def strip_unicode(text):
clean = ftfy.fix_text(ftfy.fix_encoding(text))
temp = ""
try:
temp = clean.decode('utf-8', 'ignore').encode('utf-8')
except UnicodeEncodeError or UnicodeEncodeError as e:
print "error1" + str(e)
try:
temp = text.encode('utf-8').decode('utf-8', 'ignore').encode('utf-8')
except UnicodeEncodeError or UnicodeDecodeError as e:
print "error2" + str(e)
return temp
def strip_unicode(text):
clean = ftfy.fix_text(ftfy.fix_encoding(text))
try:
temp = clean.decode('utf-8', 'ignore').encode('utf-8')
except UnicodeEncodeError or UnicodeEncodeError as e:
# print "ERROR ONE"
pass
try:
# print "first encoding failed, trying other way!"
temp = text.encode('utf-8').decode('utf-8', 'ignore').encode('utf-8')
except UnicodeEncodeError or UnicodeDecodeError as e:
# print "ERROR TWO"
pass
return temp
def fetch():
all_prices, timestamps = {}, []
for i, url in enumerate(configuration.get('cardhoarder_urls')):
s = fetcher_internal.fetch(url)
s = ftfy.fix_encoding(s)
timestamps.append(dtutil.parse_to_ts(s.split('\n', 1)[0].replace('UPDATED ', ''), '%Y-%m-%dT%H:%M:%S+00:00', dtutil.CARDHOARDER_TZ))
all_prices[i] = parse_cardhoarder_prices(s)
url = configuration.get('mtgotraders_url')
if url:
s = fetcher_internal.fetch(configuration.get('mtgotraders_url'))
timestamps.append(dtutil.dt2ts(dtutil.now()))
all_prices['mtgotraders'] = parse_mtgotraders_prices(s)
if not timestamps:
raise TooFewItemsException('Did not get any prices when fetching {urls} ({all_prices})'.format(urls=configuration.get('cardhoarder_urls') + [configuration.get('mtgotraders_url')], all_prices=all_prices))
store(min(timestamps), all_prices)
def tokenizer_word(data_dir, category, save_dir, sw_dir):
list_fn = os.listdir(os.path.join(data_dir, category))
file_train = os.path.join(save_dir, "Train", "{}.txt".format(category))
file_test = os.path.join(save_dir, "Test", "{}.txt".format(category))
# lay stopwords tu file
with open(sw_dir, "r") as f:
stopwords = f.readlines()
for i in range(len(stopwords)):
stopwords[i] = stopwords[i].strip()
# tao train file, test file theo chu de
with open(file_train, "w", encoding="utf8"):
pass
with open(file_test, "w", encoding="utf8"):
pass
# chia train, test (60/40)
mid = int(len(list_fn) * 0.6 + 0.5)
for i, fn in enumerate(list_fn):
fl = os.path.join(data_dir, category, fn)
with open(fl, "r") as f:
content = f.readlines()
content = " ".join(content)
content = content.replace("\n", " ")
content = fix_encoding(content)
content = ViTokenizer.tokenize(content)
content = gensim.utils.simple_preprocess(content)
content = [w for w in content if not w in stopwords]
content = " ".join(content)
if len(content) < 100:
continue
if i < mid:
with open(file_train, "a", encoding="utf8") as f:
f.write(content)
f.write("\n")
elif i == mid:
with open(file_train, "a", encoding="utf8") as f:
f.write(content)
else:
if i == (len(list_fn) - 1):
with open(file_test, "a", encoding="utf8") as f:
f.write(content)
else:
with open(file_test, "a", encoding="utf8") as f:
f.write(content)
f.write("\n")
def get_results_from_lines(lines, locale_template):
# lines = []
all_results = []
to_check_items = list(locale_template.items())
already_check_items = list()
value = {}
prev_field = None
for line in lines:
line = ftfy.fix_encoding(line)
found_value, value_new_entries = parse_line(line, to_check_items)
if found_value:
field, expression = found_value
if field == "matrix_sub_value":
value_new_entries = {"{}{}".format(prev_field, k): v for k, v in value_new_entries.items()}
value.update(value_new_entries)
prev_field = field
if field == "end_of_file":
if value:
result = seal_result(value, to_check_items)
all_results.append(result)
to_check_items = reset_check_items(locale_template)
value = {}
if field in permanent_matches:
continue
already_check_items.append(found_value)
