newberryai 0.2.2

NewberryAI Python Package

🫐📦 NewberryAI

NewberryAI

The complete AI toolkit that turns complex workflows into simple Python commands. From medical diagnosis to compliance checking, document analysis to face recognition - NewberryAI brings enterprise-grade AI capabilities to your fingertips.

---

Installation

pip install newberryai

Troubleshooting Installation Issues

If you encounter issues installing pyaudio or related audio dependencies, try the following:

For Ubuntu/Debian systems:

bashsudo apt-get install -y portaudio19-dev
pip install pyaudio

For other systems:

macOS:

brew install portaudio then pip install pyaudio

Windows:

Download pre-compiled wheels from "https://www.lfd.uci.edu/~gohlke/pythonlibs/#pyaudio" and

pip install PyAudio‑0.2.11‑cp312‑cp312‑win_amd64.whl

replace the filename with the one you downloaded

📚Features

🏥 Healthcare & Medical

• HealthScribe: Medical transcription with AWS HealthScribe
• Differential Diagnosis Assistant(DDX): Clinical diagnosis support
• Medical Bill Extractor: Automated medical billing analysis

🔒 Compliance & Security

• Compliance Checker: Video analysis for regulatory compliance
• PII Redactor: Remove personally identifiable information
• PII Extractor: Extract and categorize sensitive data

📊 Data & Analytics

• EDA Assistant: Automated exploratory data analysis
• NL2SQL: Natural language to SQL query conversion
• Excel Formula Generator: AI-powered Excel formula creation

📄 Document Processing

• PDF Summarizer: Intelligent document summarization
• PDF Extractor: Semantic search and content extraction
• Handwriting to Text Converter: Extract handwritten text from images using AI

🎨 Media Generation

• Video Generator: Text-to-video with Amazon Bedrock Nova
• Image Generator: Text-to-image with Titan Image Generator
• Virtual Try-On: AI-powered fashion visualization

🔍 Computer Vision

• Face Recognition: Identity management with AWS Rekognition
• Face Detection: Video face detection and tracking

💻 Development Tools

• Coding Assistant: Code review and debugging support
• Speech-to-Speech: Real-time voice interaction

1. Compliance Checker

Environment Setup

# AWS Credentials
AWS_ACCESS_KEY_ID=your_aws_access_key
AWS_SECRET_ACCESS_KEY=your_aws_secret_key
AWS_DEFAULT_REGION=us-east-1

Python SDK

#Analyze videos for regulatory compliance. Requires AWS credentials.
from newberryai import ComplianceChecker

checker = ComplianceChecker()
video_file = 'YOUR.mp4'
compliance_question = 'Is the video compliant with safety regulations such as mask?'

result, status_code = checker.check_compliance(
    video_file=video_file,
    prompt=compliance_question
)
if status_code:
    print(f'Error: {result.get("error", "Unknown error")})')
else:
    print(f'Compliant: {"Yes" if result["compliant"] else "No"}')
    print(f'Analysis: {result["analysis"]}')

CLI Usage

newberryai compliance --video_file YOUR.mp4 --question "Is the video compliant with safety regulations such as mask?"
newberryai compliance --gradio

---

2. HealthScribe

Environment Setup

# AWS Credentials
AWS_ACCESS_KEY_ID=your_aws_access_key
AWS_SECRET_ACCESS_KEY=your_aws_secret_key
AWS_DEFAULT_REGION=us-east-1
HEALTHSCRIBE_INPUT_BUCKET=your-healthscribe-input-bucket
HEALTHSCRIBE_OUTPUT_BUCKET=your-healthscribe-output-bucket
HEALTHSCRIBE_DATA_ACCESS_ROLE=arn:aws:iam::account:role/your-role

Python SDK

# Medical transcription using AWS HealthScribe. Requires AWS credentials.
from newberryai import HealthScribe
scribe = HealthScribe(
    input_s3_bucket='your-input-bucket',
    data_access_role_arn='arn:aws:iam::992382417943:role/YOUR-role'
)
result = scribe.process(
    file_path=r'YOUR_AUDIO.mp3',
    job_name='JOB-NAME',
    output_s3_bucket='your-output-bucket'
)
print(result["summary"])

