_LABELLENZ / app.py
dev2607's picture
Upload folder using huggingface_hub
4fce4af verified
import os
import subprocess
import sys
import re
import numpy as np
from PIL import Image
import gradio as gr
import requests
import json
from dotenv import load_dotenv
# Attempt to install pytesseract if not found
try:
import pytesseract
except ImportError:
subprocess.check_call([sys.executable, '-m', 'pip', 'install', 'pytesseract'])
import pytesseract
# AFTER importing pytesseract, then set the path
try:
# First try the default path
if os.path.exists('/usr/bin/tesseract'):
pytesseract.pytesseract.tesseract_cmd = '/usr/bin/tesseract'
# Try to find it on the PATH
else:
tesseract_path = subprocess.check_output(['which', 'tesseract']).decode().strip()
if tesseract_path:
pytesseract.pytesseract.tesseract_cmd = tesseract_path
except:
# If all else fails, try the default installation path
pytesseract.pytesseract.tesseract_cmd = 'tesseract'
# Load environment variables
load_dotenv()
# Mistral API Key
MISTRAL_API_KEY = "GlrVCBWyvTYjWGKl5jqtK4K41uWWJ79F"
# Import and configure Mistral API
def analyze_ingredients_with_mistral(ingredients_list, health_conditions=None):
"""
Use Mistral AI to analyze ingredients and provide health insights.
"""
if not ingredients_list:
return "No ingredients detected or provided."
# Prepare the list of ingredients for the prompt
ingredients_text = ", ".join(ingredients_list)
# Create a prompt for Mistral
if health_conditions and health_conditions.strip():
prompt = f"""
Analyze the following food ingredients for a person with these health conditions: {health_conditions}
Ingredients: {ingredients_text}
For each ingredient:
1. Provide its potential health benefits
2. Identify any potential risks
3. Note if it may affect the specified health conditions
Then provide an overall assessment of the product's suitability for someone with the specified health conditions.
Format your response in markdown with clear headings and sections.
"""
else:
prompt = f"""
Analyze the following food ingredients:
Ingredients: {ingredients_text}
For each ingredient:
1. Provide its potential health benefits
2. Identify any potential risks or common allergens associated with it
Then provide an overall assessment of the product's general health profile.
Format your response in markdown with clear headings and sections.
"""
try:
headers = {
"Authorization": f"Bearer {MISTRAL_API_KEY}",
"Content-Type": "application/json"
}
data = {
"model": "mistral-small", # Or another suitable model
"messages": [{"role": "user", "content": prompt}],
"temperature": 0.7,
}
response = requests.post("https://api.mistral.ai/v1/chat/completions", headers=headers, json=data)
if response.status_code == 200:
analysis = response.json()['choices'][0]['message']['content']
else:
return dummy_analyze(ingredients_list, health_conditions) + f"\n\n(Using fallback analysis: Mistral API Error - {response.status_code} - {response.text})"
# Add disclaimer
disclaimer = """
## Disclaimer
This analysis is provided for informational purposes only and should not replace professional medical advice.
Always consult with a healthcare provider regarding dietary restrictions, allergies, or health conditions.
"""
return analysis + disclaimer
except Exception as e:
# Fallback to basic analysis if API call fails
return dummy_analyze(ingredients_list, health_conditions) + f"\n\n(Using fallback analysis: {str(e)})"
# Dummy analysis function for when API is not available
def dummy_analyze(ingredients_list, health_conditions=None):
ingredients_text = ", ".join(ingredients_list)
report = f"""
# Ingredient Analysis Report
## Detected Ingredients
{", ".join([i.title() for i in ingredients_list])}
## Overview
This is a simulated analysis since the Mistral API call failed. In the actual application,
the ingredients would be analyzed by Mistral for their health implications.
## Health Considerations
"""
if health_conditions:
report += f"""
The analysis would specifically consider these health concerns: {health_conditions}
"""
else:
report += """
No specific health concerns were provided, so a general analysis would be performed.
"""
report += """
## Disclaimer
This analysis is provided for informational purposes only and should not replace professional medical advice.
Always consult with a healthcare provider regarding dietary restrictions, allergies, or health conditions.
"""
return report
# Function to extract text from images using OCR
def extract_text_from_image(image):
try:
if image is None:
return "No image captured. Please try again."
# Verify Tesseract executable is accessible
try:
subprocess.run([pytesseract.pytesseract.tesseract_cmd, "--version"],
check=True, capture_output=True, text=True)
except (subprocess.SubprocessError, FileNotFoundError):
return "Tesseract OCR is not installed or not properly configured. Please check installation."
