chatbot_test / app.py
gamza's picture
Update app.py
d16378b
Raw History Blame Contribute Delete
1.32 kB
import gradio as gr
import pandas as pd
from sentence_transformers import SentenceTransformer
from sklearn.metrics.pairwise import cosine_similarity
title = "๐Ÿ€๊ณ ๋ฏผ ํ•ด๊ฒฐ ๋„์„œ ์ถ”์ฒœ ์ฑ—๋ด‡๐Ÿ€"
description = "๊ณ ๋ฏผ์ด ๋ฌด์—‡์ธ๊ฐ€์š”? ๊ณ ๋ฏผ ํ•ด๊ฒฐ์„ ๋„์™€์ค„ ์ฑ…์„ ์ถ”์ฒœํ•ด๋“œ๋ฆฝ๋‹ˆ๋‹ค"
examples = [["์š”์ฆ˜ ์ž ์ด ์•ˆ ์˜จ๋‹ค"], ["ํ™”๋ถ„์ด ์ž˜ ์ž๋ผ์ง€ ์•Š์•„"]]
# model = SentenceTransformer('jhgan/ko-sroberta-multitask')
df = pd.read_pickle('BookData_emb.pkl')
df_emb = df[['์„œํ‰์ž„๋ฒ ๋”ฉ']].copy()
def recommend(message):
answer = df.loc[df_emb['์„œํ‰์ž„๋ฒ ๋”ฉ'][0]]
# embedding = model.encode(message)
# df_emb['๊ฑฐ๋ฆฌ'] = df_emb['์„œํ‰์ž„๋ฒ ๋”ฉ'].map(lambda x: cosine_similarity([embedding], [x]).squeeze())
# answer = df.loc[df_emb['๊ฑฐ๋ฆฌ'].idxmax()]
# Book_title = answer['์ œ๋ชฉ']
# Book_author = answer['์ž‘๊ฐ€']
# Book_publisher = answer['์ถœํŒ์‚ฌ']
# Book_comment = answer['์„œํ‰']
return answer
gr.ChatInterface(
fn=recommend,
textbox=gr.Textbox(placeholder="๋ง๊ฑธ์–ด์ฃผ์„ธ์š”..", container=False, scale=7),
title=title,
description=description,
theme="soft",
examples=examples,
retry_btn="๋‹ค์‹œ๋ณด๋‚ด๊ธฐ โ†ฉ",
undo_btn="์ด์ „์ฑ— ์‚ญ์ œ โŒ",
clear_btn="์ „์ฑ— ์‚ญ์ œ ๐Ÿ’ซ").launch()