Latest update seems to have similar issue as before, resources not found

[axel0681]

Hello,

Same Problem on our sites.
Package Versioncom.paperlessng.cloudronapp@1.5.2

The last step "saving the Dokument failed"

Thx
Axel

abc.pdf: The following error occurred while consuming 
abc.pdf: 
**********************************************************************
  Resource stopwords not found.
  Please use the NLTK Downloader to obtain the resource:

  >>> import nltk
  >>> nltk.download('stopwords')
  
  For more information see: https://www.nltk.org/data.html

  Attempted to load corpora/stopwords

  Searched in:
    - '/usr/share/nltk_data'
**********************************************************************

[girish]

@axel0681 can you check the contents of /usr/local/share/nltk_data ?

[girish]

@axel0681 Actually, can you set PAPERLESS_NLTK_DIR=/usr/local/share/ntlk_data in paperless.conf and restart the app ?

[scooke]

@girish Weird. Adding that to the conf, while still on v1.5.1, causes it to fail again.

The contents of that directory are: corpora, stemmers, tokenizers

Despite the error, which reads thusly:

********************************************************************** Resource stopwords not found. Please use the NLTK Downloader to obtain the resource: >>> import nltk >>> nltk.download('stopwords')  For more information see: https://www.nltk.org/data.html Attempted to load corpora/stopwords Searched in: - '/usr/local/share/ntlk_data' **********************************************************************

there is /usr/local/share/ntlk_data/corpora/stopwords, which includes a stopwords.zip file.

I noticed that in the conf file, the env PAPERLESS_NLTK_DIR=/usr/local/share/ntlk_data has the NLTK highlighted in brown, which must mean that the syntax is off, right?

Finally, I also noticed at the bottom of the env file that the following are all commented out:

# Binaries

#PAPERLESS_CONVERT_BINARY=/usr/bin/convert
#PAPERLESS_GS_BINARY=/usr/bin/gs
#PAPERLESS_OPTIPNG_BINARY=/usr/bin/optipng

Commenting out the line you suggested I add (running v1.5.1) sees the app functioning properly.

I guess I will now try to update again, then add that line to the conf file and see what happens.

[scooke]

@scooke I've updated to v1.5.2, and neither with or without that conf setting does it work. Still saying it can't find corpora/stopwords, even though it's there.

Logs:

