--- a/scrape.py +++ b/scrape.py @@ -1,1 +1,99 @@ +#http://packages.python.org/CouchDB/client.html +import couchdb +import urllib2 +from BeautifulSoup import BeautifulSoup +import re +import hashlib +from urlparse import urljoin +import time +import os +#http://diveintopython.org/http_web_services/etags.html +class NotModifiedHandler(urllib2.BaseHandler): + def http_error_304(self, req, fp, code, message, headers): + addinfourl = urllib2.addinfourl(fp, headers, req.get_full_url()) + addinfourl.code = code + return addinfourl + +def fetchURL(docsdb, url, agencyID): + hash = hashlib.md5(url).hexdigest() + req = urllib2.Request(url) + print "Fetching %s", url + doc = docsdb.get(hash) + if doc == None: + doc = {'_id': hash, 'agencyID': agencyID} + #if there is a previous version stored in couchdb, load caching helper tags + if doc.has_key('etag'): + req.add_header("If-None-Match", doc['etag']) + if doc.has_key('last_modified'): + req.add_header("If-Modified-Since", doc['last_modified']) + + opener = urllib2.build_opener(NotModifiedHandler()) + url_handle = opener.open(req) + headers = url_handle.info() # the addinfourls have the .info() too + doc['etag'] = headers.getheader("ETag") + doc['last_modified'] = headers.getheader("Last-Modified") + doc['web_server'] = headers.getheader("Server") + doc['powered_by'] = headers.getheader("X-Powered-By") + doc['file_size'] = headers.getheader("Content-Length") + doc['mime_type'] = headers.getheader("Content-Type").split(";")[0] + if hasattr(url_handle, 'code'): + if url_handle.code == 304: + print "the web page has not been modified" + return None + else: + content = url_handle.read() + docsdb.save(doc) + doc = docsdb.get(hash) # need to get a _rev + docsdb.put_attachment(doc, content, str(time.time())+"-"+os.path.basename(url), doc['mime_type']) + return (doc['mime_type'], content) + #store as attachment epoch-filename + else: + print "error %s in downloading %s", url_handle.code, URL + #record/alert error to error database + + + + +def scrapeAndStore(docsdb, url, depth, agencyID): + (mime_type,content) = fetchURL(docsdb, url, agencyID) + if content != None: + if mime_type == "text/html" or mime_type == "application/xhtml+xml" or mime_type =="application/xml": + # http://www.crummy.com/software/BeautifulSoup/documentation.html + soup = BeautifulSoup(content) + navIDs = soup.findAll(id=re.compile('nav|Nav|menu|bar')) + for nav in navIDs: + print "Removing element", nav['id'] + nav.extract() + navClasses = soup.findAll(attrs={'class' : re.compile('nav|menu|bar')}) + for nav in navClasses: + print "Removing element", nav['class'] + nav.extract() + links = soup.findAll('a') # soup.findAll('a', id=re.compile("^p-")) + for link in links: + if link.has_key("href"): + if link['href'].startswith("http"): + linkurl = link['href'] + else: + linkurl = urljoin(url,link['href']) + print linkurl + #for each unique link + # if + #if html mimetype + # go down X levels, + # diff with last stored attachment, store in document + #if not + # remember to save parentURL and title (link text that lead to document) + + +couch = couchdb.Server('http://127.0.0.1:5984/') + +# select database +agencydb = couch['disclosr-agencies'] +docsdb = couch['disclosr-documents'] + +for row in agencydb.view('app/getScrapeRequired'): #not recently scraped agencies view? + agency = agencydb.get(row.id) + print agency['name'] + scrapeAndStore(docsdb, agency['website'],1,agency['_id']) +