def _handle_page(self, title, text):
self.n_pages += 1
if not self.filter_page_raw(title, text):
return
if '==' in text:
text = text.split('==')[0]
text = self.braces_pattern.sub('', text)
text = self.wikiref_pattern.sub(lambda x: x.groups(0)[1], text)
text = self.wikiref2_pattern.sub('', text)
text = self.extraref_pattern.sub(lambda x: x.groups(0)[1], text)
text = self.extraref2_pattern.sub('', text)
text = self.xml_tag_pattern.sub('', text)
text = self.paren_fixup_pattern.sub('(', text)
text = self.space_fixup_pattern.sub(' ', text)
text = self.entity_fixup_pattern.sub(lambda x:
{ 'mdash': '-', 'ndash': '-', 'quot': '"',
'amp': '&', 'lt': '<', 'gt': '>' }[x.groups()[0]], text)
text = text.strip()
self.handle_page(title, text)
def process_gh_project_names(text):
if '-' in text:
return ' '.join(text.split('-'))
elif text[0].isupper():
return ''.join(
map(lambda x: x
if x.islower() else " " + x, text[0].lower() + text[1:]))
else:
return ''.join(map(lambda x: x if x.islower() else " " + x, text))
def Tokenize(text):
tokenizer = MWETokenizer(category.all())
for word in category:
if word.find(' '):
tokenizer.add_mwe(word.split())
for word in sub_category:
if word.find(' '):
tokenizer.add_mwe(word.split())
for word in brand:
if word.find(' '):
tokenizer.add_mwe(word.split())
for word in article:
if word.find(' '):
tokenizer.add_mwe(word.split())
token = tokenizer.tokenize(text.split())
tokens = []
for word in token:
word = word.replace("_", " ")
tokens.append(word)
return tokens
def build_tsquery( text ):
text=scalar('select gazetteer.gaz_plainText2(:text)',text=text)
return ' & '.join(map( lambda x: re.sub(r'\*$',':*',x),text.split()))
def preprocess_search_query(text):
text = re.sub(r'\W', ' ', text, re.UNICODE)
tokens = text.split()
tokens = map(lambda t: t.strip(), tokens)
return ' & '.join(['%s:*' % t for t in tokens])
def build_tsquery(text):
text = scalar('select gazetteer.gaz_plainText2(:text)', text=text)
return ' & '.join(map(lambda x: re.sub(r'\*$', ':*', x), text.split()))
def handle_app_mention(event_data):
message = event_data["event"]
# Get the user who initiated the @mention
source_user = slack_client.users_info(user=message["user"])
# Default variable values
slack_message, msg, mentioned_user, attachments = "", "", None, None
notify_user = source_user["user"]["id"]
# Iterate through message body looking for the destination user. Use the first match.
elements = message["blocks"][0]["elements"][0]["elements"]
for element in elements:
if element["type"] == "user" and element["user_id"] != SLACKBOT_USERID:
mentioned_user = slack_client.users_info(user=element["user_id"])
break
# If there was a user mention, handle the points.
if mentioned_user is not None:
if source_user["user"]["id"] == mentioned_user["user"]["id"]:
msg = "Hey <@" + source_user["user"][
"id"] + "> - nice try but you can only give points to others"
else:
# Try to get the number of points, default to 1.
text = message.get('text')
points_array = [int(s) for s in text.split() if s.isdigit()]
points = 1 if len(points_array) == 0 else points_array[0]
source_user_email = source_user["user"]["profile"]["email"]
mentioned_user_email = mentioned_user["user"]["profile"]["email"]
original_message = slack_client.chat_getPermalink(
channel=message["channel"], message_ts=message["ts"])
# Try to grant points
if points > 10:
msg = """Hey <@{source}> :wave: I couldn't do <{permalink}|this>.
You can give a maximum of 10 points at a time! You should try giving <@{mentioned}> some points again (but only up to 10, remember?)!""".format(
source=source_user["user"]["id"],
permalink=original_message["permalink"],
mentioned=mentioned_user["user"]["id"])
else:
error = try_grant_points(source_user_email,
mentioned_user_email, points)
if error is None:
msg = """Hey <@{mentioned}> :wave: <{permalink}|you got {points} points> from <@{source}>!
⚡ Check out the leaderboard <https://snitch-leaderboard.herokuapp.com/|here>! ⚡
⚡ And get to snitching! ⚡""".format(
mentioned=mentioned_user["user"]["id"],
points=str(points),
source=source_user["user"]["id"],
permalink=original_message["permalink"])
notify_user = mentioned_user["user"]["id"]
#React to the slack post now, for some sense of transparency
try:
slack_client.reactions_add(
channel=message["channel"],
timestamp=message["event_ts"],
name="thumbsup")
except:
print(
"well that's too bad I couldn't react... try again with another emoji"
)
try:
slack_client.reactions_add(
channel=message["channel"],
timestamp=message["event_ts"],
name="white_check_mark")
except:
print("something went really really wrong")
pass
# If we couldn't grant points, let people know
else:
msg = "Hey <@" + source_user["user"][
"id"] + "> - I couldn't give <@" + mentioned_user["user"][
"id"] + "> points from you, here's what the computer told me: " + error
# If there was no user mention, handle the error.
else:
msg = "Hey <@" + source_user["user"][
"id"] + "> something funny just happened... can you try granting that point again?"
print("Sending %s the following alert: %s" % (notify_user, msg))
slack_client.chat_postMessage(channel=notify_user, text=msg)
def preprocess_search_query(text):
text = re.sub(r'\W', ' ', text, re.UNICODE)
tokens = text.split()
tokens = map(lambda t: t.strip(), tokens)
return ' & '.join(['%s:*' % t for t in tokens])
def build_tsquery(text):
text = scalar("select gazetteer.gaz_plainText2(:text)", text=text)
return " & ".join([re.sub(r"\*$", ":*", x) for x in text.split()])