Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
|
@@ -0,0 +1,65 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import gradio as gr
|
| 2 |
+
from groq import Groq
|
| 3 |
+
from PyPDF2 import PdfReader
|
| 4 |
+
from docx import Document
|
| 5 |
+
import os
|
| 6 |
+
from gtts import gTTS
|
| 7 |
+
import tempfile
|
| 8 |
+
import whisper
|
| 9 |
+
|
| 10 |
+
# Set the API key for Groq
|
| 11 |
+
api_key = "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"
|
| 12 |
+
os.environ["GROQ_API_KEY"] = api_key
|
| 13 |
+
|
| 14 |
+
# Initialize Groq client with API key from environment variable
|
| 15 |
+
api_key = os.getenv("GROQ_API_KEY")
|
| 16 |
+
if not api_key:
|
| 17 |
+
raise ValueError("API key for Groq is not set. Please set the environment variable GROQ_API_KEY.")
|
| 18 |
+
|
| 19 |
+
client = Groq(api_key=api_key)
|
| 20 |
+
|
| 21 |
+
# Load Whisper model
|
| 22 |
+
model = whisper.load_model("base")
|
| 23 |
+
|
| 24 |
+
def summarize_document(file):
|
| 25 |
+
# Read file content
|
| 26 |
+
if file.name.endswith('.pdf'):
|
| 27 |
+
# Read PDF file
|
| 28 |
+
reader = PdfReader(file.name)
|
| 29 |
+
text = ''.join([page.extract_text() for page in reader.pages])
|
| 30 |
+
elif file.name.endswith('.docx'):
|
| 31 |
+
# Read DOCX file
|
| 32 |
+
doc = Document(file.name)
|
| 33 |
+
text = ''.join([para.text for para in doc.paragraphs])
|
| 34 |
+
else:
|
| 35 |
+
return "Unsupported file format. Please upload a PDF or DOCX file."
|
| 36 |
+
|
| 37 |
+
# Generate summary
|
| 38 |
+
chat_completion = client.chat.completions.create(
|
| 39 |
+
messages=[
|
| 40 |
+
{"role": "user", "content": f"Please summarize the following text: {text}"}
|
| 41 |
+
],
|
| 42 |
+
model="llama3-8b-8192",
|
| 43 |
+
)
|
| 44 |
+
|
| 45 |
+
summary = chat_completion.choices[0].message.content
|
| 46 |
+
|
| 47 |
+
# Convert summary text to speech using gTTS
|
| 48 |
+
tts = gTTS(text=summary, lang='en')
|
| 49 |
+
audio_file = tempfile.NamedTemporaryFile(delete=False, suffix='.mp3')
|
| 50 |
+
tts.save(audio_file.name)
|
| 51 |
+
|
| 52 |
+
return summary, audio_file.name
|
| 53 |
+
|
| 54 |
+
# Create Gradio interface with Monochrome theme
|
| 55 |
+
iface = gr.Interface(
|
| 56 |
+
fn=summarize_document,
|
| 57 |
+
inputs=gr.File(label="Upload a Word or PDF Document"),
|
| 58 |
+
outputs=[gr.Textbox(label="Summary"), gr.Audio(label="Audio Summary")],
|
| 59 |
+
title="Document Summarizer",
|
| 60 |
+
description="Upload a Word or PDF document and get a summary with audio playback.",
|
| 61 |
+
theme=gr.themes.Monochrome() # Apply Monochrome theme
|
| 62 |
+
)
|
| 63 |
+
|
| 64 |
+
# Launch the interface with sharing enabled
|
| 65 |
+
iface.launch(share=True, debug=True)
|