Hi @JasonKessler ,

I have the same issue here and I could not solve it with your suggestion.
Basically, the following code is used:

data = data.loc[:, ['id', 'language', 'ProcessedText', 'OriginalText']]

from sklearn.decomposition import NMF
from sklearn.feature_extraction.text import TfidfTransformer
from sklearn.pipeline import Pipeline


data['parse'] = data['ProcessedText'].apply(st.whitespace_nlp_with_sentences)

unigram_corpus = (st.CorpusFromParsedDocuments(data,
                                               category_col='language',
                                               parsed_col='parse')
                  .build().get_stoplisted_unigram_corpus())



html = st.produce_scattertext_explorer(
            unigram_corpus,
            category='French', category_name='French', not_category_name='German',
            minimum_term_frequency=0, pmi_threshold_coefficient=0,
            width_in_pixels=1000, metadata=unigram_corpus.get_df()['language'],
            alternative_text_field = 'OriginalText' ,
            transform=st.Scalers.dense_rank
)

What I expect is to see fully the text from the OriginalText column after clicking on a given word in the chart. However at the moment I only see a chunk of such text.

For example, when clicking on the word 'thank', I would see something like the following:

Thank you!

When I expect to see the following instead:

This was a great moment. Thank you!

Basically, I do not want the chunking when search for a given word among my text column. Can we achieve that? 😄

how can I use the original text in the snippets after cleaning? #54

Description

Metadata

Assignees

Labels

Projects

Milestone

Relationships

Development

Issue actions