# Import necessary libraries
import cv2
import numpy as np
from PIL import Image, ImageOps, ImageEnhance
# First approach: Invert the image for light text on dark background
inverted_image = ImageOps.invert(image)
# Try OCR on inverted image
custom_config = r'--oem 3 --psm 6 -l eng --dpi 300'
inverted_text = pytesseract.image_to_string(inverted_image, config=custom_config)
# Second approach: OpenCV processing for colored backgrounds
img_cv = cv2.cvtColor(np.array(image), cv2.COLOR_RGB2BGR)
# Convert to grayscale
gray = cv2.cvtColor(img_cv, cv2.COLOR_BGR2GRAY)
# Apply bilateral filter to preserve edges while reducing noise
filtered = cv2.bilateralFilter(gray, 11, 17, 17)
# Adaptive thresholding to handle varied lighting
thresh = cv2.adaptiveThreshold(filtered, 255, cv2.ADAPTIVE_THRESH_GAUSSIAN_C,
cv2.THRESH_BINARY, 11, 2)
# Invert the image (if text is light on dark background)
inverted_thresh = cv2.bitwise_not(thresh)
# Try OCR on processed image
cv_text = pytesseract.image_to_string(
Image.fromarray(inverted_thresh),
config=custom_config
)
# Third approach: Color filtering to isolate text from colored background
# Convert to HSV color space to better isolate colors
hsv = cv2.cvtColor(img_cv, cv2.COLOR_BGR2HSV)
# Create a mask to extract light colored text (assuming white/light text)
lower_white = np.array([0, 0, 150])
upper_white = np.array([180, 30, 255])
mask = cv2.inRange(hsv, lower_white, upper_white)
# Apply morphological operations to clean up the mask
kernel = np.ones((2, 2), np.uint8)
mask = cv2.morphologyEx(mask, cv2.MORPH_CLOSE, kernel)
mask = cv2.morphologyEx(mask, cv2.MORPH_OPEN, kernel)
# Improve character connectivity
mask = cv2.dilate(mask, kernel, iterations=1)
# Try OCR on color filtered image
color_text = pytesseract.image_to_string(
Image.fromarray(mask),
config=r'--oem 3 --psm 6 -l eng --dpi 300'
)
# Fourth approach: Try directly with the image but with different configs
direct_text = pytesseract.image_to_string(
image,
config=r'--oem 3 --psm 11 -l eng --dpi 300'
)
# Compare results and select the best one
results = [inverted_text, cv_text, color_text, direct_text]
# Select the result with the most alphanumeric characters
def count_alphanumeric(text):
return sum(c.isalnum() for c in text)
best_text = max(results, key=count_alphanumeric)
# If still poor results, try with explicit text color inversion in tesseract
if count_alphanumeric(best_text) < 20:
# Try with tesseract's built-in inversion
neg_text = pytesseract.image_to_string(
image,
config=r'--oem 3 --psm 6 -c textord_heavy_nr=1 -c textord_debug_printable=0 -l eng --dpi 300'
)
if count_alphanumeric(neg_text) > count_alphanumeric(best_text):
best_text = neg_text
# Clean up the text
best_text = re.sub(r'[^\w\s,;:%.()\n\'-]', '', best_text)
best_text = best_text.replace('\n\n', '\n')
# Special case for ingredients list format
if "ingredient" in best_text.lower() or any(x in best_text.lower() for x in ["sugar", "cocoa", "milk", "contain"]):
# Specific cleaning for ingredient lists
best_text = re.sub(r'([a-z])([A-Z])', r'\1 \2', best_text) # Add space between lowercase and uppercase
best_text = re.sub(r'(\d+)([a-zA-Z])', r'\1 \2', best_text) # Add space between number and letter
if not best_text.strip():
return "No text could be extracted. Ensure image is clear and readable."
return best_text.strip()
except Exception as e:
return f"Error extracting text: {str(e)}"
# Function to parse ingredients from text
def parse_ingredients(text):
if not text:
return []
# Clean up the text
text = re.sub(r'^ingredients:?\s*', '', text.lower(), flags=re.IGNORECASE)
# Remove common OCR errors and extraneous characters
text = re.sub(r'[|\\/@#$%^&*()_+=]', '', text)
# Replace common OCR errors
text = re.sub(r'\bngredients\b', 'ingredients', text)
# Handle common OCR misreads
replacements = {
'0': 'o', 'l': 'i', '1': 'i',
'5': 's', '8': 'b', 'Q': 'g',
}
for error, correction in replacements.items():
text = text.replace(error, correction)
# Split by common ingredient separators
ingredients = re.split(r',|;|\n', text)
# Clean up each ingredient
cleaned_ingredients = []
for i in ingredients:
i = i.strip().lower()
if i and len(i) > 1: # Ignore single characters which are likely OCR errors
cleaned_ingredients.append(i)
return cleaned_ingredients
# Function to process input based on method (camera, upload, or manual entry)
def process_input(input_method, text_input, camera_input, upload_input, health_conditions):
if input_method == "Camera":
if camera_input is not None:
extracted_text = extract_text_from_image(camera_input)
# If OCR fails, inform the user they can try manual entry
if "Error" in extracted_text or "No text could be extracted" in extracted_text:
return extracted_text + "\n\nPlease try using the 'Manual Entry' option instead."
