Spaces:
Runtime error
Runtime error
arousrihab
commited on
Commit
•
8f3023d
1
Parent(s):
616a701
Create app.py
Browse files
app.py
ADDED
@@ -0,0 +1,75 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
# app.py
|
2 |
+
import streamlit as st
|
3 |
+
from extractive import preprocess_text, get_sentence_embeddings, build_semantic_graph, apply_textrank, generate_summary
|
4 |
+
from abstractive import abstractive_summary
|
5 |
+
from utils import extract_named_entities
|
6 |
+
from transformers import AutoTokenizer, AutoModel
|
7 |
+
|
8 |
+
# Load pre-trained BERT model and tokenizer
|
9 |
+
model_name = "dmis-lab/biobert-base-cased-v1.2"
|
10 |
+
tokenizer = AutoTokenizer.from_pretrained(model_name)
|
11 |
+
model = AutoModel.from_pretrained(model_name)
|
12 |
+
|
13 |
+
# Streamlit app layout
|
14 |
+
st.title("Hybrid Summarization App")
|
15 |
+
st.write("Upload text files for multi-document summarization or enter text manually for single-document summarization.")
|
16 |
+
|
17 |
+
# Multi-document summarization
|
18 |
+
st.header("Multi-Document Summarization")
|
19 |
+
uploaded_files = st.file_uploader("Upload text files", type="txt", accept_multiple_files=True)
|
20 |
+
|
21 |
+
if uploaded_files:
|
22 |
+
texts = [file.read().decode("utf-8") for file in uploaded_files]
|
23 |
+
|
24 |
+
# Perform extractive summarization for each document
|
25 |
+
extractive_summaries = []
|
26 |
+
for text in texts:
|
27 |
+
sentences = preprocess_text(text)
|
28 |
+
embeddings = get_sentence_embeddings(sentences, model, tokenizer)
|
29 |
+
graph = build_semantic_graph(embeddings)
|
30 |
+
ranked_sentences = apply_textrank(graph, sentences)
|
31 |
+
ext_summary = generate_summary(ranked_sentences, sentences, max_length_ratio=0.5)
|
32 |
+
extractive_summaries.append(ext_summary)
|
33 |
+
|
34 |
+
# Combine extractive summaries for multi-document summarization
|
35 |
+
combined_extractive_summary = " ".join(extractive_summaries)
|
36 |
+
st.write("Combined Extractive Summary:", combined_extractive_summary)
|
37 |
+
|
38 |
+
# Extract named entities from the combined summary
|
39 |
+
entities = extract_named_entities(combined_extractive_summary)
|
40 |
+
st.write("Named Entities:", entities)
|
41 |
+
|
42 |
+
# Choose summary length ratio for abstractive summarization
|
43 |
+
abs_ratio_option = st.selectbox("Choose abstractive summary length ratio", ("1/2", "1/3", "1/4"))
|
44 |
+
abs_ratio = {"1/2": 0.5, "1/3": 0.33, "1/4": 0.25}[abs_ratio_option]
|
45 |
+
|
46 |
+
# Perform abstractive summarization
|
47 |
+
combined_input = combined_extractive_summary + " " + ' '.join([ent[0] for ent in entities])
|
48 |
+
abs_summary = abstractive_summary(combined_input, max_length_ratio=abs_ratio, min_length_ratio=abs_ratio/2)
|
49 |
+
st.write("Abstractive Summary:", abs_summary)
|
50 |
+
|
51 |
+
# Single-document summarization
|
52 |
+
st.header("Single-Document Summarization")
|
53 |
+
text_input = st.text_area("Enter text here")
|
54 |
+
|
55 |
+
if text_input:
|
56 |
+
# Extract named entities
|
57 |
+
entities = extract_named_entities(text_input)
|
58 |
+
st.write("Named Entities:", entities)
|
59 |
+
|
60 |
+
# Perform extractive summarization
|
61 |
+
sentences = preprocess_text(text_input)
|
62 |
+
embeddings = get_sentence_embeddings(sentences, model, tokenizer)
|
63 |
+
graph = build_semantic_graph(embeddings)
|
64 |
+
ranked_sentences = apply_textrank(graph, sentences)
|
65 |
+
ext_summary = generate_summary(ranked_sentences, sentences, max_length_ratio=0.5)
|
66 |
+
st.write("Extractive Summary:", ext_summary)
|
67 |
+
|
68 |
+
# Choose summary length ratio for abstractive summarization
|
69 |
+
abs_ratio_option = st.selectbox("Choose abstractive summary length ratio", ("1/2", "1/3", "1/4"))
|
70 |
+
abs_ratio = {"1/2": 0.5, "1/3": 0.33, "1/4": 0.25}[abs_ratio_option]
|
71 |
+
|
72 |
+
# Perform abstractive summarization
|
73 |
+
combined_input = ext_summary + " " + ' '.join([ent[0] for ent in entities])
|
74 |
+
abs_summary = abstractive_summary(combined_input, max_length_ratio=abs_ratio, min_length_ratio=abs_ratio/2)
|
75 |
+
st.write("Abstractive Summary:", abs_summary)
|