-
Notifications
You must be signed in to change notification settings - Fork 68
/
Copy pathsetup.py
30 lines (24 loc) · 1.01 KB
/
setup.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
# -*- coding=utf-8 -*-
from setuptools import setup
# version 0.1.0: 使用 jieba 分词,利用 word 级别 tfidf 计算
# version 0.2.0: 使用 pkuseg 分词,利用 word 级别 tfidf 计算,增加短语过滤规则
# version 0.3.0: 使用 lda 主题模型,同时利用 word 级别 tfidf 信息,增加短语过滤规则
# version 0.4.0: 增加依据词典,抽取特定短语抽取接口、过滤接口。
# reference: Teneva N , Cheng W . Salience Rank: Efficient Keyphrase Extraction with Topic Modeling[C]// Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers). 2017.
# setup
setup(name="ckpe",
version="0.4.0",
url="https://github.com/dongrixinyu/chinese_keyphrase_extractor",
author="dongrixinyu",
author_email="dongrixinyu.89@163.com",
py_modules=[],
packages=[
"ckpe",
],
include_package_data=True,
install_requires=[
"pkuseg"
],
entry_points={
},
)