Develop Reference

STANZA/ RuntimeError: Integer division of tensors using div or / is no longer supported

I want to use stanza for tokenizing, pos tagging and parsing some text I have, but it keeps giving me this error. I've tried changing the way a I call it but nothing happens. Any ideas?
My code(Here a iterate through a list of list of text and appli stanza to each one)
t = time()
data_stanza = []
for text in data:
stz = apply_stanza(text[0])
data_stanza.append(stz)
print('Time to run: {} mins'.format(round((time() - t) / 60, 2)))
This is the function I use to apply_stanza to each text:
nlp = stanza.Pipeline('pt')
def apply_stanza(text):
doc = nlp(text)
All = []
for sent in doc.sentences:
for word in sent.words:
All.append((word.id,word.text,word.lemma,word.upos,word.feats,word.head,word.deprel))
return All
The error:
---------------------------------------------------------------------------
RuntimeError Traceback (most recent call last)
<ipython-input-17-7ac303eec8e8> in <module>
3 data_staza = []
4 for text in data:
----> 5 stz = apply_stanza(text[0])
6 data_stanza.append(stz)
7
<ipython-input-16-364c3ac30f32> in apply_stanza(text)
2
3 def apply_stanza(text):
----> 4 doc = nlp(text)
5 All = []
6 for sent in doc.sentences:
~\anaconda3\lib\site-packages\stanza\pipeline\core.py in __call__(self, doc)
174 assert any([isinstance(doc, str), isinstance(doc, list),
175 isinstance(doc, Document)]), 'input should be either str, list or Document'
--> 176 doc = self.process(doc)
177 return doc
178
~\anaconda3\lib\site-packages\stanza\pipeline\core.py in process(self, doc)
168 for processor_name in PIPELINE_NAMES:
169 if self.processors.get(processor_name):
--> 170 doc = self.processors[processor_name].process(doc)
171 return doc
172
~\anaconda3\lib\site-packages\stanza\pipeline\mwt_processor.py in process(self, document)
31 preds = []
32 for i, b in enumerate(batch):
---> 33 preds += self.trainer.predict(b)
34
35 if self.config.get('ensemble_dict', False):
~\anaconda3\lib\site-packages\stanza\models\mwt\trainer.py in predict(self, batch, unsort)
77 self.model.eval()
78 batch_size = src.size(0)
---> 79 preds, _ = self.model.predict(src, src_mask, self.args['beam_size'])
80 pred_seqs = [self.vocab.unmap(ids) for ids in preds] # unmap to tokens
81 pred_seqs = utils.prune_decoded_seqs(pred_seqs)
~\anaconda3\lib\site-packages\stanza\models\common\seq2seq_model.py in predict(self, src, src_mask, pos, beam_size)
259 done = []
260 for b in range(batch_size):
--> 261 is_done = beam[b].advance(log_probs.data[b])
262 if is_done:
263 done += [b]
~\anaconda3\lib\site-packages\stanza\models\common\beam.py in advance(self, wordLk, copy_indices)
82 # bestScoresId is flattened beam x word array, so calculate which
83 # word and beam each score came from
---> 84 prevK = bestScoresId / numWords
85 self.prevKs.append(prevK)
86 self.nextYs.append(bestScoresId - prevK * numWords)
RuntimeError: Integer division of tensors using div or / is no longer supported, and in a future release div will perform
true division as in Python 3. Use true_divide or floor_divide (// in Python) instead.
ATT: It turns after all that it was and error with the mwt module of stanza pipeline, so I just specified not to use it.

Use // for division instead of /.
Try to edit your code as follows:
print('Time to run: {} mins'.format(round((time() - t) // 60, 2)))

Using floor division (//) will floor the result to the largest possible integer.
Using torch.true_divide(Dividend, Divisor) or numpy.true_divide(Dividend, Divisor) in stead.