to_check_items.remove(found_value)
if len(to_check_items) <= len(permanent_matches):
# I don't need to check anyother field
# print("No other field to check for filename {}", filename)
all_results.append(seal_result(value, to_check_items))
value = {}
else:
# potentially the line is unwanted so we need to reset the items and add results
found_value, value_new_entries = parse_line(line, already_check_items)
if found_value:
# Something wrong in the format of the file
# try create a new value starting from this line
all_results.append(seal_result(value, to_check_items))
value = value_new_entries
to_check_items = reset_check_items(locale_template)
to_check_items.remove(found_value)
# This is the end of the file
return all_results
def fetch_content(request_url, extractor):
meta = {}
# fetch using python
url, response = fetch_url(request_url)
if response is not None:
# extract content
el, txt, meta = extractor.extract(response, url)
if txt is not None and txt != '':
txt = fix_encoding(txt)
return url, response, txt, meta
# give up
return url if url else request_url, None, None, meta
def _sanitize_story_line(line):
line = ftfy.fix_encoding(line)
sentence_endings = [
".", "!", "?", "...", "'", "`", '"', ")", "\u2019", "\u201d"
]
# CNN stories always start with "(CNN)"
if line.startswith("(CNN)"):
line = line[len("(CNN)"):]
# Highlight are essentially bullet points and don't have proper sentence endings
if line[-1] not in sentence_endings:
line += "."
return line
def getDatasetList(self):
response = requests.get(self.reqUrl)
linkList = BeautifulSoup(response.text, "html.parser").find_all('a')
fixedList = list()
datasetList = list()
for link in linkList:
fixedList.append(fix_encoding(link.get('href')))
for item in fixedList:
if item.find('Bases') != -1:
fixedItem = re.sub('Bases/|.xlsx', '', item)
datasetList.append(fixedItem)
response.close()
return datasetList
def _fetch_chart(self):
req = urllib.request.Request(self.url)
page = urllib.request.urlopen(req, data=None, timeout=15)
root = lxml.html.parse(page)
rank = 1
entry = None
for element in root.iter(tag=lxml.etree.Element):
cls = element.get('class')
if cls is None:
continue
if cls == 'wrap_rank':
entry = ChartEntry()
entry.rank = rank
if rank > self.limit:
break
else:
self.append(entry)
rank += 1
entry.change = element[0].get('class').replace(
'icon_', '').replace('static',
'none').replace('rank_', '')
if entry.change is not 'new' and len(element) >= 2:
entry.change_diff = element[1].text_content().strip()
if cls == 'wrap_song_info' and entry is not None:
next = False
for a in element.iter(tag='a'):
if not next:
entry.title = ftfy.fix_encoding(
a.text_content().strip())
next = True
else:
for artist in a.text_content().split('|')[0].replace(
' & ', ',').split(','):
entry.artists.append(Artist(artist.strip()))
break
def get_articles(id_site, interval_start, interval_stop):
query = "SELECT id_article, xmltext FROM Article WHERE id_site = {} AND stimestamp between (1<<31)-unix_timestamp()+3600*24*{} AND (1<<31)-unix_timestamp()+3600*24*{}".format(
id_site, interval_stop, interval_start)
connection = pymysql.connect(host=CONNECTION_DETAILS["host"],
port=CONNECTION_DETAILS["port"],
user=CONNECTION_DETAILS["user"],
passwd=CONNECTION_DETAILS["passwd"],
db=CONNECTION_DETAILS["db"],
charset=CONNECTION_DETAILS["charset"])
cur = connection.cursor(pymysql.cursors.DictCursor)
cur.execute(query)
list_of_articles = []
for line in cur.fetchall():
text = ftfy.fix_encoding(line["xmltext"])
plain_text = remove_html_tags(text)
out = plain_text.translate(str.maketrans("", "", string.punctuation))
list_of_articles.append(out)
return list_of_articles
def clean_text(x):
'''General purpose text cleaner that reads in a string, applies Unicode sandwhich, strips special characters, and returns a list of tokens.'''
