Update app.py
Browse files
app.py
CHANGED
|
@@ -1,16 +1,25 @@
|
|
| 1 |
import gradio as gr
|
| 2 |
from transformers import pipeline
|
| 3 |
import PyPDF2
|
|
|
|
|
|
|
|
|
|
| 4 |
|
| 5 |
-
#
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 6 |
summarizer = pipeline(
|
| 7 |
"summarization",
|
| 8 |
model="m3hrdadfi/mt5-small-parsinlu-summarization-fa"
|
| 9 |
)
|
| 10 |
|
|
|
|
| 11 |
def summarize_text(text):
|
| 12 |
if not text.strip():
|
| 13 |
-
return "
|
| 14 |
result = summarizer(text, max_length=150, min_length=30)
|
| 15 |
return result[0]["summary_text"]
|
| 16 |
|
|
@@ -19,25 +28,74 @@ def summarize_pdf(file):
|
|
| 19 |
reader = PyPDF2.PdfReader(file.name)
|
| 20 |
text = ""
|
| 21 |
for page in reader.pages:
|
| 22 |
-
|
|
|
|
|
|
|
| 23 |
return summarize_text(text)
|
| 24 |
except Exception as e:
|
| 25 |
-
return f"خطا در خواندن PDF: {e}"
|
| 26 |
-
|
| 27 |
-
|
| 28 |
-
|
| 29 |
-
|
| 30 |
-
|
| 31 |
-
|
| 32 |
-
|
| 33 |
-
|
| 34 |
-
|
| 35 |
-
|
| 36 |
-
|
| 37 |
-
|
| 38 |
-
|
| 39 |
-
|
| 40 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 41 |
|
| 42 |
if __name__ == "__main__":
|
| 43 |
demo.launch(server_name="0.0.0.0", server_port=7860)
|
|
|
|
| 1 |
import gradio as gr
|
| 2 |
from transformers import pipeline
|
| 3 |
import PyPDF2
|
| 4 |
+
import pandas as pd
|
| 5 |
+
from fpdf import FPDF
|
| 6 |
+
from datetime import datetime
|
| 7 |
|
| 8 |
+
# رنگهای برند
|
| 9 |
+
PRIMARY_GRADIENT = "linear-gradient(90deg, #6a4cff, #00c9b7)"
|
| 10 |
+
HEADER_COLOR = "#1f2937"
|
| 11 |
+
TEXT_COLOR = "#111827"
|
| 12 |
+
|
| 13 |
+
# مدل خلاصهسازی
|
| 14 |
summarizer = pipeline(
|
| 15 |
"summarization",
|
| 16 |
model="m3hrdadfi/mt5-small-parsinlu-summarization-fa"
|
| 17 |
)
|
| 18 |
|
| 19 |
+
# --- توابع ---
|
| 20 |
def summarize_text(text):
|
| 21 |
if not text.strip():
|
| 22 |
+
return "⚠️ لطفاً متن وارد کنید."
|
| 23 |
result = summarizer(text, max_length=150, min_length=30)
|
| 24 |
return result[0]["summary_text"]
|
| 25 |
|
|
|
|
| 28 |
reader = PyPDF2.PdfReader(file.name)
|
| 29 |
text = ""
|
| 30 |
for page in reader.pages:
|
| 31 |
+
txt = page.extract_text()
|
| 32 |
+
if txt:
|
| 33 |
+
text += txt + "\n"
|
| 34 |
return summarize_text(text)
|
| 35 |
except Exception as e:
|
| 36 |
+
return f"❌ خطا در خواندن PDF: {e}"
|
| 37 |
+
|
| 38 |
+
def save_to_pdf(text, summary):
|
| 39 |
+
filename = f"summary_{datetime.now().strftime('%Y%m%d_%H%M%S')}.pdf"
|
| 40 |
+
pdf = FPDF()
|
| 41 |
+
pdf.add_page()
|
| 42 |
+
pdf.set_font("Arial", size=12)
|
| 43 |
+
pdf.multi_cell(0, 10, f"📄 متن اصلی:\n\n{text}\n\n---\n\n📝 خلاصه:\n\n{summary}")
|
| 44 |
+
pdf.output(filename)
|
| 45 |
+
return filename
|
| 46 |
+
|
| 47 |
+
def save_to_excel(text, summary):
|
| 48 |
+
filename = f"summary_{datetime.now().strftime('%Y%m%d_%H%M%S')}.xlsx"
|
| 49 |
+