For example: 3/4 = torch.true_divide(3, 4)
https://pytorch.org/docs/stable/generated/torch.true_divide.html
https://numpy.org/doc/stable/reference/generated/numpy.true_divide.html

Error when using google translate API to translate a dataframe

I'm trying to translate part of SQuAD 1.1 dataset to Sinhalese. I don't know whether i can use the json file straight into translation
What i tried so far is making a little dataframe of SQuAD dataset and try to translate that as a demo to myself. But i got different errors. Below is the error i'm getting now. Can you help me to fix that error or tell me a better way to complete my task using python.
```import googletrans
from googletrans import Translator
import os
from google.cloud import translate_v2 as translate
os.environ['GOOGLE_APPLICATION_CREDENTIALS']=r"C:\Users\Sathsara\Documents\Python Learning\Translation test\translationAPI\flash-medley-278816-b2012b874797.json"
# create a translator object
translator = Translator()
# use translate method to translate a string - by default, the destination language is english
translated = translator.translate('I am Sathsara Rasantha',dest='si')
# the translate method returns an object
print(translated)
# obtain translated string by using attribute .text
translated.text
import pandas as pd
translate_example = pd.read_json("example2.json")
translate_example
contexts = []
questions = []
answers_text = []
answers_start = []
for i in range(translate_example.shape[0]):
topic = translate_example.iloc[i,0]['paragraphs']
for sub_para in topic:
for q_a in sub_para['qas']:
questions.append(q_a['question'])
answers_start.append(q_a['answers'][0]['answer_start'])
answers_text.append(q_a['answers'][0]['text'])
contexts.append(sub_para['context'])
df = pd.DataFrame({"context":contexts, "question": questions, "answer_start": answers_start, "text": answers_text})
df
df=df.loc[0:2,:]
df
# make a deep copy of the data frame
df_si = df.copy()
# translate columns' name using rename function
df_si.rename(columns=lambda x: translator.translate(x).text, inplace=True)
df_si.columns
translations = {}
for column in df_si.columns:
# unique elements of the column
unique_elements = df_si[column].unique()
for element in unique_elements:
# add translation to the dictionary
translations[element] = translator.translate(element,dest='si').text
print(translations)
# modify all the terms of the data frame by using the previously created dictionary
df_si.replace(translations, inplace = True)
# check translation
df_si.head()```
This is the error i get
> --------------------------------------------------------------------------- TypeError Traceback (most recent call
> last) <ipython-input-24-f55a5ca59c36> in <module>
> 5 for element in unique_elements:
> 6 # add translation to the dictionary
> ----> 7 translations[element] = translator.translate(element,dest='si').text
> 8
> 9 print(translations)
>
> ~\Anaconda3\lib\site-packages\googletrans\client.py in translate(self,
> text, dest, src)
> 170
> 171 origin = text
> --> 172 data = self._translate(text, dest, src)
> 173
> 174 # this code will be updated when the format is changed.
>
> ~\Anaconda3\lib\site-packages\googletrans\client.py in
> _translate(self, text, dest, src)
> 73 text = text.decode('utf-8')
> 74
> ---> 75 token = self.token_acquirer.do(text)
> 76 params = utils.build_params(query=text, src=src, dest=dest,
> 77 token=token)
>
> ~\Anaconda3\lib\site-packages\googletrans\gtoken.py in do(self, text)
> 199 def do(self, text):
> 200 self._update()
> --> 201 tk = self.acquire(text)
> 202 return tk
>
> ~\Anaconda3\lib\site-packages\googletrans\gtoken.py in acquire(self,
> text)
> 144 a = []
> 145 # Convert text to ints
> --> 146 for i in text:
> 147 val = ord(i)
> 148 if val < 0x10000:
>
> TypeError: 'numpy.int64' object is not iterable

OverflowError when trying to convert generators to lists

I'm trying to extract dates from txt files using datefinder.find_dates which returns a generator object. Everything works fine until I try to convert the generator to list, when i get the following error.
