| import json | |
| import streamlit as st | |
| from pyserini.search.lucene import LuceneSearcher | |
| class SearchApplication: | |
| def __init__(self): | |
| self.title = "Awesome Japanese NLP resources search" | |
| self.set_page_config() | |
| self.searcher = self.set_searcher() | |
| st.header(self.title) | |
| col1, col2 = st.columns(2) | |
| with col1: | |
| self.query = st.text_input( | |
| "Search English or Japanese words", value="" | |
| ) | |
| with col2: | |
| st.write("#") | |
| self.search_button = st.button("🔎") | |
| st.caption( | |
| "You can search for open-source software from [400+ Japanese NLP" | |
| " repositories](https://github.com/taishi-i/awesome-japanese-nlp-resources)." | |
| ) | |
| st.write("#") | |
| self.show_popular_words() | |
| self.show_search_results() | |
| def set_page_config(self): | |
| st.set_page_config( | |
| page_title=self.title, | |
| page_icon="😎", | |
| layout="centered", | |
| ) | |
| def set_searcher(self): | |
| searcher = LuceneSearcher("indexes/docs") | |
| searcher.set_language("ja") | |
| return searcher | |
| def show_popular_words(self): | |
| st.caption("Popular words") | |
| word1, word2, word3, word4, _ = st.columns(5) | |
| with word1: | |
| button1 = st.button("Python") | |
| if button1: | |
| self.query = "Python" | |
| with word2: | |
| button2 = st.button("ChatGPT") | |
| if button2: | |
| self.query = "ChatGPT" | |
| with word3: | |
| button3 = st.button("辞書") | |
| if button3: | |
| self.query = "辞書" | |
| with word4: | |
| button4 = st.button("コーパス") | |
| if button4: | |
| self.query = "Corpus" | |
| def show_search_results(self): | |
| if self.query or self.search_button: | |
| st.write("#") | |
| search_results = self.searcher.search(self.query, k=500) | |
| num_search_results = len(search_results) | |
| st.write(f"{num_search_results} results") | |
| for result in sorted( | |
| search_results, | |
| key=lambda x: json.loads(x.raw)["stargazers_count"], | |
| reverse=True | |
| ): | |
| data_json = json.loads(result.raw) | |
| description = data_json["description"] | |
| url = data_json["url"] | |
| project_name = data_json["project_name"] | |
| main_topic = data_json["main_topic"] | |
| sub_topic = data_json["sub_topic"] | |
| st.write("---") | |
| st.subheader(f"[{project_name}]({url})") | |
| st.markdown(description) | |
| if sub_topic is None: | |
| st.caption(f"{main_topic}") | |
| else: | |
| st.caption(f"{main_topic} / {sub_topic}") | |
| st.write("#") | |
| def main(): | |
| SearchApplication() | |
| if __name__ == "__main__": | |
| main() | |