ingredients = parse_ingredients(extracted_text)
return analyze_ingredients_with_mistral(ingredients, health_conditions)
else:
return "No camera image captured. Please try again."
elif input_method == "Image Upload":
if upload_input is not None:
extracted_text = extract_text_from_image(upload_input)
# If OCR fails, inform the user they can try manual entry
if "Error" in extracted_text or "No text could be extracted" in extracted_text:
return extracted_text + "\n\nPlease try using the 'Manual Entry' option instead."
ingredients = parse_ingredients(extracted_text)
return analyze_ingredients_with_mistral(ingredients, health_conditions)
else:
return "No image uploaded. Please try again."
elif input_method == "Manual Entry":
if text_input and text_input.strip():
ingredients = parse_ingredients(text_input)
return analyze_ingredients_with_mistral(ingredients, health_conditions)
else:
return "No ingredients entered. Please try again."
return "Please provide input using one of the available methods."
# Create the Gradio interface
with gr.Blocks(title="AI Ingredient Scanner") as app:
gr.Markdown("# AI Ingredient Scanner")
gr.Markdown("Scan product ingredients and analyze them for health benefits, risks, and potential allergens.")
with gr.Row():
with gr.Column():
input_method = gr.Radio(
["Camera", "Image Upload", "Manual Entry"],
label="Input Method",
value="Camera"
)
# Camera input
camera_input = gr.Image(label="Capture ingredients with camera", type="pil", visible=True)
# Image upload
upload_input = gr.Image(label="Upload image of ingredients label", type="pil", visible=False)
# Text input
text_input = gr.Textbox(
label="Enter ingredients list (comma separated)",
placeholder="milk, sugar, flour, eggs, vanilla extract",
lines=3,
visible=False
)
# Health conditions input - now optional and more flexible
health_conditions = gr.Textbox(
label="Enter your health concerns (optional)",
placeholder="diabetes, high blood pressure, peanut allergy, etc.",
lines=2,
info="The AI will automatically analyze ingredients for these conditions"
)
analyze_button = gr.Button("Analyze Ingredients")
with gr.Column():
output = gr.Markdown(label="Analysis Results")
extracted_text_output = gr.Textbox(label="Extracted Text (for verification)", lines=3)
# Show/hide inputs based on selection
def update_visible_inputs(choice):
return {
upload_input: gr.update(visible=(choice == "Image Upload")),
camera_input: gr.update(visible=(choice == "Camera")),
text_input: gr.update(visible=(choice == "Manual Entry"))
}
input_method.change(update_visible_inputs, input_method, [upload_input, camera_input, text_input])
# Extract and display the raw text (for verification purposes)
def show_extracted_text(input_method, text_input, camera_input, upload_input):
if input_method == "Camera" and camera_input is not None:
return extract_text_from_image(camera_input)
elif input_method == "Image Upload" and upload_input is not None:
return extract_text_from_image(upload_input)
elif input_method == "Manual Entry":
return text_input
return "No input detected"
# Set up event handlers
analyze_button.click(
fn=process_input,
inputs=[input_method, text_input, camera_input, upload_input, health_conditions],
outputs=output
)
analyze_button.click(
fn=show_extracted_text,
inputs=[input_method, text_input, camera_input, upload_input],
outputs=extracted_text_output
)
gr.Markdown("### How to use")
gr.Markdown("""
1. Choose your input method (Camera, Image Upload, or Manual Entry)
2. Take a photo of the ingredients label or enter ingredients manually
3. Optionally enter your health concerns
4. Click "Analyze Ingredients" to get your personalized analysis
The AI will automatically analyze the ingredients, their health implications, and their potential impact on your specific health concerns.
""")
gr.Markdown("### Examples of what you can ask")
gr.Markdown("""
The system can handle a wide range of health concerns, such as:
- General health goals: "trying to reduce sugar intake" or "watching sodium levels"
- Medical conditions: "diabetes" or "hypertension"
- Allergies: "peanut allergy" or "shellfish allergy"
- Dietary restrictions: "vegetarian" or "gluten-free diet"
- Multiple conditions: "diabetes, high cholesterol, and lactose intolerance"
The AI will tailor its analysis to your specific needs.
""")
gr.Markdown("### Tips for best results")
gr.Markdown("""
- Hold the camera steady and ensure good lighting
- Focus directly on the ingredients list
- Make sure the text is clear and readable
- Be specific about your health concerns for more targeted analysis
""")
gr.Markdown("### Disclaimer")
gr.Markdown("""
This tool is for informational purposes only and should not replace professional medical advice.
Always consult with a healthcare provider regarding dietary restrictions, allergies, or health conditions.
""")
# Launch the app
if __name__ == "__main__":
app.launch()