demo_time_plot.py

import re

import scattertext as st
import spacy

nlp = spacy.blank('en')
nlp.add_pipe('sentencizer')

headline_df = st.SampleCorpora.GuardianHeadlines.get_data().assign(
    Parse=lambda df: df.Text.progress_apply(nlp),
    MonthNum=lambda df: df.Date.apply(lambda x: x.month),
    Month=lambda df: df.Date.apply(lambda x: x.strftime("%B-%Y")),
    DateStr=lambda df: df.Date.apply(lambda x: x.strftime("%Y-%m-%d")),
)


#def exclude_ngrams_which_do_not_start_and_end_with_function_words(ngram: spacy.tokens.Span) -> bool:
#    return len(set([ngram[0].pos_, ngram[-1].pos_]) \
#              & {"DET", "AUX", "ADP", "AUX", "PRON", "PUNCT", 'CCONJ', 'PART'}) > 0
word_number_matcher = re.compile('^[A-Za-z0-9 ]+$')

def exclude_ngrams_which_do_not_start_and_end_with_function_words(ngram: spacy.tokens.Span) -> bool:
    return any([ngram[0].lower_.strip() in st.MY_ENGLISH_STOP_WORDS,
                ngram[-1].lower_.strip() in st.MY_ENGLISH_STOP_WORDS,
                word_number_matcher.match(ngram[0].lower_.strip()) is None,
                word_number_matcher.match(ngram[-1].lower_.strip()) is None])


corpus = st.OffsetCorpusFactory(
    headline_df,
    category_col='DateStr',
    parsed_col='Parse',
    feat_and_offset_getter=st.FlexibleNGramFeatures(
        ngram_sizes=[1, 2, 3, 4, 5],
        exclude_ngram_filter=exclude_ngrams_which_do_not_start_and_end_with_function_words
    )
).build().compact(
    compactor=st.NPMICompactor(
        minimum_term_count=3,
        number_terms_per_length=2000,
    ),
    non_text=True
).compact(
    st.NgramPercentageCompactor(
        usage_portion=0.6,
    ),
    non_text=True
).filter_out(
    lambda x: len(x) == 1,
    non_text=True
).compact(
    compactor=st.AssociationCompactor(
        2000,
        scorer=st.DeltaJSDivergenceScorer,
        term_ranker=st.OncePerDocFrequencyRanker,
        use_non_text_features=True
    ),
    non_text=True
)



category_order = list(sorted(corpus.get_categories()))
heading_categories, heading_category_order = st.CharacteristicGrouper(
    corpus,
    non_text=True,
    rank_embedder=st.RankEmbedder(
        term_scorer=st.DeltaJSDivergenceScorer(corpus),
        rank_threshold=10
    ),
    window_size=1,
    to_text=' to '
).get_new_doc_categories(
    number_of_splits=5,
    category_order=category_order
)


order = list(sorted(corpus.get_categories()))

html = st.produce_scattertext_table(
    corpus=corpus,
    heading_categories=heading_categories,
    heading_category_order=heading_category_order,
    category_order=category_order,
    all_category_scorer=lambda c: st.AllCategoryTermScorer(c, term_scorer=st.DeltaJSDivergenceScorer),
    metadata=lambda c: c.get_df()['Date'].astype(str),
    ignore_categories=False,
    plot_width=1000,
    plot_height=400,
    top_terms_length=5,
    sort_doc_labels_by_name=True,
    use_offsets=True,
    non_text=True,
    trend_plot_settings=st.TimePlotSettings(
        category_order=category_order,
        dispersion_metric='DA',
        dispersion_scaler=st.Scalers.dense_rank
    )
)

fn = 'demo_time_plot_da.html'
with open(fn, 'w') as of:
    of.write(html)

print('Run ./open demo_time_plot_da.html')