#Fix encoding in case there is a problem
x = ftfy.fix_encoding(x)
#Unicode Sandwhich...
x = x.encode('UTF-8') #encode as UTF-8
input_code = chardet.detect(bytes(x))['encoding'] #Auto-detect encoding
u_string = x.decode(input_code) #Decode
re.sub(r'\w{4}', u'xxxx', u_string,
flags=re.UNICODE) #Handle funny unicode artifacts
x = u_string.encode('UTF-8') #Encode back to UTF-8
x = str(x) #Convert to string
x = re.sub(r'[\[\]\"\'\,]', ' ', x) #Remove backslashes
x = re.sub(r'[^\w\s]', ' ', x) #Insert space after end of each sentence
x = x.split() #split string to list
return x
def chat():
print("Enter your question and (quit) to stop!")
while True:
inp = input("You: ")
if inp.lower() == "quit":
break
results = model.predict([bag_of_words(inp, words)])
#print(bag_of_words(inp, words))
results_index = numpy.argmax(results)
#print(results)
tag = labels[results_index]
for tg in data["intents"]:
if tg['tag'] == tag:
responses = tg["responses"]
print(fix_encoding(random.choice(responses)))
break
def _fetch_chart(self):
req = urllib.request.Request(self.url)
page = urllib.request.urlopen(req, data=None, timeout=15)
root = lxml.html.parse(page)
rank = 1
entry = None
for element in root.iter(tag=lxml.etree.Element):
cls = element.get('class')
if cls is None:
continue
if cls == 'ranking':
entry = ChartEntry()
entry.rank = element.text_content().strip()
if rank > self.limit:
break
else:
self.append(entry)
rank += 1
if cls == 'change':
change = element[0].get('class')
change = change if change else 'none'
entry.change = change
change_diff = element.text_content().strip()
if (change == 'up'
or change == 'down') and change_diff != 'HOT':
entry.change_diff = change_diff
if cls == 'subject':
entry.title = ftfy.fix_encoding(
element[0].text_content().strip())
for artist in element[1].text_content().split('|')[0].replace(
' & ', ',').split(','):
entry.artists.append(Artist(artist.strip()))
def get_in4_from_positon(i):
data = pd.read_csv("static/data_img/data.csv")
name_product = list(data['name_product'])
link = list(data['link'])
price = list(data['price'])
try:
fix_name = fix_encoding(name_product[i])
name_product[i] = fix_name
except:
None
return name_product[i], link[i], str(f'{int(price[i]):n}') + ' Đồng'
# i = get_locate_image('static/data_img/image/1588_00002.jpg')
# data = pd.read_csv("static/data_img/data.csv")
# name_img = list(data['name_img'])
# a, b, c = get_in4_from_positon(0)
# print(a, b, c)
def run() -> None:
files = rotation.files()
n = len(files)
time_until = TIME_UNTIL_ROTATION - datetime.timedelta(weeks=1)
if n >= rotation.TOTAL_RUNS:
print(
'It is the moment of discovery, the triumph of the mind, and the end of this rotation.'
)
return
if n == 0 and TIME_UNTIL_ROTATION > datetime.timedelta(7):
print(
'The monks of the North Tree rarely saw their kodama until the rotation, when it woke like a slumbering, angry bear.'
