Official Python package for advanced voice transcription with multi-language support. Dhvagna-NPI offers an end-to-end voice transcription suite for developers building speech recognition applications. You can easily transcribe spoken audio to text in multiple languages with a beautiful command-line interface.

With this package, you can easily perform voice transcription from any Python 3.7+ application by utilizing the provided command-line interface or importing the modules directly. Currently, Dhvagna-NPI supports both quick single transcriptions and interactive sessions with customizable settings.

To learn how to use our package, check out our documentation.

---

Table of Contents

• Installation
• Get Started
• Features
• Creating Your First Transcription
• Using the Command-Line Interface
• Customizing Settings
• Supported Languages
• Examples
  • Minimal Example
  • Quick Transcribe Example
  • Custom Settings Example
  • Save to File Example
  • Multilingual Example
• Available Methods
• Documentation
• License

---

Installation

To install the latest version available:

pip install dhvagna-npi

When using this package in your application, make sure to pin to at least the major version (e.g., ==0.1.*). This helps ensure your application remains stable and avoids potential issues from breaking changes in future updates.

---

Get Started

Dhvagna-NPI provides a simple way to transcribe voice to text:

from dhvagna_npi.core import transcribe_audio
import speech_recognition as sr

# Initialize recognizer and microphone
recognizer = sr.Recognizer()
microphone = sr.Microphone()

# Record and transcribe
with microphone as source:
    print("Listening... Speak now!")
    recognizer.adjust_for_ambient_noise(source)
    audio = recognizer.listen(source)
    success, text, _ = transcribe_audio(audio, recognizer, "en-US")
    if success:
        print(f"Transcription: {text}")

---

Features

• Multi-language Support: Transcribe speech in English, Spanish, French, German, Chinese, Hindi, Telugu, and more.
• Offline Processing: All processing happens locally - your voice data never leaves your computer.
• Rich Command-line Interface: Beautiful terminal UI with the Rich library.
• Customizable Settings: Adjust language, microphone sensitivity, timeout, and more.
• Transcription History: Save and retrieve past transcriptions.
• Multiple Export Formats: Save transcriptions as text, JSON, or CSV.

---

Creating Your First Transcription

from dhvagna_npi.config import Config
from dhvagna_npi.core import run_single_transcription

def main():
    # Run a single transcription and exit
    run_single_transcription()

if __name__ == "__main__":
    main()

---

Using the Command-Line Interface

Dhvagna-NPI provides two CLI modes:

# Quick mode (single transcription)
dhvagna-npi

# Interactive mode (multiple transcriptions with settings)
dhvagna-npi-interactive

Keyboard Controls

• k (or custom hotkey): Start/Stop recording
• h: Show help
• s: Settings menu
• v: View recent transcriptions
• q: Quit

---

Customizing Settings

from dhvagna_npi.config import Config

# Create and customize configuration
config = Config()
config.language = "es-ES"  # Set language to Spanish
config.timeout = 15  # Set recording timeout to 15 seconds
config.energy_threshold = 250  # Set microphone sensitivity
config.save_transcriptions = True  # Enable saving transcriptions
config.transcription_folder = "my_transcriptions"  # Set custom folder

# Save configuration
config.save()

---

Supported Languages

• English (US) - en-US
• English (UK) - en-GB
• Spanish - es-ES
• French - fr-FR
• German - de-DE
• Chinese (Mandarin) - zh-CN
• Hindi - hi-IN
• Telugu - te-IN

---

Examples

Minimal Example

import speech_recognition as sr
from dhvagna_npi.core import transcribe_audio

# Initialize recognizer and microphone
recognizer = sr.Recognizer()
microphone = sr.Microphone()

# Record and transcribe
with microphone as source:
    print("Listening... Speak now!")
    recognizer.adjust_for_ambient_noise(source)
    audio = recognizer.listen(source)
    success, text, _ = transcribe_audio(audio, recognizer, "en-US")
    if success:
        print(f"Transcription: {text}")

Quick Transcribe Example

from dhvagna_npi.core import run_single_transcription

# Run a single transcription with default settings
run_single_transcription()

Custom Settings Example

import os
from dhvagna_npi.config import Config
from dhvagna_npi.core import transcribe_audio
import speech_recognition as sr

# Create custom configuration
config = Config()
config.language = "fr-FR"  # French
config.energy_threshold = 250  # More sensitive
config.timeout = 8  # 8 second timeout

# Initialize recognizer with custom settings
recognizer = sr.Recognizer()
recognizer.energy_threshold = config.energy_threshold
microphone = sr.Microphone()

# Record and transcribe
with microphone as source:
    print("Adjusting for ambient noise...")
    recognizer.adjust_for_ambient_noise(source)
    
    print("Listening... Speak now!")
    audio = recognizer.listen(source, timeout=config.timeout)
    
    success, text, _ = transcribe_audio(audio, recognizer, config.language)
    if success:
        print(f"French transcription: {text}")

Save to File Example

import json
import os
import datetime
from dhvagna_npi.core import transcribe_audio
import speech_recognition as sr

def save_as_text(text, language, output_dir="output"):
    """Save transcription as a text file."""
    if not os.path.exists(output_dir):
        os.makedirs(output_dir)
        
    timestamp = datetime.datetime.now().strftime("%Y%m%d_%H%M%S")
    filename = f"transcription_{timestamp}.txt"
    filepath = os.path.join(output_dir, filename)
    
    with open(filepath, 'w', encoding='utf-8') as f:
        f.write(f"Timestamp: {datetime.datetime.now().strftime('%Y-%m-%d %H:%M:%S')}\n")
        f.write(f"Language: {language}\n\n")
        f.write(text)
        
    return filepath

# Initialize recognizer and microphone
recognizer = sr.Recognizer()
microphone = sr.Microphone()

# Record and transcribe
with microphone as source:
    print("Listening...")
    recognizer.adjust_for_ambient_noise(source)
    audio = recognizer.listen(source)
    
    success, text, _ = transcribe_audio(audio, recognizer, "en-US")
    if success:
        # Save the transcription
        filepath = save_as_text(text, "en-US")
        print(f"Saved transcription to: {filepath}")

Multilingual Example

from dhvagna_npi.config import Config
from dhvagna_npi.utils import get_language_codes
from dhvagna_npi.core import transcribe_audio
import speech_recognition as sr

# Get available language codes
language_codes = get_language_codes()
print("Available languages:")
for key, language in language_codes.items():
    print(f"{key}: {language}")

# Choose a language code
language = "es-ES"  # Spanish

# Initialize recognizer and microphone
recognizer = sr.Recognizer()
microphone = sr.Microphone()

# Record and transcribe
with microphone as source:
    print(f"Recording in {language}...")
    recognizer.adjust_for_ambient_noise(source)
    audio = recognizer.listen(source)
    
    success, text, _ = transcribe_audio(audio, recognizer, language)
    if success:
        print(f"Transcription in {language}: {text}")

---

Available Methods

from dhvagna_npi.config import Config
from dhvagna_npi.history import TranscriptionHistory
from dhvagna_npi.utils import get_language_name

# Create a configuration
config = Config()

# Get language name from code
language_name = get_language_name("fr-FR")  # Returns "French"

# Create a history manager
history = TranscriptionHistory(config)

# Add a transcription to history
history.add("This is a test transcription")

# Get recent transcriptions
recent = history.get_recent(5)  # Get last 5 transcriptions

---

Documentation

For more detailed information about Dhvagna-NPI, check out the following resources:

• GitHub Repository
• PyPI Package
• Examples in the examples/ directory.

---

License

This project is licensed under the MIT License - see the LICENSE file for details.