I have been looking around for a solution but I can't figure out a solution to this, not sure I really understand the problem neither.
import datefinder
import glob
path = "some_path/*.txt"
files = glob.glob(path)
dates_dict = {}
for name in files:
with open(name, encoding='utf8') as f:
dates_dict[name] = list(datefinder.find_dates(f.read()))
Returns :
---------------------------------------------------------------------------
OverflowError Traceback (most recent call last)
<ipython-input-53-a4b508b01fe8> in <module>()
1 for name in files:
2 with open(name, encoding='utf8') as f:
----> 3 dates_dict[name] = list(datefinder.find_dates(f.read()))
C:\ProgramData\Anaconda3\lib\site-packages\datefinder\__init__.py in
find_dates(self, text, source, index, strict)
29 ):
30
---> 31 as_dt = self.parse_date_string(date_string, captures)
32 if as_dt is None:
33 ## Dateutil couldn't make heads or tails of it
C:\ProgramData\Anaconda3\lib\site-packages\datefinder\__init__.py in
parse_date_string(self, date_string, captures)
99 # otherwise self._find_and_replace method might corrupt
them
100 try:
--> 101 as_dt = parser.parse(date_string, default=self.base_date)
102 except ValueError:
103 # replace tokens that are problematic for dateutil
C:\ProgramData\Anaconda3\lib\site-packages\dateutil\parser\_parser.py in
parse(timestr, parserinfo, **kwargs)
1354 return parser(parserinfo).parse(timestr, **kwargs)
1355 else:
-> 1356 return DEFAULTPARSER.parse(timestr, **kwargs)
1357
1358
C:\ProgramData\Anaconda3\lib\site-packages\dateutil\parser\_parser.py in
parse(self, timestr, default, ignoretz, tzinfos, **kwargs)
651 raise ValueError("String does not contain a date:",
timestr)
652
--> 653 ret = self._build_naive(res, default)
654
655 if not ignoretz:
C:\ProgramData\Anaconda3\lib\site-packages\dateutil\parser\_parser.py in
_build_naive(self, res, default)
1222 cday = default.day if res.day is None else res.day
1223
-> 1224 if cday > monthrange(cyear, cmonth)[1]:
1225 repl['day'] = monthrange(cyear, cmonth)[1]
1226
C:\ProgramData\Anaconda3\lib\calendar.py in monthrange(year, month)
122 if not 1 <= month <= 12:
123 raise IllegalMonthError(month)
--> 124 day1 = weekday(year, month, 1)
125 ndays = mdays[month] + (month == February and isleap(year))
126 return day1, ndays
C:\ProgramData\Anaconda3\lib\calendar.py in weekday(year, month, day)
114 """Return weekday (0-6 ~ Mon-Sun) for year (1970-...), month(1- 12),
115 day (1-31)."""
--> 116 return datetime.date(year, month, day).weekday()
117
118
OverflowError: Python int too large to convert to C long
Can someone explain this clearly?
Thanks in advance
REEDIT : After taking into consideration the remarks that were made, I found a minimal, readable and verifiable example. The error occurs on :
import datefinder
generator = datefinder.find_dates("466990103060049")
for s in generator:
pass

This looks to be a bug in the library you are using. It is trying to parse the string as a year, but that this year is too big to be handled by Python. The library that datefinder is using says that it raises an OverflowError in this instance, but that datefinder is ignoring this possibility.
One quick and dirty hack just to get it working would be to do:
>>> datefinder.ValueError = ValueError, OverflowError
>>> list(datefinder.find_dates("2019/02/01 is a date and 466990103060049 is not"))
[datetime.datetime(2019, 2, 1, 0, 0)]

how to fix - error: bad escape \u at position 0

Hello I'm trying to export a gmap html using ipywidgets in jupyter notebook but am encountering the following error: - error: bad escape \u at position 0.
I'm new to programing and could use help fixing whatever is causing this error to occur. If there is any easier way to go about exporting the html file I'm happy to change approaches.
