Spaces:
Sleeping
Sleeping
Create app.py
Browse files
app.py
ADDED
|
@@ -0,0 +1,89 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import streamlit as st
|
| 2 |
+
from transformers import pipeline
|
| 3 |
+
|
| 4 |
+
|
| 5 |
+
@st.cache_resource # кэширование
|
| 6 |
+
def load_model():
|
| 7 |
+
return pipeline("text-classification", model="Wi/arxiv-distilbert-base-cased") # скачивание модели
|
| 8 |
+
|
| 9 |
+
model = load_model()
|
| 10 |
+
|
| 11 |
+
def top_pct(preds, threshold=.95):
|
| 12 |
+
preds = sorted(preds, key=lambda x: -x["score"])
|
| 13 |
+
|
| 14 |
+
cum_score = 0
|
| 15 |
+
for i, item in enumerate(preds):
|
| 16 |
+
cum_score += item["score"]
|
| 17 |
+
if cum_score >= threshold:
|
| 18 |
+
break
|
| 19 |
+
|
| 20 |
+
preds = preds[:(i+1)]
|
| 21 |
+
|
| 22 |
+
return preds
|
| 23 |
+
|
| 24 |
+
|
| 25 |
+
def format_predictions(preds) -> str:
|
| 26 |
+
"""
|
| 27 |
+
Prepare predictions and their scores for printing to the user
|
| 28 |
+
"""
|
| 29 |
+
out = ""
|
| 30 |
+
for i, item in enumerate(preds):
|
| 31 |
+
out += f"{i+1}. {item['label']} (score {item['score']:.2f})\n"
|
| 32 |
+
return out
|
| 33 |
+
|
| 34 |
+
|
| 35 |
+
st.markdown("""
|
| 36 |
+
<div style='text-align: center;'>
|
| 37 |
+
<img src='https://info.arxiv.org/brand/images/brand-logo-primary.jpg' alt='Centered Image' width='300'/>
|
| 38 |
+
</div>
|
| 39 |
+
""", unsafe_allow_html=True)
|
| 40 |
+
|
| 41 |
+
st.markdown("""
|
| 42 |
+
<h2 style='text-align: center; color: #e80ad8; font-family: Arial;'>
|
| 43 |
+
🚀 arXiv paper categories predictor
|
| 44 |
+
</h2>
|
| 45 |
+
""", unsafe_allow_html=True)
|
| 46 |
+
|
| 47 |
+
|
| 48 |
+
# CSS to change the background of the entire app
|
| 49 |
+
background_color_css = """
|
| 50 |
+
<style>
|
| 51 |
+
.stApp {
|
| 52 |
+
background-color: black; /* #eefcfa */
|
| 53 |
+
}
|
| 54 |
+
</style>
|
| 55 |
+
"""
|
| 56 |
+
|
| 57 |
+
st.markdown("""
|
| 58 |
+
<br><br> <!-- Adds vertical space -->
|
| 59 |
+
<p style='
|
| 60 |
+
color: white;
|
| 61 |
+
font-size: 20px;
|
| 62 |
+
font-family: "Courier New", monospace;
|
| 63 |
+
'>
|
| 64 |
+
Paste Title and Abstract of the paper and get most likely categories of the paper in the
|
| 65 |
+
<a href="https://arxiv.org/category_taxonomy" target="_blank" style="color: cyan; text-decoration: none;">
|
| 66 |
+
arXiv taxonomy
|
| 67 |
+
</a>
|
| 68 |
+
</p>
|
| 69 |
+
""", unsafe_allow_html=True)
|
| 70 |
+
|
| 71 |
+
title = st.text_input("Title", value="")
|
| 72 |
+
|
| 73 |
+
abstract = st.text_input("Abstract", value="")
|
| 74 |
+
|
| 75 |
+
query = title + '\n' + abstract
|
| 76 |
+
if query:
|
| 77 |
+
st.markdown("""
|
| 78 |
+
<br><br> <!-- Adds vertical space -->
|
| 79 |
+
<p style='
|
| 80 |
+
color: white;
|
| 81 |
+
font-size: 20px;
|
| 82 |
+
font-family: "Courier New", monospace;
|
| 83 |
+
'>
|
| 84 |
+
Most likely categories of the paper:
|
| 85 |
+
</p>
|
| 86 |
+
""", unsafe_allow_html=True)
|
| 87 |
+
result = format_predictions(top_pct(model(query)[0]))
|
| 88 |
+
|
| 89 |
+
st.write(result)
|