dhchenx / opinionx

A toolkit to extract opinions and useful information from text

Opinion Analysis Toolkit

A toolkit to extract opinions and useful information from text

Installation

pip install opinionx

Example Usage

1. Find opinions

from opinionx.text import get_opinion
text=open("test.txt",'r',encoding='utf-8').read()
opinion_words=['表示','认为','说','介绍','提出','透露','指出','强调','：']
list_opinion,_,_=get_opinion(text,lang='zh',opinion_words=opinion_words)
for opinion in list_opinion:
    print(opinion)

2. Find Leader's Opinions

from opinionx.text import get_leader_opinions

text=open("test.txt",'r',encoding='utf-8').read()

list_opinion = get_leader_opinions(text,save_path="", search_keywords_path="data/search_keywords.csv",leader_path="data/g20_leaders.csv")
print()
for opinion in list_opinion:
    print(opinion)
    print(opinion["opinion"])
    print(opinion["first_found_keyword"])
    print(opinion["first_found_leader"])
    print()

3. run tf-idf and tf models for massive text files

from opinionx.tfidf_shell import *
run_tfidf_shell(input_folder="tfidf_folder/raw_data", # a list of text files
                output_folder="tfidf_folder/output", # output folder
                user_dict_path="tfidf_folder/user_dictionaries", # the folder contains csv files with each line as a word
                font_path="utils/fonts/SimHei.ttf",# use it when analysis Chinese text
                is_html=True
                )

Credits & References

• Stanza
• jieba

License

The opinionx project is provided by Donghua Chen.