def addMemento(self, urim):
try:
r = self.session.get(urim)
if len(r.history) == 0:
raw_urim = otmt.generate_raw_urim(urim)
else:
raw_urim = otmt.generate_raw_urim(r.url)
self.session.get(raw_urim)
self.urimlist.append(urim)
except (ConnectionError, TooManyRedirects, RequestException) as e:
self.addMementoError(urim, repr(e))
def generate_raw_urim_archiveorg_happy_path(self):
urim = "https://web.archive.org/web/20070207050545/http://www.cnn.com:80/"
raw_urim = "https://web.archive.org/web/20070207050545id_/http://www.cnn.com:80/"
self.assertEquals(raw_urim, generate_raw_urim(urim))
def generate_raw_urim_archiveit_happy_path(self):
urim = "http://wayback.archive-it.org/1068/20110317183254/http://www.amnistia.org.mx/"
raw_urim = "http://wayback.archive-it.org/1068/20110317183254id_/http://www.amnistia.org.mx/"
self.assertEquals(raw_urim, generate_raw_urim(urim))
def get_newspaper_publication_date(urim, cache_storage):
import otmt
dbconn = MongoClient(cache_storage)
session = get_web_session(cache_storage)
db = dbconn.get_default_database()
try:
return db.derivedvalues.find_one({"urim":
urim})["newspaper publication date"]
except (KeyError, TypeError):
raw_urim = otmt.generate_raw_urim(urim)
r = session.get(raw_urim)
r.raise_for_status()
article = Article(urim)
article.download(r.text)
article.parse()
article.nlp()
pd = article.publish_date
if pd is None:
pd = r.headers['memento-datetime']
else:
pd = pd.strftime("%a, %d %b %Y %H:%M:%S GMT")
db.derivedvalues.update(
{"urim": urim}, {"$set": {
"newspaper publication date": str(pd)
}})
return pd
def getMementoContent(self, urim):
"""Returns the HTTP entity of memento at `urim` provided that it
was previously stored via `addMemento`.
If no data was stored via `addMemento` for `urim`, then
`CollectionModelNoSuchMementoException` is thrown.
If data was stored via `addMementoError` for `urim`, then
`CollectionModelMementoErrorException` is thrown.
"""
raw_urim = otmt.generate_raw_urim(urim)
return self.session.get(raw_urim).text
def get_raw_simhash(urim, cache_storage):
import otmt
dbconn = MongoClient(cache_storage)
session = get_web_session(cache_storage)
db = dbconn.get_default_database()
# 1 if lang of urim in cache, return it
try:
return db.derivedvalues.find_one({"urim": urim})["raw simhash"]
except (KeyError, TypeError):
r = session.get(urim)
if len(r.history) == 0:
raw_urim = otmt.generate_raw_urim(urim)
else:
raw_urim = otmt.generate_raw_urim(r.url)
r2 = session.get(raw_urim)
r2.raise_for_status()
if 'text/html' not in r2.headers['content-type']:
raise Exception(
"Hypercane currently only operates with HTML resources, refusing to compute Simhash on {}"
.format(urim))
simhash = Simhash(r2.text).value
db.derivedvalues.update({"urim": urim},
{"$set": {
"raw simhash": str(simhash)
}},
upsert=True)
return str(simhash)
def synthesize_warc(urim, session, output_directory):
import otmt
import glob
from warcio.warcwriter import WARCWriter
from warcio.statusandheaders import StatusAndHeaders
from hashlib import md5
from datetime import datetime
import traceback
m = md5()
m.update(urim.encode('utf8'))
urlhash = m.hexdigest()
if len( glob.glob('{}/{}*.warc.gz'.format(output_directory, urlhash)) ) > 0:
module_logger.warning("Detected existing WARC for URI-M, skipping {}".format(urim))
return
resp = session.get(urim, stream=True)
resp.raise_for_status()
headers_list = resp.raw.headers.items()
# we use response.url instead of urim to (hopefully) avoid raw redirects
raw_urim = otmt.generate_raw_urim(resp.url)
raw_response = session.get(raw_urim, stream=True)
warc_target_uri = None
# we have to implement this construct in case the archive combines original with other relations
for link in resp.links:
if 'original' in link:
warc_target_uri = resp.links[link]['url']
if warc_target_uri is None:
module_logger.warning("could not find this memento's original resource, skipping {}".format(urim))
return
try:
mdt = resp.headers['Memento-Datetime']
except KeyError:
module_logger.warning("could not find this memento's memento-datetime, skipping {}".format(urim))
return
http_headers = StatusAndHeaders('200 OK',
headers_list, protocol='HTTP/1.0')
module_logger.debug("mdt formatted by strptime and converted by strftime: {}".format(
datetime.strptime(
mdt, "%a, %d %b %Y %H:%M:%S GMT"
).strftime('%Y-%m-%dT%H:%M:%SZ')
))
warc_headers_dict = {}
warc_headers_dict['WARC-Date'] = datetime.strptime(
mdt, "%a, %d %b %Y %H:%M:%S GMT"
).strftime('%Y-%m-%dT%H:%M:%SZ')
with open("{}/{}-{}.warc.gz".format(output_directory, urlhash, datetime.now().strftime('%Y%m%d%H%M%S')), 'wb') as output:
writer = WARCWriter(output, gzip=True)
record = writer.create_warc_record(
warc_target_uri, 'response',
payload=raw_response.raw,
http_headers=http_headers,
warc_headers_dict=warc_headers_dict
)
writer.write_record(record)
def addManyMementos(self, urims):
module_logger.info("started with {} URI-Ms for processing...".format(len(urims)))
# protect the function from duplicates in the urims list
urims = list(set(urims))
module_logger.info("found duplicates, now using {} URI-Ms for processing...".format(len(urims)))
futuressession = FuturesSession(session=self.session)
retry = Retry(
total=10,
read=10,
connect=10,
backoff_factor=0.3,
status_forcelist=(500, 502, 504)
)
adapter = HTTPAdapter(max_retries=retry)
futuressession.mount('http://', adapter)
futuressession.mount('https://', adapter)
futures = {}
raw_futures = {}
working_urim_list = []
raw_urims = []
for uri in urims:
# raw_urim = otmt.generate_raw_urim(uri)
working_urim_list.append(uri)
futures[uri] = futuressession.get(uri)
# futures[raw_urim] = futuressession.get(raw_urim)
working_starting_size = len(working_urim_list)
def uri_generator(urilist):
while len(urilist) > 0:
uchoice = random.choice(urilist)
yield uchoice
for uri in uri_generator(working_urim_list):
if futures[uri].done():
module_logger.debug("URI-M {} is done, processing...".format(uri))
if len(working_urim_list) % 100 == 0:
module_logger.info("{}/{} mementos left to process".format(len(working_urim_list), working_starting_size))
try:
r = futures[uri].result()
if len(r.history) == 0:
raw_urim = otmt.generate_raw_urim(uri)
else:
raw_urim = otmt.generate_raw_urim(r.url)
raw_urims.append( raw_urim )
if 'memento-datetime' not in r.headers:
self.addMementoError(uri, "URI-M {} does not produce a memento".format(uri))
else:
# the content should be cached by the session
# we just need to keep track of the URI-Ms for this run
self.urimlist.append(uri)
except Exception as e:
self.addMementoError(uri, repr(e))
working_urim_list.remove(uri)
del futures[uri]
module_logger.info("done adding {} mementos, now adding corresponding {} raw mementos...".format( len(urims), len(raw_urims) ))
working_raw_urim_list = []
for raw_urim in list(set(raw_urims)):
working_raw_urim_list.append(raw_urim)
raw_futures[raw_urim] = futuressession.get(raw_urim)
working_rawurims_starting_size = len(working_raw_urim_list)
# for raw_urim in uri_generator(working_raw_urim_list):
while len(working_raw_urim_list) > 0:
raw_urim = random.choice(working_raw_urim_list)
module_logger.debug("fetching results for raw URI-M {}".format(raw_urim))
# module_logger.debug("are the keys the same as the working list: {}".format( set(working_raw_urim_list) == set(list(raw_futures.keys())) ) )
module_logger.debug("raw mementos working list size: {}".format(len(working_raw_urim_list)))