)
print('ETA: {t}'.format(
t=dtutil.display_time(int(time_until.total_seconds()))))
return
if n == 0:
rotation.clear_redis(clear_files=True)
#else:
# rotation.clear_redis()
all_prices = {}
for url in configuration.get_list('cardhoarder_urls'):
s = fetch_tools.fetch(url)
s = ftfy.fix_encoding(s)
all_prices[url] = parse_cardhoarder_prices(s)
url = configuration.get_str('mtgotraders_url')
if url:
s = fetch_tools.fetch(url)
all_prices['mtgotraders'] = parse_mtgotraders_prices(s)
run_number = process(all_prices)
if run_number == rotation.TOTAL_RUNS:
make_final_list()
try:
url = f'{fetcher.decksite_url()}/api/rotation/clear_cache'
fetch_tools.fetch(url)
except Exception as c: # pylint: disable=broad-except
print(c)
def processPage(page):
title = page['title']
if title not in page_uuids:
return
uuid = page_uuids[title]
children = []
if 'children' in page.keys():
for child in page['children']:
children.append({
'html':
applyHeading(renderMarkdown(fix_encoding(child['string'])),
child) + renderBullets(child)
})
template_data = {
'title': renderMarkdown(title, ignoreLinks=True),
'blocks': children,
'uuid': uuid,
'references': []
}
global _linksTo
for item in _linksTo:
item['link_from'] = uuid
item['title'] = renderMarkdown(title, ignoreLinks=True)
item['text'] = renderMarkdown(item['text'], ignoreLinks=True)
#if item['uuid'] == uuid:
# continue
if item['link_to'] in references.keys():
references[item['link_to']].append(item)
else:
references[item['link_to']] = [item]
_linksTo = []
page_data[title] = template_data
def fix_characters_in_string(text):
"""
Removes control characters such as \r\n \x1b \ufffd from string impl and returns a unicode
string where all control characters have been replaced by a space.
:param text: expects a unicode string
:return: unicode string
"""
# deal with encoding
new_text = fix_encoding(text)
# remove unicode characters from "Specials" block
# see: https://www.compart.com/en/unicode/block/U+FFF0
new_text = re.sub(r"\\ufff.", " ", new_text.encode("unicode-escape"))
# remove all kinds of control characters and emojis
# see: https://www.fileformat.info/info/unicode/category/index.htm
new_text = u"".join(ch if unicodedata.category(ch)[0] != "C" else " " for ch in new_text.decode("unicode-escape"))
return new_text
def ocr(filename):
"""
This function will handle the core OCR processing of images.
"""
i = cv2.imread(filename)
# Convert to gray
i = cv2.cvtColor(i, cv2.COLOR_BGR2GRAY)
# Apply dilation and erosion to remove some noise
kernel = np.ones((1, 1), np.uint8)
i = cv2.dilate(i, kernel, iterations=1)
i = cv2.erode(i, kernel, iterations=1)
text = pytesseract.image_to_string(i)
#return text
# Cleaning all the gibberish text
text = ftfy.fix_text(text)
text = ftfy.fix_encoding(text)
return text
def process_target_tables(tar_tables, doc):
# only the first 3 tables are of interest
for x in range(0, 3):
tablenum = x
tr_list = tar_tables[x].find_all("tr")
print("--- Row list length :" + str(len(tr_list)))
# time.sleep(1)
# tr_list[0] is just the column headers... can be ignored
for y in range(1, len(tr_list)):
curr_rowdata = tr_list[y].find_all("td")
print(curr_rowdata)
print("^^^^^^^^^")
tar_info = process_row_info(curr_rowdata, tablenum)
# exceptions for wonky seasonal table
# if any of these trigger, it's an empty data thing to be ignored
try:
if (tar_info[1].strip()) == "":
pass
# print("Detected nothing at idx 1!")
elif (tar_info[2].strip()) == "":
pass
# print("Detected nothing at idx 2!")
elif (tar_info[3].strip()) == "":
pass
# print("Detected nothing at idx 3!")
else:
for z in tar_info:
try:
purified_line = ftfy.fix_encoding(z)
fixed_line = ftfy.fix_text(purified_line)
doc.add_paragraph(fixed_line)
# print(x.encode("ascii", "replace"))
# run printing operation here
except:
print("Check for line info processing error...")
