Dekker1/Fourmi
Archived
1
0
Fork 0
Code

NIST now scrapes the symbol table for later use

Browse Source
This commit is contained in:
Rob tB 2014-05-04 21:20:46 +02:00
parent 0cec4bd2d8
commit 930eb6cad5
1 changed files with 11 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

13
FourmiCrawler/sources/NIST.py
View File

@ -13,8 +13,17 @@ class NIST(Source):
def __init__(self): def __init__(self):
Source.__init__(self) Source.__init__(self)
def parse(self, reponse): def parse(self, response):
pass sel = Selector(response)
symbol_table = {}
tds = sel.xpath('//table[@class="symbol_table"]/tr/td')
for (symbol_td, name_td) in zip(tds[::2], tds[1::2]):
symbol = ''.join(symbol_td.xpath('node()').extract())
name = name_td.xpath('text()').extract()[0]
symbol_table[symbol] = name
log.msg('NIST symbol: |%s|, name: |%s|' % (symbol, name),
level=log.DEBUG)
def new_compound_request(self, compound): def new_compound_request(self, compound):
return Request(url=self.website[:-1] + self.search % compound, return Request(url=self.website[:-1] + self.search % compound,