Jan 31 22:31:41 [2023-01-31 21:31:41,532] [INFO] [celery.worker.strategy] Task documents.tasks.consume_file[3e572bd2-9ccc-4ee4-9201-fb350b47cfd9] received
Jan 31 22:31:41 [2023-01-31 21:31:41,652] [INFO] [paperless.consumer] Consuming Doc - May 25, 2014, 11-08 AM.pdf
Jan 31 22:31:44 [2023-01-31 21:31:44,150] [INFO] [ocrmypdf._pipeline] page is facing ⇧, confidence 4.01 - no change
Jan 31 22:31:53 [2023-01-31 21:31:53,935] [INFO] [ocrmypdf._sync] Postprocessing...
Jan 31 22:31:54 [2023-01-31 21:31:54,867] [INFO] [ocrmypdf._pipeline] Optimize ratio: 1.40 savings: 28.3%
Jan 31 22:31:54 [2023-01-31 21:31:54,872] [INFO] [ocrmypdf._sync] Output file is a PDF/A-2B (as expected)
Jan 31 22:31:57 1263:M 31 Jan 2023 21:31:57.078 * 10 changes in 300 seconds. Saving...
Jan 31 22:31:57 1263:M 31 Jan 2023 21:31:57.079 * Background saving started by pid 1303
Jan 31 22:31:57 1303:C 31 Jan 2023 21:31:57.083 * DB saved on disk
Jan 31 22:31:57 1303:C 31 Jan 2023 21:31:57.084 * RDB: 0 MB of memory used by copy-on-write
Jan 31 22:31:57 1263:M 31 Jan 2023 21:31:57.180 * Background saving terminated with success
Jan 31 22:32:00 [2023-01-31 21:32:00,895] [ERROR] [paperless.consumer] The following error occurred while consuming Doc - May 25, 2014, 11-08 AM.pdf:
Jan 31 22:32:00 **********************************************************************
Jan 31 22:32:00 Resource stopwords not found.
Jan 31 22:32:00 Please use the NLTK Downloader to obtain the resource:
Jan 31 22:32:00
Jan 31 22:32:00 >>> import nltk
Jan 31 22:32:00 >>> nltk.download('stopwords')
Jan 31 22:32:00
Jan 31 22:32:00 For more information see: https://www.nltk.org/data.html
Jan 31 22:32:00
Jan 31 22:32:00 Attempted to load corpora/stopwords
Jan 31 22:32:00
Jan 31 22:32:00 Searched in:
Jan 31 22:32:00 - '/usr/local/share/ntlk_data'
Jan 31 22:32:00 **********************************************************************
Jan 31 22:32:00 Traceback (most recent call last):
Jan 31 22:32:00 File "/usr/local/lib/python3.10/dist-packages/nltk/corpus/util.py", line 84, in __load
Jan 31 22:32:00 root = nltk.data.find(f"{self.subdir}/{zip_name}")
Jan 31 22:32:00 File "/usr/local/lib/python3.10/dist-packages/nltk/data.py", line 583, in find
Jan 31 22:32:00 raise LookupError(resource_not_found)
Jan 31 22:32:00 LookupError:
Jan 31 22:32:00 **********************************************************************
Jan 31 22:32:00 Resource stopwords not found.
Jan 31 22:32:00 Please use the NLTK Downloader to obtain the resource:
Jan 31 22:32:00
Jan 31 22:32:00 >>> import nltk
Jan 31 22:32:00 >>> nltk.download('stopwords')
Jan 31 22:32:00
Jan 31 22:32:00 For more information see: https://www.nltk.org/data.html
Jan 31 22:32:00
Jan 31 22:32:00 Attempted to load corpora/stopwords.zip/stopwords/
Jan 31 22:32:00
Jan 31 22:32:00 Searched in:
Jan 31 22:32:00 - '/usr/local/share/ntlk_data'
Jan 31 22:32:00 **********************************************************************
Jan 31 22:32:00
Jan 31 22:32:00
Jan 31 22:32:00 During handling of the above exception, another exception occurred:
Jan 31 22:32:00
Jan 31 22:32:00 Traceback (most recent call last):
Jan 31 22:32:00 File "/app/code/src/documents/consumer.py", line 410, in try_consume_file
Jan 31 22:32:00 document_consumption_finished.send(
Jan 31 22:32:00 File "/usr/local/lib/python3.10/dist-packages/django/dispatch/dispatcher.py", line 176, in send
Jan 31 22:32:00 return [
Jan 31 22:32:00 File "/usr/local/lib/python3.10/dist-packages/django/dispatch/dispatcher.py", line 177, in <listcomp>
Jan 31 22:32:00 (receiver, receiver(signal=self, sender=sender, **named))
Jan 31 22:32:00 File "/app/code/src/documents/signals/handlers.py", line 194, in set_tags
Jan 31 22:32:00 matched_tags = matching.match_tags(document, classifier)
Jan 31 22:32:00 File "/app/code/src/documents/matching.py", line 50, in match_tags
Jan 31 22:32:00 predicted_tag_ids = classifier.predict_tags(document.content)
