|
|
!pip install transformers accelerate torch PyPDF2 gradio --quiet
|
|
|
|
|
|
import gradio as gr
|
|
|
from transformers import pipeline
|
|
|
import PyPDF2
|
|
|
|
|
|
|
|
|
|
|
|
persian_summarizer = pipeline(
|
|
|
"summarization",
|
|
|
model="m3hrdadfi/bert2bert-fa-summarization",
|
|
|
tokenizer="m3hrdadfi/bert2bert-fa-summarization"
|
|
|
)
|
|
|
|
|
|
|
|
|
english_summarizer = pipeline(
|
|
|
"summarization",
|
|
|
model="facebook/bart-large-cnn"
|
|
|
)
|
|
|
|
|
|
|
|
|
def extract_text_from_pdf(pdf_file):
|
|
|
"""استخراج متن از فایل PDF"""
|
|
|
pdf_reader = PyPDF2.PdfReader(pdf_file)
|
|
|
text = ""
|
|
|
for page in pdf_reader.pages:
|
|
|
page_text = page.extract_text()
|
|
|
if page_text:
|
|
|
text += page_text + "\n"
|
|
|
return text
|
|
|
|
|
|
def summarize_document(language, text, pdf):
|
|
|
"""خلاصه‌سازی متن یا PDF بر اساس زبان انتخابی"""
|
|
|
if pdf is not None:
|
|
|
text = extract_text_from_pdf(pdf)
|
|
|
|
|
|
if not text or len(text.strip()) < 50:
|
|
|
return "❗ لطفاً یک متن یا فایل PDF معتبر و حداقل ۵۰ کاراکتر وارد کنید."
|
|
|
|
|
|
|
|
|
if language == "Persian":
|
|
|
summarizer = persian_summarizer
|
|
|
max_chunk = 500
|
|
|
else:
|
|
|
summarizer = english_summarizer
|
|
|
max_chunk = 1024
|
|
|
|
|
|
|
|
|
text = text.strip().replace("\n", " ")
|
|
|
chunks = [text[i:i+max_chunk] for i in range(0, len(text), max_chunk)]
|
|
|
|
|
|
summary = ""
|
|
|
for chunk in chunks:
|
|
|
result = summarizer(
|
|
|
chunk,
|
|
|
max_length=150,
|
|
|
min_length=40,
|
|
|
do_sample=False
|
|
|
)[0]['summary_text']
|
|
|
summary += result + "\n"
|
|
|
|
|
|
return summary.strip()
|
|
|
|
|
|
|
|
|
with gr.Blocks() as demo:
|
|
|
gr.Markdown("## 📄 SummarizeX – AI Text & PDF Summarizer 🌐")
|
|
|
gr.Markdown("خلاصه‌ساز هوش مصنوعی برای متن و PDF – پشتیبانی از فارسی و انگلیسی")
|
|
|
|
|
|
with gr.Row():
|
|
|
language = gr.Radio(["Persian", "English"], label="زبان خلاصه‌سازی", value="Persian")
|
|
|
|
|
|
with gr.Row():
|
|
|
pdf_input = gr.File(label="آپلود PDF", file_types=[".pdf"])
|
|
|
text_input = gr.Textbox(label="یا متن را وارد کنید", lines=8, placeholder="متن را اینجا وارد کنید...")
|
|
|
|
|
|
summarize_button = gr.Button("خلاصه کن / Summarize")
|
|
|
output_box = gr.Textbox(label="خلاصه نهایی", lines=8)
|
|
|
|
|
|
summarize_button.click(summarize_document, inputs=[language, text_input, pdf_input], outputs=output_box)
|
|
|
|
|
|
demo.launch(share=True) |