CLI Usage

newberryai healthscribe --file_path YOUR_AUDIO.mp3 --job_name sdktest --output_s3_bucket dax-healthscribe-v2
newberryai healthscribe --gradio

---

3. Differential Diagnosis (DDx) Assistant

Environment Setup

# OpenAI Configuration
OPENAI_API_KEY=your_openai_api_key

Python SDK

#Get assistance with clinical diagnosis.
from newberryai import DDxChat

ddx_chat = DDxChat()
response = ddx_chat.ask('Patient presents with fever, cough, and fatigue for 5 days')
print(response)

CLI Usage

newberryai ddx --question "Patient presents with fever, cough, and fatigue for 5 days"
newberryai ddx --interactive
newberryai ddx --gradio

---

4. Excel Formula Generator AI Assistant

Environment Setup

# OpenAI Configuration
OPENAI_API_KEY=your_openai_api_key

Python SDK

#Get assistance with Excel formulas.
from newberryai import ExcelExp

excel_expert = ExcelExp()
response = excel_expert.ask(
"Calculate average sales for products that meet specific criteria E.g: give me excel formula to calculate average of my sale for year 2010,2011 sales is in col A, Year in Col B  and Months in Col C"
)
print(response)

CLI Usage

newberryai excel --question "Calculate average sales for products that meet specific criteria E.g: give me excel formula to calculate average of my sale for year 2010,2011 sales is in col A, Year in Col B  and Months in Col C"
newberryai excel --interactive
newberryai excel --gradio

---

5. Medical Bill Extractor

Environment Setup

# OpenAI Configuration
OPENAI_API_KEY=your_openai_api_key
# AWS Credentials
AWS_ACCESS_KEY_ID=your_aws_access_key
AWS_SECRET_ACCESS_KEY=your_aws_secret_key

Python SDK

#Extract and analyze data from medical bills.
from newberryai import Bill_extractor

extractor = Bill_extractor()
analysis = extractor.analyze_document('Billimg.jpg')
print(analysis)

CLI Usage

newberryai bill --file_path Billimg.jpg
newberryai bill --interactive
newberryai bill --gradio

---

6. Coding and Debugging AI Assistant

Environment Setup

# OpenAI Configuration
OPENAI_API_KEY=your_openai_api_key

Python SDK

#Analyze code and help you with coding as debugger.
from newberryai import CodeReviewAssistant

code_debugger = CodeReviewAssistant()
response = code_debugger.ask('''Explain and correct below code
def calculate_average(nums):
sum = 0
for num in nums:
sum += num
average = sum / len(nums)
return average

numbers = [10, 20, 30, 40, 50]
result = calculate_average(numbers)
print("The average is:", results)''')
print(response)

CLI Usage

newberryai code --question "Explain and correct below code
def calculate_average(nums):
sum = 0
for num in nums:
sum += num
average = sum / len(nums)
return average

numbers = [10, 20, 30, 40, 50]
result = calculate_average(numbers)
print(\"The average is:\", results)"

newberryai code --interactive
newberryai code --gradio

---

7. PII Redactor AI Assistant

Environment Setup

# OpenAI Configuration
OPENAI_API_KEY=your_openai_api_key

Python SDK

#Analyze text and remove PII (personally identifiable information) from the text.
from newberryai import PII_Redaction

pii_red = PII_Redaction()
response = pii_red.ask("Patient name is John Doe with fever. he is from Austin,Texas.His email id is [e-mail]")
print(response)

CLI Usage

newberryai pii --text "Patient name is John Doe with fever. he is from Austin,Texas.His email id is [e-mail]"
newberryai pii --interactive
newberryai pii --gradio

---

8. PII Extractor AI Assistant

Environment Setup

# OpenAI Configuration
OPENAI_API_KEY=your_openai_api_key

Python SDK

#Analyze text and extract PII (personally identifiable information) from the text.
from newberryai import PII_extraction

pii_extract = PII_extraction()
response = pii_extract.ask("Patient name is John Doe with fever. he is from Austin,Texas.His email id is [e-mail]")
print(response)

