Spaces:
Sleeping
Sleeping
Upload 2 files
Browse files- app.py +80 -0
- requirements.txt +9 -0
app.py
ADDED
@@ -0,0 +1,80 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import gradio as gr
|
2 |
+
import easyocr
|
3 |
+
import cv2
|
4 |
+
import numpy as np
|
5 |
+
from PIL import Image
|
6 |
+
import pdf2image
|
7 |
+
import tempfile
|
8 |
+
from langchain_groq import ChatGroq
|
9 |
+
import re
|
10 |
+
|
11 |
+
# Initialize OCR reader
|
12 |
+
reader = easyocr.Reader(['en'])
|
13 |
+
|
14 |
+
# Initialize LLM
|
15 |
+
llm = ChatGroq(
|
16 |
+
temperature=0,
|
17 |
+
groq_api_key="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",
|
18 |
+
model_name="llama-3.3-70b-versatile"
|
19 |
+
)
|
20 |
+
|
21 |
+
# Utility to clean up unwanted characters
|
22 |
+
def clean_text(text):
|
23 |
+
text = re.sub(r"[*β’ββͺοΈβ¦β‘οΈ~]+", "", text) # remove bullet points and asterisks
|
24 |
+
text = re.sub(r"\s{2,}", " ", text) # remove excessive spacing
|
25 |
+
return text.strip()
|
26 |
+
|
27 |
+
def extract_text_and_summarize(file):
|
28 |
+
file_path = file.name
|
29 |
+
|
30 |
+
# If it's a PDF, convert to image
|
31 |
+
if file_path.lower().endswith(".pdf"):
|
32 |
+
images = pdf2image.convert_from_path(file_path)
|
33 |
+
image = np.array(images[0])
|
34 |
+
else:
|
35 |
+
image = cv2.imread(file_path)
|
36 |
+
|
37 |
+
# OCR
|
38 |
+
results = reader.readtext(image)
|
39 |
+
extracted_text = ' '.join([text[1] for text in results])
|
40 |
+
extracted_text = clean_text(extracted_text)
|
41 |
+
|
42 |
+
if not extracted_text.strip():
|
43 |
+
return "No readable text found.", ""
|
44 |
+
|
45 |
+
# LLM summarization
|
46 |
+
messages = [
|
47 |
+
{"role": "system", "content": "Your job is to summarize the given research paper and list its key sub-domains and topics clearly."},
|
48 |
+
{"role": "user", "content": extracted_text}
|
49 |
+
]
|
50 |
+
result = llm.invoke(messages)
|
51 |
+
|
52 |
+
summarized_text = clean_text(result.content)
|
53 |
+
return extracted_text, summarized_text
|
54 |
+
|
55 |
+
# Build Gradio UI
|
56 |
+
with gr.Blocks(title="Research Paper Summarizer") as iface:
|
57 |
+
gr.Markdown(
|
58 |
+
"""
|
59 |
+
# π§ Research Paper Summarizer
|
60 |
+
Upload an **image** or **PDF** of a research paper. This app will:
|
61 |
+
- Extract text using OCR
|
62 |
+
- Summarize the content
|
63 |
+
- List key subdomains and research topics
|
64 |
+
|
65 |
+
β‘ Powered by EasyOCR & LLaMA-3 via Groq
|
66 |
+
"""
|
67 |
+
)
|
68 |
+
|
69 |
+
with gr.Row():
|
70 |
+
file_input = gr.File(label="π Upload Research Paper (Image or PDF)", file_types=[".png", ".jpg", ".jpeg", ".pdf"])
|
71 |
+
|
72 |
+
with gr.Row():
|
73 |
+
extracted_box = gr.Textbox(label="π Extracted Text", lines=10, interactive=False)
|
74 |
+
summary_box = gr.Textbox(label="π Summarized Topics & Subdomains", lines=10, interactive=False)
|
75 |
+
|
76 |
+
run_button = gr.Button("π Run Summarizer")
|
77 |
+
|
78 |
+
run_button.click(fn=extract_text_and_summarize, inputs=file_input, outputs=[extracted_box, summary_box])
|
79 |
+
|
80 |
+
iface.launch(share=True)
|
requirements.txt
ADDED
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
gradio
|
2 |
+
easyocr
|
3 |
+
cv2
|
4 |
+
numpy
|
5 |
+
PIL
|
6 |
+
pdf2imag
|
7 |
+
tempfile
|
8 |
+
langchain_groq
|
9 |
+
re
|