Jan 31 22:32:00 File "/app/code/src/documents/classifier.py", line 370, in predict_tags
Jan 31 22:32:00 X = self.data_vectorizer.transform([self.preprocess_content(content)])
Jan 31 22:32:00 File "/app/code/src/documents/classifier.py", line 331, in preprocess_content
Jan 31 22:32:00 self._stop_words = set(stopwords.words(settings.NLTK_LANGUAGE))
Jan 31 22:32:00 File "/usr/local/lib/python3.10/dist-packages/nltk/corpus/util.py", line 121, in __getattr__
Jan 31 22:32:00 self.__load()
Jan 31 22:32:00 File "/usr/local/lib/python3.10/dist-packages/nltk/corpus/util.py", line 86, in __load
Jan 31 22:32:00 raise e
Jan 31 22:32:00 File "/usr/local/lib/python3.10/dist-packages/nltk/corpus/util.py", line 81, in __load
Jan 31 22:32:00 root = nltk.data.find(f"{self.subdir}/{self.__name}")
Jan 31 22:32:00 File "/usr/local/lib/python3.10/dist-packages/nltk/data.py", line 583, in find
Jan 31 22:32:00 raise LookupError(resource_not_found)
Jan 31 22:32:00 LookupError:
Jan 31 22:32:00 **********************************************************************
Jan 31 22:32:00 Resource stopwords not found.
Jan 31 22:32:00 Please use the NLTK Downloader to obtain the resource:
Jan 31 22:32:00
Jan 31 22:32:00 >>> import nltk
Jan 31 22:32:00 >>> nltk.download('stopwords')
Jan 31 22:32:00
Jan 31 22:32:00 For more information see: https://www.nltk.org/data.html
Jan 31 22:32:00
Jan 31 22:32:00 Attempted to load corpora/stopwords
Jan 31 22:32:00
Jan 31 22:32:00 Searched in:
Jan 31 22:32:00 - '/usr/local/share/ntlk_data'
Jan 31 22:32:00 **********************************************************************
Jan 31 22:32:00
Jan 31 22:32:00 [2023-01-31 21:32:00,915] [ERROR] [celery.app.trace] Task documents.tasks.consume_file[3e572bd2-9ccc-4ee4-9201-fb350b47cfd9] raised unexpected: ConsumerError("Doc - May 25, 2014, 11-08 AM.pdf: The following error occurred while consuming Doc - May 25, 2014, 11-08 AM.pdf: \n**********************************************************************\n Resource \x1b[93mstopwords\x1b[0m not found.\n Please use the NLTK Downloader to obtain the resource:\n\n \x1b[31m>>> import nltk\n >>> nltk.download('stopwords')\n \x1b[0m\n For more information see: https://www.nltk.org/data.html\n\n Attempted to load \x1b[93mcorpora/stopwords\x1b[0m\n\n Searched in:\n - '/usr/local/share/ntlk_data'\n**********************************************************************\n")
Jan 31 22:32:00 Traceback (most recent call last):
Jan 31 22:32:00 File "/usr/local/lib/python3.10/dist-packages/nltk/corpus/util.py", line 84, in __load
Jan 31 22:32:00 root = nltk.data.find(f"{self.subdir}/{zip_name}")
Jan 31 22:32:00 File "/usr/local/lib/python3.10/dist-packages/nltk/data.py", line 583, in find
Jan 31 22:32:00 raise LookupError(resource_not_found)
Jan 31 22:32:00 LookupError:
Jan 31 22:32:00 **********************************************************************
Jan 31 22:32:00 Resource stopwords not found.
Jan 31 22:32:00 Please use the NLTK Downloader to obtain the resource:
Jan 31 22:32:00
Jan 31 22:32:00 >>> import nltk
Jan 31 22:32:00 >>> nltk.download('stopwords')
Jan 31 22:32:00
Jan 31 22:32:00 For more information see: https://www.nltk.org/data.html
Jan 31 22:32:00
Jan 31 22:32:00 Attempted to load corpora/stopwords.zip/stopwords/
Jan 31 22:32:00
Jan 31 22:32:00 Searched in:
Jan 31 22:32:00 - '/usr/local/share/ntlk_data'
Jan 31 22:32:00 **********************************************************************
Jan 31 22:32:00
Jan 31 22:32:00
Jan 31 22:32:00 During handling of the above exception, another exception occurred:
Jan 31 22:32:00
Jan 31 22:32:00 Traceback (most recent call last):
Jan 31 22:32:00 File "/usr/local/lib/python3.10/dist-packages/asgiref/sync.py", line 302, in main_wrap
Jan 31 22:32:00 raise exc_info[1]
Jan 31 22:32:00 File "/app/code/src/documents/consumer.py", line 410, in try_consume_file
Jan 31 22:32:00 document_consumption_finished.send(
Jan 31 22:32:00 File "/usr/local/lib/python3.10/dist-packages/django/dispatch/dispatcher.py", line 176, in send
