def scrape ( self ):
acts, datemodified = self.getActs()
lastroot = self.getRoot()
if datemodified is not lastroot.released:
root = self.makeRoot(datemodified=datemodified)
else:
root = lastroot
for act in acts:
self.processAct( act, root )
root.analyze()
store.commit()
def processFile( self, fn, root ):
log.debug('Processing file %s...' % fn)
self.progress(label=fn)
soup = BS( open(fn, 'r').read() )
if fn.lower().endswith('a.xml'):
# we think this is an appendix, let's try to find an appendix
# any appendix not identified in the form /us/usc/tN[aA] will NOT be processed
appendices = soup.select('appendix[identifier$=a]')
appendices += soup.select('appendix[identifier$=A]')
self.progress(i=len(appendices))
for appendix in appendices:
log.debug('Processing appendix: %s@%s' % ( soup.appendix['identifier'], self.rp ) )
self.processAppendix(appendix, root)
else:
self.processTitle(soup.title, root)
log.debug('Finished processing file %s...' % fn)
store.commit()
def scrape (self):
if not self.rp:
r = Cache.get(self.nation.cfg['entrypoint'])
soup = BS(str(r))
#find current release point
log.info("No release point specified, retreiving latest...")
# this failed fantastically - we'll get the RP from the zipurl
#self.rp = utf8(soup.findAll('h3', attrs={'class': 'releasepointinformation'})[0].text.split()[-1])
log.info("Found release point %s" % self.rp)
#find the download url
self.zipurl = self.nation.cfg['entrypoint'].rpartition('/')[0] + '/' + soup.findAll('a', title='All USC Titles in XML')[0]['href']
# new way to set the rp using the zipurl's filename
self.rp = utf8( self.zipurl.rpartition('@')[-1].partition('.')[0] )
else:
log.info('Using specified release point %s...' % self.rp)
# don't actually need this
# rpurl = 'http://uscode.house.gov/download/releasepoints/us/pl/%s/%s/usc-rp@%s.htm' % (tuple(self.rp.split()) + (self.rp,))
self.zipurl = 'http://uscode.house.gov/download/releasepoints/us/pl/%s/%s/xml_uscAll@%s.zip' % (tuple(self.rp.split('-')) + (self.rp,))
log.debug("Using zipurl: %s" % self.zipurl)
class FileNotThere (Exception): pass
class XMLNotThere( Exception ): pass
class AllGood( Exception ): pass
filename = self.zipurl.rpartition('/')[-1]
xmldir = self._workdir + os.sep + 'xml' + os.sep
# check to see if we have xml that works
# if we don't check to see if we have a zip file
# if we don't, download it
# if we do, extract it
# check the xml again, if it's good, proceed
# if it's not, error out
try:
assert os.path.exists(xmldir + 'usc01.xml')
soup = BS(open(xmldir + os.sep + 'usc01.xml', 'r').read())
xmlrp = soup.find('docpublicationname').text.split('@')[-1]
#old way to get rp, the new way is much better
# xmlrp = soup.title.first("note", topic="miscellaneous").text.split()[-1]
if xmlrp == self.rp:
raise AllGood
else:
raise XMLNotThere
except (XMLNotThere,AssertionError):
# delete directory if it exists
if os.path.exists(xmldir):
shutil.rmtree(xmldir)
# if there's no xml file, download it
if not os.path.exists(self._workdir + os.sep + filename):
log.info('No zipfile found for release point, downloading...')
self.downloadFile(self.zipurl, filename)
# now we should have a zipfile and no existing xmldir
log.info('Extracting file %s...' % filename)
zf = ZipFile(self._workdir + os.sep + filename, 'r')
# older release points do not have an interior xml/ dir
if not all( [ n.startswith('xml/') for n in zf.namelist()]):
zf.extractall(xmldir)
else:
zf.extractall(self._workdir)
# double check the xml now...
assert os.path.exists(xmldir + 'usc01.xml')
# it may be problematic to rely on the RP information in the XML documents provided
# rp 113-21 (the earliest presently available) does not include this in the
# docpublicationname meta tag
#soup = BS(open(xmldir + os.sep + 'usc01.xml', 'r').read())
#xmlrp = soup.find('docpublicationname').text.split('@')[-1]
#if xmlrp != self.rp:
# raise XMLNotThere('XML did not check out after extraction.')
except AllGood:
pass
except:
raise
log.info('All checks passed...')
