|
import json |
|
|
|
import streamlit as st |
|
from pyserini.search.lucene import LuceneSearcher |
|
|
|
|
|
searcher = LuceneSearcher("indexes/docs") |
|
searcher.set_language("ja") |
|
|
|
|
|
st.set_page_config( |
|
page_title="awesome-japanese-nlp-resources-search", |
|
page_icon="😎", |
|
layout="centered", |
|
) |
|
|
|
|
|
cola, colb, colc = st.columns([5, 4, 5]) |
|
|
|
st.header("awesome-japanese-nlp-resources-search 😎") |
|
st.markdown( |
|
"You can search for open-source software from [400+ Japanese NLP" |
|
" repositories](https://github.com/taishi-i/awesome-japanese-nlp-resources)." |
|
) |
|
|
|
col1, col2 = st.columns([9, 1]) |
|
with col1: |
|
search_query = st.text_input(label="", placeholder="Search") |
|
|
|
with col2: |
|
st.write("#") |
|
button_clicked = st.button("🔎") |
|
|
|
|
|
if search_query or button_clicked: |
|
search_results = searcher.search(search_query, k=100_000) |
|
|
|
st.write( |
|
'<p align="light" style="color:grey;">' |
|
f" {len(search_results):,.0f} repositories </p>", |
|
unsafe_allow_html=True, |
|
) |
|
|
|
for result in search_results: |
|
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"] |
|
|
|
try: |
|
st.markdown(f"### [{project_name}]({url})") |
|
st.markdown(f"{description}") |
|
if sub_topic is None: |
|
st.text(f"{main_topic}") |
|
else: |
|
st.text(f"{main_topic} / {sub_topic}") |
|
st.markdown("") |
|
st.markdown("") |
|
|
|
except: |
|
pass |
|
|