Thanks
Here is a snippet of the code: I can add the entire thing if its helpful.
import pandas as pd
import gmaps
from ipywidgets.embed import embed_minimal_html
from ipywidgets import IntSlider
gmaps.configure(api_key='XXXX')
pd.options.mode.chained_assignment = None # default='warn'
file2 = '005 lat:long.csv'
state2 = pd.read_csv(file2)
state2 = state2.rename(columns={'Address1': 'address', 'City':'city',
'State':'state', 'Zip': 'zip'})
storenumbs = state2['Store'].str.split('#', expand=True)
state2 = state2.join(storenumbs)
state2 = state2.drop(['Store', 0], axis=1)
state2 = state2.rename(columns={1: 'store_#'})
state2['store_#'] = state2['store_#'].astype(int)
fig = gmaps.figure(center=(42.5, -71.4), map_type='TERRAIN', zoom_level=9.8)
scale = 4
one_layer = (gmaps.symbol_layer(low_points_lat_long, fill_color='red', stroke_color='red', scale= scale))
two_layer = (gmaps.symbol_layer(low_med_points_lat_long, fill_color='red', stroke_color='yellow', scale= scale))
three_layer = (gmaps.symbol_layer(med_high_points_lat_long, fill_color='yellow', stroke_color='green', scale= scale))
four_layer = (gmaps.symbol_layer(high_points_lat_long, fill_color='green', stroke_color='green', scale= scale))
fig.add_layer(one_layer)
fig.add_layer(two_layer)
fig.add_layer(three_layer)
fig.add_layer(four_layer)
fig
embed_minimal_html('export.html', views=[fig]
Long Form Error Bellow
)
KeyError Traceback (most recent call last)
~/miniconda3/lib/python3.7/sre_parse.py in parse_template(source, pattern)
1020 try:
-> 1021 this = chr(ESCAPES[this][1])
1022 except KeyError:
KeyError: '\\u'
During handling of the above exception, another exception occurred:
error Traceback (most recent call last)
<ipython-input-7-c096ac365396> in <module>
20
21 slider = IntSlider(value=40)
---> 22 embed_minimal_html('export.html', views=[slider], title='Widgets export')
~/miniconda3/lib/python3.7/site-packages/ipywidgets/embed.py in embed_minimal_html(fp, views, title, template, **kwargs)
300 {embed_kwargs}
301 """
--> 302 snippet = embed_snippet(views, **kwargs)
303
304 values = {
~/miniconda3/lib/python3.7/site-packages/ipywidgets/embed.py in embed_snippet(views, drop_defaults, state, indent, embed_url, requirejs, cors)
266 widget_views = u'\n'.join(
267 widget_view_template.format(view_spec=escape_script(json.dumps(view_spec)))
--> 268 for view_spec in data['view_specs']
269 )
270
~/miniconda3/lib/python3.7/site-packages/ipywidgets/embed.py in <genexpr>(.0)
266 widget_views = u'\n'.join(
267 widget_view_template.format(view_spec=escape_script(json.dumps(view_spec)))
--> 268 for view_spec in data['view_specs']
269 )
270
~/miniconda3/lib/python3.7/site-packages/ipywidgets/embed.py in escape_script(s)
239 involving `<` is readable.
240 """
--> 241 return script_escape_re.sub(r'\u003c\1', s)
242
243 #doc_subst(_doc_snippets)
~/miniconda3/lib/python3.7/re.py in _subx(pattern, template)
307 def _subx(pattern, template):
308 # internal: Pattern.sub/subn implementation helper
--> 309 template = _compile_repl(template, pattern)
310 if not template[0] and len(template[1]) == 1:
311 # literal replacement
~/miniconda3/lib/python3.7/re.py in _compile_repl(repl, pattern)
298 def _compile_repl(repl, pattern):
299 # internal: compile replacement pattern
--> 300 return sre_parse.parse_template(repl, pattern)
301
302 def _expand(pattern, match, template):
~/miniconda3/lib/python3.7/sre_parse.py in parse_template(source, pattern)
1022 except KeyError:
1023 if c in ASCIILETTERS:
-> 1024 raise s.error('bad escape %s' % this, len(this))
1025 lappend(this)
1026 else:
error: bad escape \u at position 0

This is an error in Python 3.7, and an issue with Python 3.6 (but it is OK with Python 2.7).