time.sleep(1)
except:
pass
print("[+] Page processing complete! ---")
def clean_string(s):
s = str(s)
if isnull(s):
return None
elif re.search('[a-zA-Z]', s) is None:
return None
else:
s = remove_bom(s)
s = remove_control_chars(s)
s = fix_encoding(s)
s = fix_text(s)
s = fix_partial_utf8_punct_in_1252(s)
s = decode_escapes(s)
s = fix_latin_ligatures(s)
s = uncurl_quotes(s)
s = s.replace("Äu0087", "ć")
s = s.replace("Äu0090", "Đ")
s = s.replace("Ãu0096", "Ö")
s = s.replace("Åu008D", "ō")
s = s.replace("\\", " ")
s = s.replace("/", " ")
s = s.replace("ö", "ö")
p = re.compile("^\w+[A-Z]{1}\w*$")
if p.search(s):
# From: https://stackoverflow.com/a/37697078
s = re.sub('(?!^)([A-Z][a-z]+)', r'\1', s)
new_string = ""
p = False
for letter in s:
if letter in "([":
p = True
elif letter in ")]":
p = False
continue
if not p:
new_string += letter
return new_string.strip()
def clean_data(x,
rem_stopwords=True,
ner=True,
rem_punc=True,
lemmatize=False):
'''Text cleaner that applies preprocessing rules to text. Preprocessing options include, ner, stop word removal, lemmatization, html cleaning, and punctuation removal.'''
#Unicode Sandwhich
#Fix encoding in case there is a problem
x = ftfy.fix_encoding(x)
#Unicode Sandwhich...
x = x.encode('UTF-8') #encode as UTF-8
input_code = chardet.detect(bytes(x))['encoding'] #Auto-detect encoding
try:
u_string = x.decode(input_code) #Decode
except:
u_string = x.decode('UTF-8')
re.sub(r'\w{4}', u'xxxx', u_string,
flags=re.UNICODE) #Handle funny unicode artifacts
x = u_string.encode('UTF-8') #Encode back to UTF-8
x_clean = str(x) #Convert to string
if ner == True:
x_clean = capitalize_named_entities(x_clean)
if rem_stopwords == True:
x_clean = remove_stopwords(x_clean)
if lemmatize == True:
x_clean = lemmatize_text(x_clean)
soup = BeautifulSoup(x_clean, 'html.parser')
x_clean = soup.text
if rem_punc == True:
x_clean = remove_punctuation(x_clean)
#tokens = word_tokenize(x_clrunean)
return x_clean
def cleanText(x):
'''General purpose text cleaner that turns text into a list of tokens fit for Word2Vec and Doc2Vec
INPUT: string;
OUTPUT: list of tokens.'''
#Fix encoding in case there is a problem
x = ftfy.fix_encoding(x)
#Unicode Sandwhich...
