From: Maxious Date: Sun, 25 Nov 2012 23:29:37 +0000 Subject: misc disclogs X-Git-Url: https://maxious.lambdacomplex.org/git/?p=disclosr.git&a=commitdiff&h=e3ed0543748b7d3cd69f581b4ea2876056ad0549 --- misc disclogs Former-commit-id: d326d97b180d7cc17008504f71240adc17cd2804 --- --- /dev/null +++ b/documents/scrapers/157cbe654bdaa0a48e6650152ae34489.py @@ -1,1 +1,21 @@ +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 5 + def getColumns(self,columns): + (id, date, title, description, notes) = columns + return (id, date, description, title, notes) + def getTable(self,soup): + return soup.find_all('table')[4] + +if __name__ == '__main__': + print 'Subclass:', issubclass(ScraperImplementation, genericScrapers.GenericOAICDisclogScraper) + print 'Instance:', isinstance(ScraperImplementation(), genericScrapers.GenericOAICDisclogScraper) + ScraperImplementation().doScrape() + --- /dev/null +++ b/documents/scrapers/3b626d4c0ce3b3b47ee197e072f36023.py @@ -1,1 +1,19 @@ +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 3 + def getColumns(self,columns): + (id, date, title) = columns + return (id, date, title,title,title) + +if __name__ == '__main__': + print 'Subclass:', issubclass(ScraperImplementation, genericScrapers.GenericOAICDisclogScraper) + print 'Instance:', isinstance(ScraperImplementation(), genericScrapers.GenericOAICDisclogScraper) + ScraperImplementation().doScrape() + --- /dev/null +++ b/documents/scrapers/795c33ed030293dbdb155c909ea47e37.py @@ -1,1 +1,21 @@ +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, notes, deletedate, otherinfo) = columns + return (id, date, description, title, notes) + #def getTable(self,soup): + # return soup.find(class_ = "box").table + +if __name__ == '__main__': + print 'Subclass:', issubclass(ScraperImplementation, genericScrapers.GenericOAICDisclogScraper) + print 'Instance:', isinstance(ScraperImplementation(), genericScrapers.GenericOAICDisclogScraper) + ScraperImplementation().doScrape() +