'''

test

'''

import json

import spacy

import pytextrank

from collections import defaultdict

nlp = spacy.load('en_core_web_sm')

# load

nlp = spacy.load("en_core_web_sm")

# add PyTextRank to the spaCy pipeline

tr = pytextrank.TextRank()

nlp.add_pipe(tr.PipelineComponent, name='textrank', last=True)

# method

def pytextrank_extract(free_text,topk=30):

    query_set = defaultdict(float)

    'textrank extraction'

    doc = nlp(free_text)

    for p in doc._.phrases:

        if len(p.text) > 5:

            query_set[p.text] = query_set[p.text] + p.rank

    ordered_query_set = [(k,v) for k, v in sorted(query_set.items(), key=lambda item: item[1],reverse=True)][:topk]

    result_list = []

    for query, score in ordered_query_set:

        # print(query,score)

        result_list.append(query)

    return result_list

# ordered_query_set = extract(test_free_text)

#

# # print out