inpho/topic-explorer

NoOptionError after upgrade

oguzhanalasehir opened this issue · 6 comments

Hello,
I'm using Windows 7 and Python 2.7.14 :: Anaconda.
I have just updated topicexplorer version as 1.0b214-1-g06a83c5 and later on I got the following error while importing corpus ( from corpus import *);

NoOptionError Traceback (most recent call last)
in ()
----> 1 from corpus import *

C:\Users\ENF_RP\Desktop\deneme\notebooks\corpus.py in ()
25 else:
26 c = Corpus.load(config.get('main', 'corpus_file'))
---> 27 context_type = config.get('main', 'context_type')
28 ctx_metadata = c.view_metadata(context_type)
29 all_ids = ctx_metadata[doc_label_name(context_type)]

C:\Users\ENF_RP\Anaconda2\lib\site-packages\backports\configparser_init_.pyc in get(self, section, option, **kwargs)
801 except KeyError:
802 if fallback is _UNSET:
--> 803 raise NoOptionError(option, section)
804 else:
805 return fallback

NoOptionError: No option 'context_type' in section: 'main'

Could you please help me about the issue?

Best regards,

Would it be possible to include the .ini file associated with your notebook instance?

When there is not a context_type, often that indicates that the LDA models have not yet been trained.

Regardless, it should not error in this user-unfriendly way, so a more helpful message will be included :)

I attached the ini file
pure_terms.zip

Btw,I checked training process again and I realized that it produced a warning as given below.

Traceback (most recent call last):
File "C:\Users\ENF_RP\Anaconda2\Scripts\topicexplorer-script.py", line 11, in

load_entry_point('topicexplorer', 'console_scripts', 'topicexplorer')()
File "c:\windows\system32\topicexplorer\topicexplorer_main_.py", line 243,
in main
benchmark(train.main)(args)
File "c:\windows\system32\topicexplorer\topicexplorer\train.py", line 243, in
main
os.remove(cluster_path)
TypeError: remove() argument 1 must be string, not None

The error has been fixed in github, but is not yet in a release. I should be able to push that in the next 2 hours.

Also, because of that error, the train hadn't been working, so I also pushed a change to that.

The issues should be fixed in the releases as well now

Thank u Jaimie, it is OK.

Best regards