xf = os.listdir(xmldir)
root = self.findOrCreateRoot()
xf = [xmldir + f for f in xf if f.endswith('.xml')]
xf.sort()
log.info("Processing %i files..." % len(xf))
self.bar = progress.Bar(label='US', expected_size=1000*len(xf))
self.progress( i=len(xf) )
for fn in xf:
self.processFile(fn, root)
self.progress(rollup=1000)
log.info('Analyzing code...')
self.progress(label="Analyzing")
root.analyze(commit=True, bar=self.bar)
store.commit()
log.info('Scrape completed.')
def processAct( self, actcid, parent ):
log.info('Processing act: %s' % actcid)
soup = self.getActSoup(actcid)
act = self.findOrCreateAct(parent.released, actcid, parent.rev)
act.parent = parent
act.cid = actcid
act.released = parent.released
act.rev = parent.rev
act.depth = 1
act.pre = Text.make( soup.title.text )
act.text = Text.make( soup.select("section.intro")[0].text )
act.meta = utf8(soup.select("p#assentedDate")[0].text.rpartition('.')[0])
doc = soup.select("div.docContents div")[0]
#so much easier to use the CSS selector
#sections = [i['id'] for i in doc.select("[id]") if i['id'].startswith('h-')]
id_prefix = 'h-'
sections = [i['id'] for i in doc.select('[id^={}]'.format(id_prefix))]
classAndTag = lambda o: isinstance(o, Tag) and o.has_attr('class')
if sections:
for secid in progress.bar(sections, label=act.cid):
sec = self.findOrCreateSection(act.released, secid, act)
soup = doc.select("[id=%s]" % secid)[0]
sec.pre = Text.make(soup.text)
sec.cid = secid
sec.depth=2
sec.parent = act
sec.released = act.released
sec.rev = act.rev
stop = False
sib = soup.nextSibling
content = ""
for t in soup.select(".wb-invisible"):
t.clear()
while not stop:
if classAndTag(sib):
if sib.has_attr('id') and sib['id'].startswith('h-'):
stop = True
elif sib.name == 'section':
stop = True
elif any( c in ['Definition',
'Section',
'MarginalNote',
'ProvisionList',
'Part',
'Subheading',
'MarginalNoteDefinedTerm',
'ContinuedSectionSubsection',
'Oath'] for c in sib['class']):
content += sib.text
elif sib['class'][0].startswith('indent'):
content += sib.text
elif sib['class'][0] == 'HistoricalNote':
sec.meta = utf8(sib.text)
elif sib['class'][0] in ['PITLink',
'nif']:
pass
else:
log.info('Unhandled case in parsing section %s/%s' % (act.cid, secid))
log.debug(sib.name)
log.debug(sib.attrs)
if not sib or not sib.nextSibling:
stop = True
if not stop:
sib = sib.nextSibling
sec.text = Text.make(content)
sec.stored = now()
schedules = soup.select('div [class=Schedule]')
post = ''
for sched in schedules:
post += sched.text
act.post = Text.make(post)
act.stored = now()
else:
#alternative section method
#for this method we switch to the XML version and pull identifying information
#out of the code = attribute. Annecdotally, this seems to need to be done for
#very small acts
log.info('Switching to alternate section method')
soup = self.getActXMLSoup(act.cid)
sections = soup.select("section[code^=se=]")
for section in sections:
try:
secid = section['code'].replace('=', '-').replace('"', '')
pre = ''
pre = section.label.text + ' ' if section.label else pre
pre = pre + section.marginalnote.text if section.marginalnote else pre
text = section.select_one('text').text
except:
log.warn('ERROR in alternate parsing method for {}.{}'.format(act.cid, secid))
raise
if 'repealed' in text.lower():
pass
else:
sec = self.findOrCreateSection(act.released, secid, act)
sec.setPre(pre)
sec.setText(text)
sec.parent = act
sec.depth = 2
sec.released = act.released
sec.rev = act.rev
sec.cid = secid
act.analyze()
store.commit()
return act