x = x.encode('UTF-8') #encode as UTF-8
input_code = chardet.detect(bytes(x))['encoding'] #Auto-detect encoding
try:
u_string = x.decode(input_code) #Decode
except:
u_string = x.decode('UTF-8')
re.sub(r'\w{4}', u'xxxx', u_string, flags=re.UNICODE) #Handle funny unicode artifacts
x = u_string.encode('UTF-8') #Encode back to UTF-8
x = str(x) #Convert to string
x = re.sub(r'[\[\]\"\'\,]',' ', x) #Remove backslashes
x = re.sub(r'[^\w\s]',' ',x) #Insert space after end of each sentence
x = x.split() #split string to list
return x
def horo(self, irc, msg, args, midnight_check, psign):
"""[bool: Only print when local hour is 0 - default False] [sign: if not set will dump all signs]
Returns horoscope in french for one zodiacal sign or all
"""
signs = []
tsigns = [
'belier', 'taureau', 'gemeaux', 'cancer', 'lion', 'vierge',
'balance', 'scorpion', 'sagittaire', 'capricorne', 'verseau',
'poissons'
]
if midnight_check and int(
datetime.datetime.fromtimestamp(time()).strftime('%H')):
self.log.info(
"Horoscope plugin: checked and not [00:00-00:59] local time")
return
if psign:
if psign not in tsigns:
irc.error("Same player try again!")
else:
signs = [psign]
else:
signs = tsigns
for sign in signs:
url = "https://www.lhoroscope.com/horoscope-general/horoscope-%s-du-jour.asp" % sign
try:
result = requests.get(url)
soup = BeautifulSoup(result.content)
h = soup.find(class_="panel-body").text.strip()
h = fix_encoding(h)
except Exception as e:
irc.error("Error {}".format(e))
break
else:
irc.reply(ircutils.bold(sign.title()) + " : " + h,
prefixNick=False)
sleep(2)
def run() -> None:
files = rotation.files()
n = len(files)
time_until = min(
TIME_UNTIL_FULL_ROTATION,
TIME_UNTIL_SUPPLEMENTAL_ROTATION) - datetime.timedelta(weeks=1)
if n >= TOTAL_RUNS:
print(
'It is the moment of discovery, the triumph of the mind, and the end of this rotation.'
)
if time_until < datetime.timedelta(days=3):
for f in files:
print('Removing {f}'.format(f=f))
os.remove(f)
return
if n == 0 and TIME_UNTIL_FULL_ROTATION > datetime.timedelta(
7) and TIME_UNTIL_SUPPLEMENTAL_ROTATION > datetime.timedelta(7):
print(
'The monks of the North Tree rarely saw their kodama until the rotation, when it woke like a slumbering, angry bear.'
)
print('ETA: {t}'.format(
t=dtutil.display_time(time_until.total_seconds())))
return
all_prices = {}
for url in configuration.get_list('cardhoarder_urls'):
s = fetcher_internal.fetch(url)
s = ftfy.fix_encoding(s)
all_prices[url] = parse_cardhoarder_prices(s)
url = configuration.get_str('mtgotraders_url')
if url:
s = fetcher_internal.fetch(url)
all_prices['mtgotraders'] = parse_mtgotraders_prices(s)
run_number = process(all_prices)
if run_number == TOTAL_RUNS:
make_final_list()
def stream_json_file(self,
json_file_name,
output_file_name="parsed_tweets.csv",
stop_at=1000000000,
verbose=False):
# Create an empty csv file that we will append to
# Create a header row for it
if verbose:
print("Initalizing Output File: %s" % output_file_name)
print("Generating Header Row")
with open('%s' % output_file_name, 'w') as f:
f.write(
'id,text,created_at,language,retweet_count,screen_name,country,user_followers_count,time_zone,user_account_location,longitude,lattitude,name\n'
) # Column headers and a trailing new line . MAKE SURE the /n is attached to the last field: eg. text/n
tweet_counter = 0
for i in open(json_file_name):
if tweet_counter > stop_at:
break
try:
# Put the data from the current tweet into a list
# Parse the current tweet
current_tweet = json.loads(i)
##################################################################
## Elements that are 1 level deep ## ##
# Get the id or insert a nan if not found
if 'id' in current_tweet:
id = current_tweet['id']
else:
id = np.nan
# Get text or insert nan
if 'text' in current_tweet:
text = current_tweet['text']
# the fix_encoding function from the FTFY package takes care of weird characters
text = ftfy.fix_encoding(text)
else:
text = np.nan
# Get created_at or insert nan
if 'created_at' in current_tweet:
created_at = current_tweet['created_at']
