期刊论文投稿推荐系统.py

from transformers import PegasusForConditionalGeneration
from tokenizers_pegasus import PegasusTokenizer
import re
import pandas as pd
import jieba
import jieba.analyse as jay
import paddle
from pyecharts import options as opts
from pyecharts.charts import Funnel


paddle.enable_static()
# -*- coding: utf-8 -*-

ARTICLE_PATH = '2.xlsx'
CACHE_PATH = ''
def get_keywords_in_article():
    def key_word(data):
        data = str(data).strip()
        result = re.findall(pattern='pan>(.*?)</t', string=data, flags=re.S)
        str_n = []
        for str_ in result:
            str_ = re.sub(r'[0-9a-zA-Z“”]', '', str_)
            if str_ == '':
                continue
            str_n.append(str_)
        return str_n

    excel_pd = pd.read_excel(ARTICLE_PATH, index_col=None, sheet_name=0)
    # keyword_excel_pd = excel_pd['字段1'].values.tolist()
    keyword_excel_list = []
    for fil in excel_pd['字段1'].values.tolist():
        keyword_excel_list.append(key_word(fil))
    keyword_excel_pd = pd.DataFrame(keyword_excel_list)

    excel_pd['字段1'] = excel_pd['字段1'].map(key_word)
    excel_pd.columns = ['期刊名称', '2022复合影响因子', '2022综合影响因子', '关键词']

    keyword_excel_new_pd = pd.concat([excel_pd, keyword_excel_pd.iloc[:, :20]], axis=1)
    columns_list = [
        '期刊名称',
        '2022复合影响因子',
        '2022综合影响因子',
        '关键词',
        '关键词1',
        '关键词2',
        '关键词3',
        '关键词4',
        '关键词5',
        '关键词6',
        '关键词7',
        '关键词8',
        '关键词9',
        '关键词10',
        '关键词11',
        '关键词12',
        '关键词13',
        '关键词14',
        '关键词15',
        '关键词16',
        '关键词17',
        '关键词18',
        '关键词19',
        '关键词20',
    ]
    keyword_excel_new_pd.columns = columns_list
    keyword_excel_new_pd.to_excel('3.xlsx')
    excel_pd_1 = excel_pd.iloc[:, 3:].values.tolist()
    new_list = [tokens for st in excel_pd_1 for tokens in st]
    new_list = [tokens for st in new_list for tokens in st]
    new_list = list(set(new_list))
    return new_list, keyword_excel_new_pd

def extract_tags(text):

    new_list, keyword_excel_new_pd = get_keywords_in_article()
    model = PegasusForConditionalGeneration.from_pretrained("IDEA-CCNL/Randeng-Pegasus-238M-Summary-Chinese")
    tokenizer = PegasusTokenizer.from_pretrained("IDEA-CCNL/Randeng-Pegasus-238M-Summary-Chinese")


    inputs = tokenizer(text, max_length=1024, return_tensors="pt")


    summary_ids = model.generate(inputs["input_ids"])
    data_decode = tokenizer.batch_decode(summary_ids, skip_special_tokens=True, clean_up_tokenization_spaces=False)[0]




    print(data_decode)
    for i in range(len(new_list)):
        jieba.add_word(new_list[i])
    cut_result_str = jieba.cut(data_decode, use_paddle=True)
    print("Paddle Mode: " + '/'.join(cut_result_str))
    tok_result = jay.extract_tags(data_decode, topK=1, withWeight=False)
    # print(tok_result[0])
    # key_word_search_result = keyword_excel_new_pd.loc[keyword_excel_new_pd['关键词2'].isin(['数字经济'])]
    # print(key_word_search_result)

    # empty_pd = pd.DataFrame()
    cache_path = '4.csv'
    # empty_list = []
    for i in range(1, 21):
        key_word_search_result = keyword_excel_new_pd.loc[keyword_excel_new_pd[f'关键词{i}'].isin([str(tok_result[0])])]
        key_word_search_result.to_csv(cache_path, mode='a+', columns=None)
    pro_pd = pd.read_csv(cache_path)

    pro_pd = pro_pd.dropna()
    pro_new_pd = pro_pd.iloc[:, 1:]
    pro_new_pd = pro_new_pd.drop_duplicates('2022综合影响因子')
    pro_new_pd['2022复合影响因子'] = pro_new_pd['2022复合影响因子'].str[13:].astype('float64')
    pro_new_pd.sort_values(by=['2022复合影响因子'], ascending=False, inplace=True)

    pro_new_pd.to_excel('4.xlsx')
    labels = []
    values_1 = []
    values_2 = []
    for index, row in pro_new_pd.iterrows():
        labels.append(row['期刊名称'])
        values_1.append(row['2022复合影响因子'])
        values_2.append(row['2022综合影响因子'].split('：')[1])
    # figure_get(labels, values_1)

    return [labels, values_1]

def figure_get(labels, values_1):

    c = (
        Funnel(
        )
        .add(
            '根据您的论文所推荐期刊',
            [list(z) for z in zip(labels[:10], values_1[:10])],
            label_opts=opts.LabelOpts(position="inside", is_show=True),
        )
        .set_global_opts(title_opts=opts.TitleOpts(title="根据您的论文所推荐的期刊"),
                         legend_opts=opts.LegendOpts(is_show=False),
                         toolbox_opts=opts.ToolboxOpts(is_show=False),
                         )
        .render("根据您的论文所推荐的期刊.html")
    )