def __init__(self, desc_len, summ_len):
self.desc_length = desc_len
self.summ_length = summ_len
self.t_processor = Text_Process()
self.t_creator = TopicManager()
self.tl_creator = TopicLinkManager()
self.rd_creator = TopicRedirectManager()
def __init__(self, desc_len, summ_len):
self.desc_length = desc_len
self.summ_length = summ_len
self.t_processor = Text_Process()
self.t_creator = TopicManager()
self.tl_creator = TopicLinkManager()
self.rd_creator = TopicRedirectManager()
def __init__(self, desc_len, summ_len):
self.desc_length = desc_len
self.summ_length = summ_len
self.t_processor = Text_Process(self.desc_length, self.summ_length)
self.wk_mgt = WikipediaHelper(self.desc_length, self.summ_length)
self.t_creator = TopicManager()
self.tl_creator = TopicLinkManager()
class WikiAPIUtils():
@classmethod
def __init__(self, desc_len, summ_len):
self.desc_length = desc_len
self.summ_length = summ_len
self.t_processor = Text_Process()
self.t_creator = TopicManager()
self.tl_creator = TopicLinkManager()
self.rd_creator = TopicRedirectManager()
# util functions
def get_rand_score(self, source, target):
#TODO: implement algorithm for topic relatedness
i = 0
stop = random.randrange(20, 99)
for i in range(stop):
i /= 100.0
return i # randomly generated score
def fetch_relevant_topics(self, title_list, num_res):
res = []
while (len(res) < num_res):
max_index = len(title_list)
index = random.randrange(0, max_index)
title = title_list[index]
if not title in res:
res.append(title)
title_list.remove(title)
title_list.extend(res)
return res
def parse_linked_pages(self, source_id, pages):
if pages != None:
pattern = re.compile(r'\d\$,')
titles = []
links = pages[str(source_id)]["links"]
if links != None:
for lk in links:
lk_title = lk["title"]
# if re.match(r'^\w+$', lk_title):
if re.match(r'^[a-zA-Z-_() ]+$', lk_title):
clean_title = self.clean_topic_title(lk_title)
if clean_title is None or not clean_title: # prevent null and empty titles
continue
titles.append(
clean_title) # get title of linked article
return titles
def parse_and_save_topic(self, pages, redirects):
topics = []
for pid in pages:
# TODO: trim content length from api call
topic = None
try:
topic = Topic.objects.get(article_id=pid)
except Topic.DoesNotExist:
pcontent = pages[pid]
title = pcontent['title']
content = pcontent['extract'].encode(
'utf-8') # TODO: fix this to make bug proof
parsedtext = self.t_processor.get_desc_summ(content)
data = {
"article_id":
int(pid),
"article_title":
title,
"description":
parsedtext[self.t_processor.desc_text],
"summary":
parsedtext[self.t_processor.summ_text],
"wiki_link":
WIKI_URL_BASE.format(self.clean_topic_title(title)),
"linked_topics": []
}
topic = self.t_creator.create_topic(data)
# continue if there are no page redirects
if redirects != None or bool(redirects) != False:
# print "\n\nREDIRECTS -->", redirects
for rd in redirects:
rdfrom = rd['from']
rdto = rd['to']
if title == rdto:
topic_rd = self.create_topic_redirect(
rdfrom, topic)
topic_rd.save()
redirects.remove(rd)
topic.save()
if topic != None:
topics.append(topic)
return topics
def create_topic_redirect(self, source_rd_title, target_rd_topic):
source_rd_id = self.convert_titles_to_ids([source_rd_title])[0]
data = {"source_id": source_rd_id, "redirect_topic": target_rd_topic}
topic_rd = self.rd_creator.create_topic_redirect(data)
return topic_rd
def request_api_pages_plain(self, url):
print "\nREQUEST URL --->\n", url
resp = requests.get(url)
if resp.status_code != 200:
# TODO: handle exception better
raise requests.HTTPError(
'GET: Wikipedia api request error\n{}'.format(
resp.status_code))
return None
# TODO: catch Exception for "No JSON object could be decoded"
json_response = resp.json()
pages = json_response['query']['pages']
return pages
def make_api_request(self, r_url, cont_flag, retrieve_flag):
all_pages = {}
c_flag = "continue"
ct_flag = self.format_flag(cont_flag)
next_flag = ""
is_first = True
is_continue = False
redirects = None
while (is_continue == True or is_first == True):
url = r_url
if is_continue == True:
url = url + ct_flag.format(next_flag)
print "\nREQUEST URL --->\n", url
resp = requests.get(url)
if resp.status_code != 200:
# TODO: handle exception better
raise requests.HTTPError('GET: api request error\n{}'.format(
resp.status_code))
return None
# TODO: catch Exception for "No JSON object could be decoded"
json_response = resp.json()
query = json_response['query']
pages = query['pages']
if 'redirects' in query:
redirects = query['redirects']
curr_pid = -1
for key, val in pages.iteritems():
if retrieve_flag in val:
curr_pid = key
break # TODO: break here or continue running?
