Spaces:
Sleeping
Sleeping
Create app.py
Browse files
app.py
ADDED
|
@@ -0,0 +1,73 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import gradio as gr
|
| 2 |
+
import google.generativeai as genai
|
| 3 |
+
from PIL import Image
|
| 4 |
+
import os
|
| 5 |
+
|
| 6 |
+
# Direct API key placement (NOT RECOMMENDED)
|
| 7 |
+
API_KEY = 'AIzaSyBMZrhMXRpQKp7M-JcN2Qk73afeta5Mv5Y'
|
| 8 |
+
|
| 9 |
+
def extract_malayalam_text(image):
|
| 10 |
+
"""
|
| 11 |
+
Extract handwritten Malayalam text from an image using Gemini 2.0 Flash
|
| 12 |
+
|
| 13 |
+
Args:
|
| 14 |
+
image (PIL.Image): Uploaded image
|
| 15 |
+
|
| 16 |
+
Returns:
|
| 17 |
+
str: Extracted text from the image
|
| 18 |
+
"""
|
| 19 |
+
# Validate input
|
| 20 |
+
if image is None:
|
| 21 |
+
return "Please upload an image first."
|
| 22 |
+
|
| 23 |
+
try:
|
| 24 |
+
# Configure the Gemini API
|
| 25 |
+
genai.configure(api_key=API_KEY)
|
| 26 |
+
|
| 27 |
+
# Use Gemini 2.0 Flash model
|
| 28 |
+
model = genai.GenerativeModel('gemini-2.0-flash')
|
| 29 |
+
|
| 30 |
+
# Highly specific prompt for Malayalam text extraction
|
| 31 |
+
response = model.generate_content(
|
| 32 |
+
[
|
| 33 |
+
"CRITICAL INSTRUCTIONS: "
|
| 34 |
+
"This is a handwritten Malayalam grocery list. "
|
| 35 |
+
"You MUST extract the text ONLY in Malayalam script. "
|
| 36 |
+
"Rules for extraction: "
|
| 37 |
+
"1. Use pure Malayalam Unicode characters "
|
| 38 |
+
"2. Maintain original handwriting style "
|
| 39 |
+
"3. If any text is unclear, mark with [uncertain] "
|
| 40 |
+
"4. Do not translate or modify original text "
|
| 41 |
+
"5. Preserve exact Malayalam script as written",
|
| 42 |
+
image
|
| 43 |
+
],
|
| 44 |
+
generation_config=genai.types.GenerationConfig(
|
| 45 |
+
temperature=0.1, # Ultra-low temperature for precise extraction
|
| 46 |
+
max_output_tokens=300 # Adjust based on expected list length
|
| 47 |
+
)
|
| 48 |
+
)
|
| 49 |
+
|
| 50 |
+
# Return the extracted text
|
| 51 |
+
extracted_text = response.text
|
| 52 |
+
|
| 53 |
+
# Additional validation
|
| 54 |
+
if not any('\u0D00' <= char <= '\u0D7F' for char in extracted_text):
|
| 55 |
+
return "No Malayalam characters detected. Please check the image quality."
|
| 56 |
+
|
| 57 |
+
return extracted_text
|
| 58 |
+
|
| 59 |
+
except Exception as e:
|
| 60 |
+
return f"An error occurred: {str(e)}"
|
| 61 |
+
|
| 62 |
+
# Create Gradio Interface
|
| 63 |
+
demo = gr.Interface(
|
| 64 |
+
fn=extract_malayalam_text,
|
| 65 |
+
inputs=gr.Image(type="pil", label="Upload Malayalam Grocery List Image"),
|
| 66 |
+
outputs=gr.Textbox(label="Extracted Malayalam Text", lines=10),
|
| 67 |
+
title="Malayalam Handwritten Text Extractor",
|
| 68 |
+
description="Upload a handwritten Malayalam grocery list image for text extraction."
|
| 69 |
+
)
|
| 70 |
+
|
| 71 |
+
# Launch the app
|
| 72 |
+
if __name__ == "__main__":
|
| 73 |
+
demo.launch()
|