common-ml-functions

Common ML Functionalities API

Python package which provides useful and common interactions with ML functionalities
such as detecting language, chatbots, link transcriptions, audio enhancements, and much more.

Powered by Sieve, Langchain

Features

• Translation (sieve/seamless_text2text): Translates text from a source language to a target language.
• Youtube Link Transcription (langchain/youtube_transcripts): Transcribes audio from a given URL.
• Detect Language (sieve): Identifies the language of the given text.
• Text to Speech (sieve): Converts text into spoken words.
• Translate and Talk (sieve): Translates text from one language to another and then converts the translated text to speech.
• Youtube Video Download (sieve): Downloads a Youtube video in MP4 format given its URL.
• Detect Language and Translate: Detects the language of the text and translates it to a specified target language.
• Audio Enhancement (sieve): Enhances the quality of an mp3 or wav file.
• Visual question answering (sieve/llava-vl-13b): Visual question answering with GPT-4 level capabilities.

Setup

To set up the package, follow these steps:

1. Install the Package: Install this package using pip: pip install common-ml-functions

2. Import in Your Project: Import the required functionalities in your Python project:

from common-ml-functions.translation import translate
from common-ml-functions.text_to_speech import convert_text_to_speech
# ... other imports as needed

Usage

The package offers a range of functionalities that can be easily integrated into Python projects. Here are some of the features and how to use them:

Translation: Use translate(original_text, source_language, target_language) to translate text.

from common-ml-functions.language import translate

translated_text = translate("Hello, world!", "en", "es")
print(translated_text)

Youtube Link Transcription: Use yb_transcript(url) to get transcriptions of YouTube videos.

from common-ml-functions.yb import yb_transcript

transcript = yb_transcript("https://www.youtube.com/watch?v=example")
print("Transcript:", transcript)

Detect Language: Use detect(text) to identify the language of a given text.

from common-ml-functions.language import detect

language_code = detect("Ceci est un texte en français.")
print(f"The detected language code is: {language_code}")

Text to Speech: Use convert_text_to_speech(text) to convert text into speech.

from common-ml-functions.text_to_speech import convert_text_to_speech

audio_file_path = convert_text_to_speech("Hello, this is a test.")
print(f"Audio file saved at: {audio_file_path}")

Translate and Talk: Use translate_and_talk(text, target_language) to translate and convert text to speech.

from common-ml-functions.text_to_speech import translate_and_talk

audio_file_path = translate_and_talk("This is a test.", "es")
print(f"Translated and converted to speech, file saved at: {audio_file_path}")

Youtube Video Download: Use yb_download(url) to download YouTube videos.

from common-ml-functions.yb import yb_download

video_path = yb_download("https://www.youtube.com/watch?v=example")
print(f"Video downloaded at: {video_path}")

Detect Language and Translate: Combine detect and translate for detecting and translating text.

from common-ml-functions.language import detect_and_translate

translated_text = detect_and_translate("Bonjour le monde!", "en")
print(translated_text)

Audio Enhancement: Use enhance_audio(audio_url, speed_boost, steps) for audio enhancement.

from common-ml-functions.audio import enhance_audio

enhanced_audio_path = enhance_audio("http://example.com/audio.mp3", True, 50)
print(f"Enhanced audio saved at: {enhanced_audio_path}")

Visual question answering: Use visual_question_answering(image_url, prompt) for visual Q&A.

from common-ml-functions.visual import visual_question_answering

answer = visual_question_answering("http://example.com/image.jpg", "What is depicted in this image?")
print(answer)

Contributing!

Please consider contributing to this young and humble project! Email any questions at [e-mail]