If you use a raw string (prefixed by "r") for the replacement in re.sub function, then the \u is escaped. For instance, r'\u003c\1' is like '\\u003c\\1': this is a string '\u', followed by '003c' and \1.
The solution is to write:
return script_escape_re.sub('\u003c\\1', s)
Quoting the documentation:
Changed in version 3.7: Unknown escapes in repl consisting of '\' and an ASCII letter now are errors.

I was facing a similar issue while trying to escape Unicode characters that have the pattern \uXXXX. Let's take a string containing Unicode characters:
>>> text = "The \u201c\u3010\u3011\u201d in this template are used to mark the variables"
>>> text
'The “【】” in this template are used to mark the variables'
Escape the Unicode characters:
>>> text = text.encode('unicode_escape').decode('ascii')
>>> text
'The \\u201c\\u3010\\u3011\\u201d in this template are used to mark the variables'
And then replace them using re.sub(r'\\u(.){4}', '', text):
>>> import re
>>> re.sub(r'\\u(.){4}', '', text)
'The in this template are used to mark the variables'

I have had the same issue during
[m.start() for m in re.finditer('Valuation Date")', 'dummytext')]
*** sre_constants.error: unbalanced parenthesis at position 15
But it was solved with re.escape help
[m.start() for m in re.finditer(re.escape('Valuation Date")'), 'dummytext')]
Enjoy.

Pandas Series.apply doesn't work consist of strings

It's seems possible to relate with Japanese Language problem,
So I asked in Japanese StackOverflow also.
When I use string just object, it works fine.
I tried to encode but I couldn't find the reason of this error.
Could you please give me advice?
MeCab is an open source text segmentation library for use with text written in the Japanese language originally developed by the Nara Institute of Science and Technology and currently maintained by Taku Kudou (工藤拓) as part of his work on the Google Japanese Input project.
https://en.wikipedia.org/wiki/MeCab
sample.csv
0,今日も夜まで働きました。
1,オフィスには誰もいませんが、エラーと格闘中
2,デバッグばかりしていますが、どうにもなりません。
This is Pandas Python3 code
import pandas as pd
import MeCab
# https://en.wikipedia.org/wiki/MeCab
from tqdm import tqdm_notebook as tqdm
# This is working...
df = pd.read_csv('sample.csv', encoding='utf-8')
m = MeCab.Tagger ("-Ochasen")
text = "りんごを食べました、そして、みかんも食べました"
a = m.parse(text)
print(a)# working!
# But I want to use Pandas's Series
def extractKeyword(text):
"""Morphological analysis of text and returning a list of only nouns"""
tagger = MeCab.Tagger('-Ochasen')
node = tagger.parseToNode(text)
keywords = []
while node:
if node.feature.split(",")[0] == u"名詞": # this means noun
keywords.append(node.surface)
node = node.next
return keywords
aa = extractKeyword(text) #working!!
me = df.apply(lambda x: extractKeyword(x))
#TypeError: ("in method 'Tagger_parseToNode', argument 2 of type 'char const *'", 'occurred at index 0')
This is the trace error
りんご リンゴ りんご 名詞-一般
を ヲ を 助詞-格助詞-一般
食べ タベ 食べる 動詞-自立 一段 連用形
まし マシ ます 助動詞 特殊・マス 連用形
た タ た 助動詞 特殊・タ 基本形
、 、 、 記号-読点
そして ソシテ そして 接続詞
、 、 、 記号-読点
みかん ミカン みかん 名詞-一般
も モ も 助詞-係助詞
食べ タベ 食べる 動詞-自立 一段 連用形
まし マシ ます 助動詞 特殊・マス 連用形
た タ た 助動詞 特殊・タ 基本形
EOS
---------------------------------------------------------------------------
TypeError Traceback (most recent call last)
<ipython-input-174-81a0d5d62dc4> in <module>()
32 aa = extractKeyword(text) #working!!