module_logger.debug("raw mementos futures keys size: {}".format(len(raw_futures)))
# try:
# raw_futures[raw_urim]
# except KeyError:
# module_logger.error("{} is not in futures".format(raw_urim))
# module_logger.error("is it: {}".format( raw_urim in raw_futures ))
# module_logger.error("")
# module_logger.error("working list follows:")
# module_logger.error(pp.pformat(working_raw_urim_list))
# module_logger.error("")
# module_logger.error("raw_futures keys follows:")
# module_logger.error(pp.pformat(list(raw_futures.keys())))
if raw_futures[raw_urim].done():
module_logger.debug("raw URI-M {} is done, processing...".format(raw_urim))
if len(working_raw_urim_list) % 100 == 0:
module_logger.info("{}/{} raw mementos left to process".format(len(working_raw_urim_list), working_rawurims_starting_size))
try:
r = raw_futures[raw_urim].result()
if 'memento-datetime' not in r.headers:
self.addMementoError(uri, "raw URI-M {} does not produce a memento".format(raw_urim))
else:
# the content should be cached by the session
# we just need to keep track of the raw URI-Ms for this run
self.urimlist.append(raw_urim)
except Exception as e:
self.addMementoError(raw_urim, repr(e))
# module_logger.debug("removing {} from working raw URI-M list and raw futures keys".format(raw_urim))
working_raw_urim_list.remove(raw_urim)
del raw_futures[raw_urim]
# module_logger.debug("raw URI-M {} in working raw URI-M list still? {}".format( raw_urim, raw_urim in working_raw_urim_list ))
time.sleep(1)
def generate_raw_urim_archiveorg_raw_already(self):
raw_urim = "https://web.archive.org/web/20070207050545id_/http://www.cnn.com:80/"
self.assertEquals(raw_urim, generate_raw_urim(raw_urim))
def generate_raw_urim_archiveit_raw_already(self):
raw_urim = "http://wayback.archive-it.org/1068/20110317183254id_/http://www.amnistia.org.mx/"
self.assertEquals(raw_urim, generate_raw_urim(raw_urim))
def get_boilerplate_free_content(urim,
cache_storage="",
dbconn=None,
session=None):
import otmt
from boilerpy3 import extractors
if dbconn is None:
dbconn = MongoClient(cache_storage)
if session is None:
session = get_web_session(cache_storage)
db = dbconn.get_default_database()
# 1. if boilerplate free content in cache, return it
try:
module_logger.info(
"returing boilerplate free content from cache for {}".format(urim))
bpfree = db.derivedvalues.find_one({"urim":
urim})["boilerplate free content"]
return bytes(bpfree, "utf8")
except (KeyError, TypeError):
module_logger.info(
"generating boilerplate free content for {}".format(urim))
r = session.get(urim)
if len(r.history) == 0:
raw_urim = otmt.generate_raw_urim(urim)
else:
raw_urim = otmt.generate_raw_urim(r.url)
r2 = session.get(raw_urim)
r2.raise_for_status()
module_logger.info("content-type is {}".format(
r2.headers['content-type']))
if 'text/html' not in r2.headers['content-type']:
module_logger.warning(
"we can only remove boilerplate from HTML, returning zero bytes"
)
return bytes()
# paragraphs = justext(
# r.text, get_stoplist('English')
# )
# bpfree = ""
# for paragraph in paragraphs:
# bpfree += "{}\n".format(paragraph.text)
module_logger.debug(
"attempting to extract boilerplate free content from {}".format(
urim))
extractor = extractors.ArticleExtractor()
try:
bpfree = extractor.get_content(r2.text)
module_logger.info(
"storing boilerplate free content in cache {}".format(urim))
db.derivedvalues.update(
{"urim": urim}, {"$set": {
"boilerplate free content": bpfree
}},
upsert=True)
except Exception:
module_logger.exception(
"failed to extract boilerplate from {}, setting value to empty string"
.format(urim))
hypercane.errors.errorstore.add(urim, traceback.format_exc())
return bytes()
return bytes(bpfree, "utf8")