df = pd.DataFrame({"متن اصلی": [text], "خلاصه": [summary]})
|
| 50 |
+
df.to_excel(filename, index=False)
|
| 51 |
+
return filename
|
| 52 |
+
|
| 53 |
+
# --- رابط کاربری ---
|
| 54 |
+
with gr.Blocks(css="""
|
| 55 |
+
body { font-family: Vazir, sans-serif; background: #f9fafb; }
|
| 56 |
+
h1 { font-weight: bold; color: white; text-align: center; padding: 20px;
|
| 57 |
+
background: linear-gradient(90deg, #6a4cff, #00c9b7); border-radius: 8px; }
|
| 58 |
+
.tab { background-color: white; padding: 20px; border-radius: 8px;
|
| 59 |
+
box-shadow: 0px 2px 8px rgba(0,0,0,0.05); }
|
| 60 |
+
button { border-radius: 8px !important; font-weight: bold; }
|
| 61 |
+
""") as demo:
|
| 62 |
+
|
| 63 |
+
with gr.Row():
|
| 64 |
+
gr.Markdown("<h1>📝 SummarizeX — خلاصهساز متن و PDF</h1>")
|
| 65 |
+
|
| 66 |
+
with gr.Tab("خلاصه متن") as tab1:
|
| 67 |
+
with gr.Row(elem_classes="tab"):
|
| 68 |
+
text_input = gr.Textbox(lines=10, placeholder="متن خود را اینجا وارد کنید...")
|
| 69 |
+
summary_output = gr.Textbox(lines=8, label="خلاصه متن")
|
| 70 |
+
with gr.Row():
|
| 71 |
+
btn_summary = gr.Button("✨ خلاصه کن", elem_id="btn-summary")
|
| 72 |
+
with gr.Row():
|
| 73 |
+
pdf_btn = gr.Button("📄 ذخیره به PDF")
|
| 74 |
+
excel_btn = gr.Button("📊 ذخیره به Excel")
|
| 75 |
+
with gr.Row():
|
| 76 |
+
file_pdf_out = gr.File(label="دانلود PDF")
|
| 77 |
+
file_excel_out = gr.File(label="دانلود Excel")
|
| 78 |
+
|
| 79 |
+
btn_summary.click(summarize_text, inputs=text_input, outputs=summary_output)
|
| 80 |
+
pdf_btn.click(lambda t, s: save_to_pdf(t, s), inputs=[text_input, summary_output], outputs=file_pdf_out)
|
| 81 |
+
excel_btn.click(lambda t, s: save_to_excel(t, s), inputs=[text_input, summary_output], outputs=file_excel_out)
|
| 82 |
+
|
| 83 |
+
with gr.Tab("خلاصه PDF") as tab2:
|
| 84 |
+
with gr.Row(elem_classes="tab"):
|
| 85 |
+
pdf_input = gr.File(type="file", file_types=[".pdf"], label="انتخاب فایل PDF")
|
| 86 |
+
pdf_summary_output = gr.Textbox(lines=8, label="خلاصه PDF")
|
| 87 |
+
with gr.Row():
|
| 88 |
+
btn_pdf_summary = gr.Button("✨ خلاصه PDF")
|
| 89 |
+
with gr.Row():
|
| 90 |
+
pdf_btn2 = gr.Button("📄 ذخیره به PDF")
|
| 91 |
+
excel_btn2 = gr.Button("📊 ذخیره به Excel")
|
| 92 |
+
with gr.Row():
|
| 93 |
+
file_pdf_out2 = gr.File(label="دانلود PDF")
|
| 94 |
+
file_excel_out2 = gr.File(label="دانلود Excel")
|
| 95 |
+
|
| 96 |
+
btn_pdf_summary.click(summarize_pdf, inputs=pdf_input, outputs=pdf_summary_output)
|
| 97 |
+
pdf_btn2.click(lambda f, s: save_to_pdf("PDF File", s), inputs=[pdf_input, pdf_summary_output], outputs=file_pdf_out2)
|
| 98 |
+
excel_btn2.click(lambda f, s: save_to_excel("PDF File", s), inputs=[pdf_input, pdf_summary_output], outputs=file_excel_out2)
|
| 99 |
|
| 100 |
if __name__ == "__main__":
|
| 101 |
demo.launch(server_name="0.0.0.0", server_port=7860)
|