parse():

def parse(self, response):     if 'person.xml' in response.url:         item = personitem()     elif 'product.xml' in response.url:         item = productitem()     else:         raise exception('could not determine item type') 

update:

if want pro-active override start_requests():

class myspider(basespider):     name = "myspider"     allowed_domains = ["example.com"]     pid = "4545-fw"     person = "bob"      def start_requests(self):         start_urls = (             ('http://www.example.com/person/%s/person.xml' % self.person, personitem),             ('http://www.example.com/%s/product.xml' % self.pid, productitem),         )         url, cls in start_urls:             yield request(url, meta=dict(cls=cls))      def parse(self, response):         item = response.meta['cls']()