Jan 31 22:32:00 return [
Jan 31 22:32:00 File "/usr/local/lib/python3.10/dist-packages/django/dispatch/dispatcher.py", line 177, in <listcomp>
Jan 31 22:32:00 (receiver, receiver(signal=self, sender=sender, **named))
Jan 31 22:32:00 File "/app/code/src/documents/signals/handlers.py", line 194, in set_tags
Jan 31 22:32:00 matched_tags = matching.match_tags(document, classifier)
Jan 31 22:32:00 File "/app/code/src/documents/matching.py", line 50, in match_tags
Jan 31 22:32:00 predicted_tag_ids = classifier.predict_tags(document.content)
Jan 31 22:32:00 File "/app/code/src/documents/classifier.py", line 370, in predict_tags
Jan 31 22:32:00 X = self.data_vectorizer.transform([self.preprocess_content(content)])
Jan 31 22:32:00 File "/app/code/src/documents/classifier.py", line 331, in preprocess_content
Jan 31 22:32:00 self._stop_words = set(stopwords.words(settings.NLTK_LANGUAGE))
Jan 31 22:32:00 File "/usr/local/lib/python3.10/dist-packages/nltk/corpus/util.py", line 121, in __getattr__
Jan 31 22:32:00 self.__load()
Jan 31 22:32:00 File "/usr/local/lib/python3.10/dist-packages/nltk/corpus/util.py", line 86, in __load
Jan 31 22:32:00 raise e
Jan 31 22:32:00 File "/usr/local/lib/python3.10/dist-packages/nltk/corpus/util.py", line 81, in __load
Jan 31 22:32:00 root = nltk.data.find(f"{self.subdir}/{self.__name}")
Jan 31 22:32:00 File "/usr/local/lib/python3.10/dist-packages/nltk/data.py", line 583, in find
Jan 31 22:32:00 raise LookupError(resource_not_found)
Jan 31 22:32:00 LookupError:
Jan 31 22:32:00 **********************************************************************
Jan 31 22:32:00 Resource stopwords not found.
Jan 31 22:32:00 Please use the NLTK Downloader to obtain the resource:
Jan 31 22:32:00
Jan 31 22:32:00 >>> import nltk
Jan 31 22:32:00 >>> nltk.download('stopwords')
Jan 31 22:32:00
Jan 31 22:32:00 For more information see: https://www.nltk.org/data.html
Jan 31 22:32:00
Jan 31 22:32:00 Attempted to load corpora/stopwords
Jan 31 22:32:00
Jan 31 22:32:00 Searched in:
Jan 31 22:32:00 - '/usr/local/share/ntlk_data'
Jan 31 22:32:00 **********************************************************************
Jan 31 22:32:00
Jan 31 22:32:00
Jan 31 22:32:00 The above exception was the direct cause of the following exception:
Jan 31 22:32:00
Jan 31 22:32:00 Traceback (most recent call last):
Jan 31 22:32:00 File "/usr/local/lib/python3.10/dist-packages/celery/app/trace.py", line 451, in trace_task
Jan 31 22:32:00 R = retval = fun(*args, **kwargs)
Jan 31 22:32:00 File "/usr/local/lib/python3.10/dist-packages/celery/app/trace.py", line 734, in __protected_call__
Jan 31 22:32:00 return self.run(*args, **kwargs)
Jan 31 22:32:00 File "/app/code/src/documents/tasks.py", line 192, in consume_file
Jan 31 22:32:00 document = Consumer().try_consume_file(
Jan 31 22:32:00 File "/app/code/src/documents/consumer.py", line 468, in try_consume_file
Jan 31 22:32:00 self._fail(
Jan 31 22:32:00 File "/app/code/src/documents/consumer.py", line 93, in _fail
Jan 31 22:32:00 raise ConsumerError(f"{self.filename}: {log_message or message}") from exception
Jan 31 22:32:00 documents.consumer.ConsumerError: Doc - May 25, 2014, 11-08 AM.pdf: The following error occurred while consuming Doc - May 25, 2014, 11-08 AM.pdf:
Jan 31 22:32:00 **********************************************************************
Jan 31 22:32:00 Resource stopwords not found.
Jan 31 22:32:00 Please use the NLTK Downloader to obtain the resource:
Jan 31 22:32:00
Jan 31 22:32:00 >>> import nltk
Jan 31 22:32:00 >>> nltk.download('stopwords')
Jan 31 22:32:00
Jan 31 22:32:00 For more information see: https://www.nltk.org/data.html
Jan 31 22:32:00
Jan 31 22:32:00 Attempted to load corpora/stopwords
Jan 31 22:32:00
Jan 31 22:32:00 Searched in:
Jan 31 22:32:00 - '/usr/local/share/ntlk_data'
Jan 31 22:32:00 **********************************************************************
Jan 31 22:32:00