else:
created_at = np.nan
# Get language or insert nan
if 'lang' in current_tweet:
language = current_tweet['lang']
else:
language = np.nan
# get retweet count or insert nan
if 'retweet_count' in current_tweet:
retweet_count = current_tweet['retweet_count']
else:
retweet_count = np.nan
## Elements that are 2 levels deep ### ###
# For elements that are 2 layers deep use != None when searching because javascript uses None as its null operator
# get screen name or insert nan
if 'user' in current_tweet and 'screen_name' in current_tweet[
'user']:
screen_name = current_tweet['user']['screen_name']
else:
screen_name = np.nan
# get name or insert nan
if 'user' in current_tweet and 'name' in current_tweet['user']:
name = current_tweet['user']['name']
else:
name = np.nan
# get country or insert nan
if current_tweet['place'] != None and current_tweet['place'][
'country'] != None:
country = current_tweet['place']['country']
else:
country = np.nan
# get the author's follower count or nan
if current_tweet['user'] != None and current_tweet['user'][
'followers_count'] != None:
followers_count = current_tweet['user']['followers_count']
else:
followers_count = np.nan
# get the timezone or nan
if current_tweet['user'] != None and current_tweet['user'][
'time_zone'] != None:
time_zone = current_tweet['user']['time_zone']
else:
time_zone = np.nan
# get the account location or insert nan
if current_tweet['user'] != None and current_tweet['user'][
'location'] != None:
account_location = current_tweet['user']['location']
account_location = ftfy.fix_encoding(account_location)
else:
account_location = np.nan
###### Elements that are 3 levels deep ##################################
if current_tweet['coordinates'] != None and current_tweet[
'coordinates']['coordinates'] != None and len(
current_tweet['coordinates']['coordinates']) == 2:
longitude = current_tweet['coordinates']['coordinates'][0]
else:
longitude = np.nan
if current_tweet['coordinates'] != None and current_tweet[
'coordinates']['coordinates'] != None and len(
current_tweet['coordinates']['coordinates']) == 2:
lattitude = current_tweet['coordinates']['coordinates'][1]
else:
lattitude = np.nan
######################################################################################################
# Assemble the row
cleaned_current_tweet = [
id, text, created_at, language, retweet_count, screen_name,
country, followers_count, time_zone, account_location,
longitude, lattitude, name
]
# Increment the Tweet Counter
tweet_counter = tweet_counter + 1
# Give the user a progress update
if tweet_counter % 1000 == 0 and verbose:
print(" %d Tweets Parsed so far....." % tweet_counter)
# append the current tweet as a row to the csv
with open('%s' % output_file_name, 'a', newline='') as f:
writer = csv.writer(f)
writer.writerow(cleaned_current_tweet)
except:
pass
if verbose:
print(" ")
print(" Parsing Complete: %d Tweets Parsed " % tweet_counter)
def add_items(pickle_data, items, descriptions):
for i in range(len(items)):
pickle_data['Food'].append(
[ftfy.fix_encoding(items[i]),
ftfy.fix_encoding(descriptions[i])])
return pickle_data
def enc(text):
if text:
return ftfy.fix_encoding(HTMLParser().unescape(text)).encode(
"utf-16be", "replace")
#cv2.imshow("Image", image)
#cv2.imshow("Output", gray)
#cv2.waitKey(5000)
# writing extracted data into a text file
text_output = open('outputbase.txt', 'w', encoding='utf-8')
text_output.write(text)
text_output.close()
file = open('outputbase.txt', 'r', encoding='utf-8')
text = file.read()
# print(text)
# Cleaning all the gibberish text
text = ftfy.fix_text(text)
text = ftfy.fix_encoding(text)
print(text)
# Initializing data variable
name = None
fname = None
dob = None
pan = None
nameline = []
dobline = []
panline = []
text0 = []
text1 = []
text2 = []
# Searching for PAN
def clean_html_string(x):
return ftfy.fix_encoding(
ftfy.fix_text(x.replace("\n", "").replace("\t", "").strip(),
normalization='NFKC'))
def main():
print(ftfy.fix_encoding('주문하다 - to intent for?'))
