File size: 2,688 Bytes
5e7efef |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 |
import gradio as gr
import arxiv
from transformers import pipeline
from sentence_transformers import SentenceTransformer, util
qa_pipeline = pipeline("question-answering", model="distilbert-base-cased-distilled-squad")
embedder = SentenceTransformer("all-MiniLM-L6-v2")
def fetch_arxiv_papers(query, max_results=5):
search = arxiv.Search(
query=query,
max_results=max_results,
sort_by=arxiv.SortCriterion.Relevance
)
return list(search.results())
def generate_answer(question, context):
return qa_pipeline(question=question, context=context)["answer"]
def evaluate_retrieval(query, context):
query_embed = embedder.encode(query, convert_to_tensor=True)
context_embed = embedder.encode(context, convert_to_tensor=True)
similarity = util.pytorch_cos_sim(query_embed, context_embed)
score = float(similarity[0][0])
if score > 0.7:
return "High"
elif score > 0.4:
return "Medium"
else:
return "Low"
def evaluate_summary(context, answer):
if answer.lower() in context.lower():
return "Good"
elif len(answer.split()) > 5:
return "Fair"
else:
return "Poor"
def rag_pipeline(user_query):
papers = fetch_arxiv_papers(user_query)
if not papers:
return "No relevant papers found.", "", ""
output_blocks = []
for i, paper in enumerate(papers, 1):
title = paper.title
abstract = paper.summary
link = paper.entry_id
answer = generate_answer(user_query, abstract)
retrieval_score = evaluate_retrieval(user_query, abstract)
summary_score = evaluate_summary(abstract, answer)
block = f"""
### Paper {i}
Title: {title}
Abstract: {abstract}
Summarized Answer: {answer}
Link: {link}
Retrieval Accuracy: {retrieval_score}
Summary Quality: {summary_score}
---"""
output_blocks.append(block)
final_output = "\n".join(output_blocks)
return final_output, "Multiple Papers Shown", "See Above"
# Gradio UI
with gr.Blocks(theme=gr.themes.Soft(primary_hue="blue", secondary_hue="green")) as demo:
gr.Markdown("## Scientific Paper Discovery — RAG-Based QA & Evaluation")
gr.Markdown("Enter your research topic below to explore papers from arXiv and get summarized insights.")
query = gr.Textbox(label="Enter a research topic")
run_btn = gr.Button("Search and Summarize")
output = gr.Markdown(label="Output")
retrieval_accuracy = gr.Textbox(label="Retrieval Accuracy")
summary_quality = gr.Textbox(label="Summary Quality")
run_btn.click(fn=rag_pipeline, inputs=query, outputs=[output, retrieval_accuracy, summary_quality])
demo.launch()
|