Language Issue :"French" & clarifications about inputs about pke HOT 3 CLOSED

Hi @fatimalaoui,

The df file contains the document frequency counts (i.e. the number of documents that contain each word) and the model_file contains the parameter for the classifier.

pke only ships with pre-trained english models, so that is maybe why you are facing with issue. Nevertheless, can you give me a code snippet so I can reproduce your error?

f.

from pke.

Hi @boudinfl , Thank you for your reply.
Here is a code snippet that i use for english but that doesn't work also.
`#Kea keyphrase extraction model. Parameterized example::
import pke
from nltk.corpus import stopwords
from pke.supervised.feature_based.kea import Kea

define a list of stopwords
stoplist = stopwords.words('english')

1. create a Kea extractor.
   extractor = Kea()

2. load the content of the document.
   extractor.load_document(input='C-1.txt',
   language='./Programs/Python/Python37/Lib/site-packages/spacy/lang/en',
   normalization=None)

3. select 1-3 grams that do not start or end with a stopword as
   candidates. Candidates that contain punctuation marks as words
   are discarded.
   extractor.candidate_selection(stoplist=stoplist)

4. classify candidates as keyphrase or not keyphrase.
   df = pke.load_document_frequency_file(input_file='./Desktop/model.txt')
   model_file = './Desktop/df-semeval2010.tsv'
   extractor.candidate_weighting(self,model_file=model_file,df=df)

5. get the 10-highest scored candidates as keyphrases
   keyphrases = extractor.get_n_best(n=10)`

• For the df and model_file , i'm not sure what to put there exactly. (refering to the files that already exist in the package pke)

• Here is the spacy file i have in my python directory .. is there something missing maybe or is it installed correctly?
  

• The error i'm getting is the following:

Could not read meta.json from C:\Users\Fatima\AppData\Local\Programs\Python\Python37\Lib\site-packages\spacy\lang\en\meta.json

from pke.

The language parameter of load_document() should be set to en. I also think you have issues with the df and model files (df should be document frequency counts and model should be a sklearn model parameters file). You can simply try to use Kea with the standard models by simply calling extractor.candidate_weighting()

from pke.