# should always be true
if curr_pid > -1:
if curr_pid in all_pages:
all_pages[curr_pid][retrieve_flag].extend(
pages[curr_pid][retrieve_flag])
else:
all_pages.update({curr_pid: pages[curr_pid]})
if c_flag in json_response:
is_continue = True
next_flag = json_response[c_flag][cont_flag]
else:
is_continue = False
is_first = False
results = {PAGE_RESP_KEY: all_pages}
if redirects != None:
results.update({REDIRECT_RESP_KEY: redirects})
return results
def convert_titles_to_ids(self, titles):
resources = self.format_req(titles)
url = PAGE_INFO_URL.format(resources)
pages = self.request_api_pages_plain(url)
if pages == None:
# TODO: error handling
return []
ids = []
for pgval in pages.values():
tid = pgval['pageid']
ids.append(tid)
return ids
def convert_ids_to_titles(self, ids):
resources = self.format_req(ids)
url = PAGE_INFO_URL.format(resources)
pages = self.request_pages_plain(url)
if pages == None:
# TODO: error handling
return []
titles = []
for pgval in pages.values():
clean_title = self.clean_topic_title(pgval['title'])
titles.append(clean_title)
return titles
def clean_topic_title(self, title):
# capitalize = re.sub(r'\w+', lambda m:m.group(0).capitalize(), title)
clean_title = title.replace(" ", TITLE_SEP)
return clean_title
def is_pageid(self, arg):
try:
pid = int(arg)
return True
except ValueError:
return False
def is_seq(self, arg):
return (not hasattr(arg, "strip") and hasattr(arg, "__getitem__")
or hasattr(arg, "__iter__"))
def count_articles(self, r_args):
if self.is_seq(r_args):
return len(r_args)
# default
return 1
def format_req(self, r_args):
if self.is_seq(r_args): # check if arg is a list
if self.is_pageid(r_args[0]):
return R_PAGEID.format(ARG_SEP.join(r_args))
else:
return R_TITLE.format(ARG_SEP.join(r_args))
if self.is_pageid(r_args):
return R_PAGEID.format(r_args)
# default return single page title format
return R_TITLE.format(r_args)
def format_flag(self, flag):
newflag = "&{}="
return newflag.format(flag) + "{}"
class WikiAPIUtils():
@classmethod
def __init__(self, desc_len, summ_len):
self.desc_length = desc_len
self.summ_length = summ_len
self.t_processor = Text_Process()
self.t_creator = TopicManager()
self.tl_creator = TopicLinkManager()
self.rd_creator = TopicRedirectManager()
# util functions
def get_rand_score(self, source, target):
#TODO: implement algorithm for topic relatedness
i = 0
stop = random.randrange(20, 99)
for i in range(stop):
i /= 100.0
return i # randomly generated score
def fetch_relevant_topics(self, title_list, num_res):
res = []
while (len(res) < num_res):
max_index = len(title_list)
index = random.randrange(0, max_index)
title = title_list[index]
if not title in res:
res.append(title)
title_list.remove(title)
title_list.extend(res)
return res
def parse_linked_pages(self, source_id, pages):
if pages != None:
pattern = re.compile(r'\d\$,')
titles = []
links = pages[str(source_id)]["links"]
if links != None:
for lk in links:
lk_title = lk["title"]
# if re.match(r'^\w+$', lk_title):
if re.match(r'^[a-zA-Z-_() ]+$', lk_title):
clean_title = self.clean_topic_title(lk_title)
if clean_title is None or not clean_title: # prevent null and empty titles
continue
titles.append(clean_title) # get title of linked article
return titles
def parse_and_save_topic(self, pages, redirects):
topics = []
for pid in pages:
# TODO: trim content length from api call
topic = None
try:
topic = Topic.objects.get(article_id=pid)
except Topic.DoesNotExist:
pcontent = pages[pid]
title = pcontent['title']
content = pcontent['extract'].encode('utf-8') # TODO: fix this to make bug proof
parsedtext = self.t_processor.get_desc_summ(content)
data = {"article_id" : int(pid),
"article_title" : title,
"description" : parsedtext[self.t_processor.desc_text],
"summary" : parsedtext[self.t_processor.summ_text],
"wiki_link" : WIKI_URL_BASE.format(self.clean_topic_title(title)),
"linked_topics" : []
}
topic = self.t_creator.create_topic(data)