[girish]

@scooke said in Latest update seems to have similar issue as before, resources not found:

Jan 31 22:32:00 Attempted to load corpora/stopwords.zip/stopwords/

So, this is wrong. It should be corpora/stopwords/ , I think.

root@07a7fcf3-f9ff-4051-9d89-dec6ed4a4777:/usr/local/share/nltk_data/corpora/stopwords# ls
README       basque   chinese  english  german  hinglish    italian  norwegian   russian  swedish
arabic       bengali  danish   finnish  greek   hungarian   kazakh   portuguese  slovene  tajik
azerbaijani  catalan  dutch    french   hebrew  indonesian  nepali   romanian    spanish  turkish

[girish]

@scooke do you even have /usr/local/share/nltk_data/corpora/stopwords ?

[scooke]

@girish Although there is definitely a stopwords.zip in the corpora/stopwords directory, there is also a stopwords folder (unzipped, I suppose). That is one thing to correct then.

I found this link: https://aur.archlinux.org/packages/paperless-ngx. Six comments from the bottom ammo wrote that executing ln -s /usr/share/nltk_data /usr/local/share/nltk_data fixed it for him. I tried it but, of course, these are read-only directories. So maybe this is a second thing to fix?

[scooke]

@girish said in Latest update seems to have similar issue as before, resources not found:

/usr/local/share/nltk_data/corpora/stopwords

Yep

root@mypaperlessimagelongnumbernamethingy:/usr/local/share/nltk_data/corpora/stopwords# ls -al
total 160
drwxr-xr-x 2 root root  4096 Jan 26 09:05 .
drwxr-xr-x 3 root root  4096 Jan 26 09:05 ..
-rw-r--r-- 1 root root   909 Jan 26 09:05 README
-rw-r--r-- 1 root root  6348 Jan 26 09:05 arabic
-rw-r--r-- 1 root root   967 Jan 26 09:05 azerbaijani
-rw-r--r-- 1 root root  2202 Jan 26 09:05 basque
-rw-r--r-- 1 root root  5443 Jan 26 09:05 bengali
-rw-r--r-- 1 root root  1558 Jan 26 09:05 catalan
-rw-r--r-- 1 root root  5560 Jan 26 09:05 chinese
-rw-r--r-- 1 root root   424 Jan 26 09:05 danish
-rw-r--r-- 1 root root   453 Jan 26 09:05 dutch
-rw-r--r-- 1 root root   936 Jan 26 09:05 english
-rw-r--r-- 1 root root  1579 Jan 26 09:05 finnish
-rw-r--r-- 1 root root   813 Jan 26 09:05 french
-rw-r--r-- 1 root root  1362 Jan 26 09:05 german
-rw-r--r-- 1 root root  2167 Jan 26 09:05 greek
-rw-r--r-- 1 root root  1836 Jan 26 09:05 hebrew
-rw-r--r-- 1 root root  5958 Jan 26 09:05 hinglish
-rw-r--r-- 1 root root  1227 Jan 26 09:05 hungarian
-rw-r--r-- 1 root root  6446 Jan 26 09:05 indonesian
-rw-r--r-- 1 root root  1654 Jan 26 09:05 italian
-rw-r--r-- 1 root root  3880 Jan 26 09:05 kazakh
-rw-r--r-- 1 root root  3610 Jan 26 09:05 nepali
-rw-r--r-- 1 root root   851 Jan 26 09:05 norwegian
-rw-r--r-- 1 root root  1286 Jan 26 09:05 portuguese
-rw-r--r-- 1 root root  1910 Jan 26 09:05 romanian
-rw-r--r-- 1 root root  1235 Jan 26 09:05 russian
-rw-r--r-- 1 root root 15980 Jan 26 09:05 slovene
-rw-r--r-- 1 root root  2176 Jan 26 09:05 spanish
-rw-r--r-- 1 root root   559 Jan 26 09:05 swedish
-rw-r--r-- 1 root root  1818 Jan 26 09:05 tajik
-rw-r--r-- 1 root root   260 Jan 26 09:05 turkish

[scooke]

Thank you for troubleshooting this with me. Weird that I'm the only one!

[girish]

@scooke I think I am not testing this correctly because as per the code atleast it should fail for me but it doesn't.

Should I enable some flag inside paperless for nltk handling ? There is a complete absence of nltk in my logs.

[girish]

@scooke Can you try with 1.5.4 please?

I think I found the issue. First, the classifier needs to be created with document_create_classifier . One has to also put some tags, categories etc for documents. Once all that is setup, the nltk stuff kicks in. Without the classifier, nltk is just skipped.

[scooke]

@girish I"m not seeing v.1.5.4... is there a way to force it?

[axel0681]

Latest Update 1.5.4 is working for me .

Thanks

[girish]

@scooke said in Latest update seems to have similar issue as before, resources not found:

@girish I"m not seeing v.1.5.4... is there a way to force it?

Strange, it's published normally. Did you check for updates manually?

[scooke]

@girish Do I need to be upgraded to v1.5.2 to see it? Or do I need to have auto-updates turned on. I've turned them off and am checking manually as I'd rather have an older version that works than a newer one that doesn't.

[girish]

@scooke ah, that's correct. you have to update from 1.5.1 to 1.5.2 to 1.5.3 .

[scooke]

@girish I'm happy to report that once upgraded to v1.5.4 that the PDF upload and processing completes without any problems. Whatever you did worked! Thank you!