Closed Doooooris closed 4 years ago
Hi @Doooooris .
Thanks for your feedback. Here is a response for all of them:
The only problem that we are facing now is getting code coverage above 90%. The code now is complex and handles lots of cases, such as when pre-trained model doesn't exist or some dependencies have not been installed. We think that some cases are not possible to handle so, the capacity of the code coverage (maximum that we can cover) is not 100% but less. Is there any possibility to reduce this threshold that was written in milestone 3?
For all 4 functions: what if the input is not data frame (string or number or array)? or it's a data frame without column name? Try to image the craze user's input.
When I run code : df = pd.DataFrame({'text_col' : ['whatever']}); get_part_of_speech(df); the result gives me: adjective noun verb 0 0.0 1.0 0.0; Basically, whatever my input is, the output always 0 1 0. Not sure if it's the problem of my local library? Please message me a screen shot of your result.
Test function of all 4 functions: What if the input is not a language? eg. '12 4 57. 243434 dfef'. What if the input is not a data frame (string or number or array)?
Test function of part_of_speech: The tested output type is restricted to ['adjective', 'noun', 'verb']. If you can try different outputs like prepositions, articles would be great.
When I try to run detect_language.py locally, the error occurs; ValueError: model/lid.176.bin cannot be opened for loading! Seems like the package does not have pretrained model.
Test the format of the output is as expected too.