CLI Usage

newberryai pii --text "Patient name is John Doe with fever. he is from Austin,Texas.His email id is [e-mail]"
newberryai pii --interactive
newberryai pii --gradio

---

9. EDA AI assistant

Environment Setup

# OpenAI Configuration
OPENAI_API_KEY=your_openai_api_key

Python SDK

#Perform detailed data exploration with real statistics, hypothesis testing, and actionable insights—no code, just direct analysis

from newberryai import EDA
import pandas as pd

eda = EDA()
eda.current_data = pd.read_csv(r'your_csv.csv')

response = eda.ask("What is the average value of column 'xyz'?")
print(response)

CLI Usage

newberryai eda --file_path your_csv.csv --question "What is the average value of column 'target'?"
newberryai eda --interactive
newberryai eda --gradio

---

10. PDF Document Summarizer

Environment Setup

# OpenAI Configuration
OPENAI_API_KEY=your_openai_api_key

Python SDK

#Extract and summarize content from PDF documents.
from newberryai import DocSummarizer

summarizer = DocSummarizer()
response = summarizer.ask(r'YOUR-pdf.pdf')
print(response)

CLI Usage

newberryai doc --file_path YOUR-pdf.pdf --question "Extract and summarize content from PDF documents."
newberryai doc --interactive
newberryai doc --gradio

---

11. PDF Extractor

Environment Setup

# OpenAI Configuration
OPENAI_API_KEY=your_openai_api_key

Python SDK

#Extract and query content from PDF documents using embeddings and semantic search. (Async usage)

import asyncio
from newberryai import PDFExtractor

async def pdf_extract_demo():
    extractor = PDFExtractor()
    pdf_id = await extractor.process_pdf(r'YOUR-pdf.pdf')
    response = await extractor.ask_question(pdf_id, 'What is the mode of review in the document?')
    print(response['answer'])
    print("\nSource Chunks:")
    for chunk in response['source_chunks']:
        print(f"\n---\n{chunk}")

#To run the async demo in a notebook cell:
await pdf_extract_demo()

CLI Usage

newberryai pdf --file_path YOUR-pdf.pdf --question "What is the mode of review in the document?"
newberryai pdf --interactive
newberryai pdf --gradio

---

12. Video Generator

Environment Setup

# AWS Credentials
AWS_ACCESS_KEY_ID=your_aws_access_key
AWS_SECRET_ACCESS_KEY=your_aws_secret_key
BEDROCK_REGION=us-east-1
BEDROCK_MODEL_ID=amazon.nova-canvas-v1:0

Python SDK

#Generate videos from text using Amazon Bedrock's Nova model. Requires AWS credentials.
# Example usage
from newberryai import VideoGenerator
generator = VideoGenerator()
prompt = "A cat dancing on a wall"
async def run_video():
    response = await generator.generate(
        text=prompt,
        duration_seconds=6,
        fps=24,
        dimension="1280x720",
        seed=42
    )
    print(response["message"])
    print("Waiting for video to complete...")
    final_response = await generator.wait_for_completion(response["job_id"])
    print(final_response["message"])
    print(f"Video URL: {final_response['video_url']}")
await run_video()

CLI Usage

newberryai video --text "A cat dancing on a wall" --duration_seconds 6 --fps 24 --dimension 1280x720 --seed 42
newberryai video --interactive
newberryai video --gradio

---

13. Image Generator

Environment Setup

# AWS Credentials
AWS_ACCESS_KEY_ID=your_aws_access_key
AWS_SECRET_ACCESS_KEY=your_aws_secret_key
BEDROCK_REGION=us-east-1
BEDROCK_MODEL_ID=amazon.titan-image-generator-v1

Python SDK

from newberryai.image_generator import ImageGenerator
import asyncio

generator = ImageGenerator()
prompt = "A lotus in a pond"
result = await generator.generate(
    text= prompt,
    width=512,
    height=512,
    number_of_images=1,
    cfg_scale=8,
    seed=42,
    quality="standard"
)

print(result["message"])
for path in result["images"]:
    print(f"Generated image path: {path}")