33
---> 34 me = df.apply(lambda x: extractKeyword(x))
~/anaconda3/lib/python3.6/site-packages/pandas/core/frame.py in apply(self, func, axis, broadcast, raw, reduce, args, **kwds)
4260 f, axis,
4261 reduce=reduce,
-> 4262 ignore_failures=ignore_failures)
4263 else:
4264 return self._apply_broadcast(f, axis)
~/anaconda3/lib/python3.6/site-packages/pandas/core/frame.py in _apply_standard(self, func, axis, ignore_failures, reduce)
4356 try:
4357 for i, v in enumerate(series_gen):
-> 4358 results[i] = func(v)
4359 keys.append(v.name)
4360 except Exception as e:
<ipython-input-174-81a0d5d62dc4> in <lambda>(x)
32 aa = extractKeyword(text) #working!!
33
---> 34 me = df.apply(lambda x: extractKeyword(x))
<ipython-input-174-81a0d5d62dc4> in extractKeyword(text)
20 """Morphological analysis of text and returning a list of only nouns"""
21 tagger = MeCab.Tagger('-Ochasen')
---> 22 node = tagger.parseToNode(text)
23 keywords = []
24 while node:
~/anaconda3/lib/python3.6/site-packages/MeCab.py in parseToNode(self, *args)
280 __repr__ = _swig_repr
281 def parse(self, *args): return _MeCab.Tagger_parse(self, *args)
--> 282 def parseToNode(self, *args): return _MeCab.Tagger_parseToNode(self, *args)
283 def parseNBest(self, *args): return _MeCab.Tagger_parseNBest(self, *args)
284 def parseNBestInit(self, *args): return _MeCab.Tagger_parseNBestInit(self, *args)
TypeError: ("in method 'Tagger_parseToNode', argument 2 of type 'char const *'", 'occurred at index 0')w

I see you got some help on the Japanese StackOverflow, but here's an answer in English:
The first thing to fix is that read_csv was treating the first line of your example.csv as the header. To fix that, use the names argument in read_csv.
Next, df.apply will by default apply the function on columns of the dataframe. You need to do something like df.apply(lambda x: extractKeyword(x['String']), axis=1), but this won't work because each sentence will have a different number of nouns and Pandas will complain it cannot stack a 1x2 array on top of a 1x5 array. The simplest way is to apply on the Series of String.
The final problem is, there's a bug in the MeCab Python3 bindings: see https://github.com/SamuraiT/mecab-python3/issues/3 You found a workaround by running parseToNode twice, you can also call parse before parseToNode.
Putting all these three things together:
import pandas as pd
import MeCab
df = pd.read_csv('sample.csv', encoding='utf-8', names=['Number', 'String'])
def extractKeyword(text):
"""Morphological analysis of text and returning a list of only nouns"""
tagger = MeCab.Tagger('-Ochasen')
tagger.parse(text)
node = tagger.parseToNode(text)
keywords = []
while node:
if node.feature.split(",")[0] == u"名詞": # this means noun
keywords.append(node.surface)
node = node.next
return keywords
me = df['String'].apply(extractKeyword)
print(me)
When you run this script, with the example.csv you provide:
➜ python3 demo.py
0 [今日, 夜]
1 [オフィス, 誰, エラー, 格闘, 中]
2 [デバッグ]
Name: String, dtype: object

parseToNode fail everytime ,
so needed to put this code
tagger.parseToNode('dummy')
before
node = tagger.parseToNode(text)
and It's worked!
But I don't know the reason, maybe parseToNode method has bug..
def extractKeyword(text):
"""Morphological analysis of text and returning a list of only nouns"""
tagger = MeCab.Tagger('-Ochasen')
tagger.parseToNode('ダミー')
node = tagger.parseToNode(text)
keywords = []
while node:
if node.feature.split(",")[0] == u"名詞": # this means noun
keywords.append(node.surface)
node = node.next
return keywords