def fix_encoding(s):
s = ftfy.fix_encoding(s)
better, _ = codecs.escape_decode(s)
return better.decode("utf-8").strip()
def fix_encoding(elem):
if isinstance(elem, unicode):
return ftfy.fix_encoding(elem)
else:
return elem
response = urllib.urlopen('http://www.gutenberg.org/files/2600/2600-0.txt')
print(response.info())
html = response.read()
response.close()
words = {}
html = html.split()
print(len(html))
def addWord(words,word):
if not word in words:
words[word] = 0
words[word] += 1
for word in html:
word = fix_encoding(word.decode('utf-8').lower())
word = ''.join(ch for ch in word if ch not in exclude)
if word.find(u"\u2014") > 0:
word = word.split(u"\u2014")
if not type(word) is list:
addWord(words,word)
else:
for w in word:
addWord(words,w)
print(words)
def __init__(self, name):
name = ftfy.fix_encoding(name)
self._name = self._english_artist(name)
def fix_component_encodings(cls, components):
return {k: ftfy.fix_encoding(safe_decode(v)) for k, v in six.iteritems(components)}
def normalize_text(text, fix_encoding=False, strip_emojis=False):
"""
Normalize text:
* normalize accents (using NFC convention),
* strip control/invisible chars and leftover combining diacritics,
* undo ligatures,
* normalize quotes, apostrophes and unicode characters (dashes, etc.),
* normalize spaces (all spaces, including nbsp and tabs, will be encoded as 0x20),
* strip and collapse multiple spaces into one,
* etc.
Optionally:
* try to detect and fix encoding issues (see `ftfy.fix_encoding <https://ftfy.readthedocs.io/en/latest/>`_)
on a per-sentence basis (delimited by newlines);
* strip emojis (using a simple regex, not all cases are covered !)
:param text: the text to normalize, newlines will be preserved;
:param fix_encoding: if set, use ftfy to fix encoding issues on a per-sentence basis;
:param strip_emojis: if set, try to find and strip unicode emojis;
:return: the normalized text
"""
# optionally fix encoding using ftfy
if fix_encoding and wrong_encoding_pattern.search(text) is not None:
try:
import ftfy
text = '\n'.join(
ftfy.fix_encoding(t) if wrong_encoding_pattern.
search(t) is not None else t for t in text.split('\n'))
except ModuleNotFoundError:
print(
'WARNING: norm_punc.py, fixing encoding requires the ftfy package: pip install ftfy.'
)
# normalize (e.g. combining diacritics)
text = unicodedata.normalize('NFC', text)
# optionally strip emojis
if strip_emojis:
# I formally used the emoji library, which is really nice but slooooow (and doesn't cover ASCII misc symbols).
# I thus preferred to use a simpler regex that covers most cases is is waaaay faster
# (203ms to process 164343 short sentences, against 31s with emoji)
text = emoji_pattern.sub(' ', text)
# apply patterns in order
for i, (typ, pattern, replace) in enumerate(normalization_patterns):
if typ == REG:
text = pattern.sub(replace, text)
else:
text = text.replace(pattern, replace)
# normalize spaces
text = spaces_pattern.sub(' ', text)
# don't forget to normalise spaces in the beginning and end
text = re.sub(r'(^|\n)\s+', r'\1', text)
text = re.sub(r'\s+(\n|$)', r'\1', text)
return text
import ftfy
with open("data/small_text.json") as fp:
for line in fp:
print(ftfy.fix_encoding(line), end='')
def fix_encoding(string):
return jsonify({
"fixed": ftfy.fix_encoding(string)
})