CLI Usage

newberryai image --text "A lotus in a pond" --width 512 --height 512 --number_of_images 1 --cfg_scale 8 --seed 42 --quality standard
newberryai image --interactive
newberryai image --gradio

---

14. Face Recognition

Environment Setup

# AWS Credentials
AWS_ACCESS_KEY_ID=your_aws_access_key
AWS_SECRET_ACCESS_KEY=your_aws_secret_key
REKOGNITION_COLLECTION_ID=your-collection-id
REKOGNITION_REGION=us-east-1

Python SDK

# Import the FaceRecognition class
from newberryai import FaceRecognition

# Initialize face recognition
face_recognition = FaceRecognition()

# Add a face to the collection and recognize a face in one go
add_response = face_recognition.add_to_collect("yourimg.jpeg", "Name")
print(add_response["message"], f"Face ID: {add_response.get('face_id', 'N/A')}")

# Recognize a face from another image
recognize_response = face_recognition.recognize_image("yourimg2.jpeg")
print(recognize_response["message"])
if recognize_response["success"]:
    print(f"Recognized: {recognize_response['name']} (Confidence: {recognize_response['confidence']:.2f}%)")

CLI Usage

newberryai face --image_path yourimg.jpeg --name Name
newberryai face --image_path yourimg2.jpeg --recognize
newberryai face --interactive
newberryai face --gradio

---

15. Face Detection

Environment Setup

# AWS Credentials
AWS_ACCESS_KEY_ID=your_aws_access_key
AWS_SECRET_ACCESS_KEY=your_aws_secret_key
REKOGNITION_COLLECTION_ID=your-collection-id
REKOGNITION_REGION=us-east-1

Python SDK

# Import and initialize FaceDetection class
from newberryai.face_detection import FaceDetection
face_detector = FaceDetection()

# Add face to collection
add_response = face_detector.add_face_to_collection("yourimg.jpeg", "kirti")
print(add_response["message"])
if add_response["success"]:
    print(f"Face ID: {add_response['face_id']}")

# Process video
results = face_detector.process_video("yourvideo.mp4", max_frames=20)
for detection in results:
    print(f"\nTimestamp: {detection['timestamp']}s")
    if detection.get('external_image_id'):
        print(f"Matched Face: {detection['external_image_id']}")
        print(f"Face ID: {detection['face_id']}")
        print(f"Confidence: {detection['confidence']:.2f}%")
    else:
        print("No match found in collection")

CLI Usage

newberryai face --image_path yourimg.jpeg --add_to_collection
newberryai face --video_path yourvideo.mp4 --max_frames 20
newberryai face --interactive
newberryai face --gradio

---

16. NL2SQL

Environment Setup

# OpenAI Configuration
OPENAI_API_KEY=your_openai_api_key
# Database Configuration
DB_HOST=localhost
DB_USER=your_db_username
DB_PASSWORD=your_db_password
DB_NAME=your_database_name
DB_PORT=3306

Python SDK

#Natural Language to SQL Query Assistant
# Import the NL2SQL class
from newberryai import NL2SQL
import json
# Initialize NL2SQL processor
nl2sql = NL2SQL()

# Set up the database connection parameters (adjust these accordingly)
host = "127.0.0.1"
user = "user-name"
password = "passward"
database = "your-database-name"
port = 3306

# Connect to the database
nl2sql.connect_to_database(host, user, password, database, port)

# Test a natural language question to SQL conversion
question = "Show all tables"
response = nl2sql.process_query(question)

# Print the results: SQL query, data, and summary
if response["success"]:
    print(f"Generated SQL Query: {response['sql_query']}")
    print(f"Data: {json.dumps(response['data'], indent=2)}")
    print(f"Summary: {response['summary']}")
else:
    print(f"Error: {response['message']}")

CLI Usage

newberryai sql --question "Show all tables"
newberryai sql --interactive
newberryai sql --gradio

