more scrapers
Former-commit-id: 624a5af265cf19be61f44ec7a025acad87daa5f8
|
import sys,os |
|
sys.path.insert(0, os.path.join(os.path.dirname(__file__) or '.', '../')) |
|
import genericScrapers |
|
import scrape |
|
from bs4 import BeautifulSoup |
|
|
|
#http://www.doughellmann.com/PyMOTW/abc/ |
|
class ScraperImplementation(genericScrapers.GenericOAICDisclogScraper): |
|
def getTable(self,soup): |
|
return soup.find(id = "maincontentcontainer").table |
|
def getColumnCount(self): |
|
return 5 |
|
def getColumns(self,columns): |
|
(date, disclogdate, title, description, notes) = columns |
|
return (date, date, title, description, notes) |
|
|
|
if __name__ == '__main__': |
|
print 'Subclass:', issubclass(ScraperImplementation, genericScrapers.GenericOAICDisclogScraper) |
|
print 'Instance:', isinstance(ScraperImplementation(), genericScrapers.GenericOAICDisclogScraper) |
|
ScraperImplementation().doScrape() |
|
|
|
import sys,os |
|
sys.path.insert(0, os.path.join(os.path.dirname(__file__) or '.', '../')) |
|
import genericScrapers |
|
import scrape |
|
from bs4 import BeautifulSoup |
|
|
|
#http://www.doughellmann.com/PyMOTW/abc/ |
|
class ScraperImplementation(genericScrapers.GenericOAICDisclogScraper): |
|
def getColumns(self,columns): |
|
(date, id, title, description, notes) = columns |
|
return (id, date, title, description, notes) |
|
|
|
if __name__ == '__main__': |
|
print 'Subclass:', issubclass(ScraperImplementation, genericScrapers.GenericOAICDisclogScraper) |
|
print 'Instance:', isinstance(ScraperImplementation(), genericScrapers.GenericOAICDisclogScraper) |
|
ScraperImplementation().doScrape() |
|
|
|
import sys,os |
|
sys.path.insert(0, os.path.join(os.path.dirname(__file__) or '.', '../')) |
|
import genericScrapers |
|
import scrape |
|
from bs4 import BeautifulSoup |
|
|
|
#http://www.doughellmann.com/PyMOTW/abc/ |
|
class ScraperImplementation(genericScrapers.GenericOAICDisclogScraper): |
|
def getColumnCount(self): |
|
return 7 |
|
def getColumns(self,columns): |
|
(id, date, title, description,link,deldate, notes) = columns |
|
return (id, date, title, description, notes) |
|
|
|
if __name__ == '__main__': |
|
print 'Subclass:', issubclass(ScraperImplementation, genericScrapers.GenericOAICDisclogScraper) |
|
print 'Instance:', isinstance(ScraperImplementation(), genericScrapers.GenericOAICDisclogScraper) |
|
ScraperImplementation().doScrape() |
|
|
|
import sys,os |
|
sys.path.insert(0, os.path.join(os.path.dirname(__file__) or '.', '../')) |
|
import genericScrapers |
|
import scrape |
|
from bs4 import BeautifulSoup |
|
|
|
#http://www.doughellmann.com/PyMOTW/abc/ |
|
class ScraperImplementation(genericScrapers.GenericOAICDisclogScraper): |
|
def getColumns(self,columns): |
|
(id, date, title, description, notes) = columns |
|
return (id, date, title, description, notes) |
|
|
|
if __name__ == '__main__': |
|
print 'Subclass:', issubclass(ScraperImplementation, genericScrapers.GenericOAICDisclogScraper) |
|
print 'Instance:', isinstance(ScraperImplementation(), genericScrapers.GenericOAICDisclogScraper) |
|
ScraperImplementation().doScrape() |
|
|