# continue if there are no page redirects
if redirects != None or bool(redirects) != False:
# print "\n\nREDIRECTS -->", redirects
for rd in redirects:
rdfrom = rd['from']
rdto = rd['to']
if title == rdto:
topic_rd = self.create_topic_redirect(rdfrom, topic)
topic_rd.save()
redirects.remove(rd)
topic.save()
if topic != None:
topics.append(topic)
return topics
def create_topic_redirect(self, source_rd_title, target_rd_topic):
source_rd_id = self.convert_titles_to_ids([source_rd_title])[0]
data = {"source_id" : source_rd_id, "redirect_topic" : target_rd_topic}
topic_rd = self.rd_creator.create_topic_redirect(data)
return topic_rd
def request_api_pages_plain(self, url):
print "\nREQUEST URL --->\n", url
resp = requests.get(url)
if resp.status_code != 200:
# TODO: handle exception better
raise requests.HTTPError('GET: Wikipedia api request error\n{}'.format(resp.status_code))
return None
# TODO: catch Exception for "No JSON object could be decoded"
json_response = resp.json()
pages = json_response['query']['pages']
return pages
def make_api_request(self, r_url, cont_flag, retrieve_flag):
all_pages = {}
c_flag = "continue"
ct_flag = self.format_flag(cont_flag)
next_flag = ""
is_first = True
is_continue = False
redirects = None
while (is_continue == True or is_first == True):
url = r_url
if is_continue == True:
url = url + ct_flag.format(next_flag)
print "\nREQUEST URL --->\n", url
resp = requests.get(url)
if resp.status_code != 200:
# TODO: handle exception better
raise requests.HTTPError('GET: api request error\n{}'.format(resp.status_code))
return None
# TODO: catch Exception for "No JSON object could be decoded"
json_response = resp.json()
query = json_response['query']
pages = query['pages']
if 'redirects' in query:
redirects = query['redirects']
curr_pid = -1
for key,val in pages.iteritems():
if retrieve_flag in val:
curr_pid = key
break # TODO: break here or continue running?
# should always be true
if curr_pid > -1:
if curr_pid in all_pages:
all_pages[curr_pid][retrieve_flag].extend(pages[curr_pid][retrieve_flag])
else:
all_pages.update({curr_pid : pages[curr_pid]})
if c_flag in json_response:
is_continue = True
next_flag = json_response[c_flag][cont_flag]
else:
is_continue = False
is_first = False
results = {PAGE_RESP_KEY : all_pages}
if redirects != None:
results.update({REDIRECT_RESP_KEY : redirects})
return results
def convert_titles_to_ids(self, titles):
resources = self.format_req(titles)
url = PAGE_INFO_URL.format(resources)
pages = self.request_api_pages_plain(url)
if pages == None:
# TODO: error handling
return []
ids = []
for pgval in pages.values():
tid = pgval['pageid']
ids.append(tid)
return ids
def convert_ids_to_titles(self, ids):
resources = self.format_req(ids)
url = PAGE_INFO_URL.format(resources)
pages = self.request_pages_plain(url)
if pages == None:
# TODO: error handling
return []
titles = []
for pgval in pages.values():
clean_title = self.clean_topic_title(pgval['title'])
titles.append(clean_title)
return titles
def clean_topic_title(self, title):
# capitalize = re.sub(r'\w+', lambda m:m.group(0).capitalize(), title)
clean_title = title.replace(" ", TITLE_SEP)
return clean_title
def is_pageid(self, arg):
try:
pid = int(arg)
return True;
except ValueError:
return False;
def is_seq(self, arg):
return (not hasattr(arg, "strip") and
hasattr(arg, "__getitem__") or
hasattr(arg, "__iter__"))
def count_articles(self, r_args):
if self.is_seq(r_args):
return len(r_args)
# default
return 1
def format_req(self, r_args):
if self.is_seq(r_args): # check if arg is a list
if self.is_pageid(r_args[0]):
return R_PAGEID.format(ARG_SEP.join(r_args))
else:
return R_TITLE.format(ARG_SEP.join(r_args))
if self.is_pageid(r_args):
return R_PAGEID.format(r_args)
# default return single page title format
return R_TITLE.format(r_args)
def format_flag(self, flag):
newflag = "&{}="
return newflag.format(flag) + "{}"