---

17. Virtual Try-On

Environment Setup

# Fashn API Configuration
FASHN_API_KEY=your_fashn_api_key
FASHN_API_URL=https://api.fashn.ai/v1

Python SDK

# Generate virtual try-on images using AI. Requires Fashn API credentials.
import base64
import asyncio
from newberryai import VirtualTryOn
try_on = VirtualTryOn()
request = await try_on.process(
    model_image='model.jpg',
    garment_image='image.png',
    category='tops'
)
async def tryon_demo():
    job_id = request["job_id"]
    while True:
        status = await try_on.get_status(job_id)
        if status["status"] in ['completed', 'failed']:
            break
        await asyncio.sleep(3)
    if status["status"] == "completed" and status["output"]:
        print('Generated images:')
        for url in status["output"]:
            print(url)
# Run the demo
await tryon_demo()

CLI Usage

newberryai tryon --model_image model.jpg --garment_image image.png --category tops
newberryai tryon --interactive
newberryai tryon --gradio

---

18. Speech-to-Speech

Environment Setup

# OpenAI Configuration
OPENAI_API_KEY=your_openai_api_key

Python SDK

# Real-time voice interaction with an AI assistant.
from newberryai.speechtospeech import RealtimeApp

# Ensure you have set your OPENAI_API_KEY environment variable
# and installed necessary audio drivers and dependencies (`pip install "openai[realtime]"`).
app = RealtimeApp()
app.run()

CLI Usage

newberryai speech_to_speech

---

19. Handwriting to Text Converter

Environment Setup

# AWS Credentials (for S3 access or AWS-powered features)
AWS_ACCESS_KEY_ID=your_aws_access_key
AWS_SECRET_ACCESS_KEY=your_aws_secret_key
AWS_DEFAULT_REGION=us-east-1

Python SDK

# Extract handwritten text from an image using AI.
from newberryai import Handwrite2Text

# Initialize the handwriting-to-text converter
handwriter = Handwrite2Text()

# Path to your handwritten document image
image_path = 'handwritten_note.jpg'

# Extract handwritten text from the image
extracted_text = handwriter.extract_text(image_path)

print("Extracted Handwritten Text:")
print(extracted_text)

CLI Usage

newberryai handwritten2text --file_path handwritten_note.jpg
newberryai handwritten2text --gradio

---

20. Image Search

Environment Setup

# AWS Credentials (for S3 access)
AWS_ACCESS_KEY_ID=your_aws_access_key
AWS_SECRET_ACCESS_KEY=your_aws_secret_key
AWS_DEFAULT_REGION=us-east-1

How it works

• Upload your images to your S3 bucket (using AWS Console, CLI, or script).
• Build the index using the CLI or Python API (this creates a local FAISS index from your S3 images, using Amazon Titan Multimodal Embeddings G1 for all embeddings).
• Search images using natural language (text-to-image, powered by Titan text embeddings) or by image (image-to-image, powered by Titan image embeddings) via CLI, Gradio, or Python API.

Python SDK

from newberryai import ImageSearch
from PIL import Image

# Initialize with your S3 bucket name
searcher = ImageSearch(s3_bucket='your-bucket-name')

# Build the index (create FAISS index from S3 images)
searcher.build_index(prefix='optional/folder/')

# Search for images by text
results = searcher.search('A cat sitting on a sofa', k=5)
for r in results:
    print(r['image_url'], r['distance'], r['folder'])

# Search for images by image
query_image = Image.open('query.jpg')
results = searcher.search_by_image(query_image, k=5)
for r in results:
    print(r['image_url'], r['distance'], r['folder'])

CLI Usage

# Build the index from your S3 images
newberryai img_search --s3_bucket your-bucket-name --build_index

# Search via CLI (choose text or image search at prompt)
newberryai img_search --s3_bucket your-bucket-name --cli

# Launch Gradio UI (text and image search, with tabs)
newberryai img_search --s3_bucket your-bucket-name --gradio

Note: You must upload your images to S3 before building the index. The tool does not upload images for you.