smart-cloud-tag 0.1.2

AI-powered cloud storage object tagging using LLMs

Smart Cloud Tag

Automatically tag cloud files across AWS, Azure, and Google Cloud with the least amount of effort.

Use Case

smart-cloud-tag is a multi-cloud tagging solution that automatically applies tags to objects in batch across AWS S3, Azure Blob Storage, and Google Cloud Storage using LLMs (GenAI). It provides end-to-end automation, from reading file content to applying tags. As expected, LLMs do a great job in predicting tags. This tool eradicates the need to manually go through files one by one to add metadata to them, or to build your own custom solution. Now the work you would need to do to tag several objects in the cloud of your choice, would take less time and effort than making your morning cup of coffee.

Architecture

Features

• Multi-Cloud Support: AWS S3, Azure Blob Storage, Google Cloud Storage
• AI-Powered: Uses OpenAI, Anthropic Claude, or Google Gemini for intelligent tagging
• File Type Support: Currently supports .txt, .json, .csv, and .md files
• Simple & Flexible: Designed to work out-of-the-box while remaining flexible for custom requirements
• Auto-Detection: Automatically detects storage provider from URI prefix
• Batch Processing: Process multiple files with one command
• Preview Mode: Preview tags before applying them (optional)
• Custom Prompts: Ability to use your own custom LLM prompt templates (optional)

Quick Start

Installation

Basic Installation

pip install smart_cloud_tag

Note: Basic installation includes AWS S3 and OpenAI support. For other cloud providers or LLM providers, use the optional dependencies below.

Installation with Optional Dependencies

You can install additional dependencies based on your needs:

# Install with all optional dependencies (recommended)
pip install smart_cloud_tag[all]

# Install with specific cloud providers
pip install smart_cloud_tag[aws]      # AWS S3 (included by default)
pip install smart_cloud_tag[azure]    # Azure Blob Storage
pip install smart_cloud_tag[gcp]      # Google Cloud Storage

# Install with specific LLM providers
pip install smart_cloud_tag[openai]   # OpenAI (included by default)
pip install smart_cloud_tag[anthropic] # Anthropic Claude
pip install smart_cloud_tag[gemini]   # Google Gemini

# Combine multiple options
pip install smart_cloud_tag[azure,anthropic]  # Azure + Anthropic
pip install smart_cloud_tag[gcp,gemini]       # GCP + Gemini

Installation Options:

• [all] - Installs all optional dependencies (all cloud providers + LLM providers)
• [aws] - AWS S3 support (included by default)
• [azure] - Azure Blob Storage support
• [gcp] - Google Cloud Storage support
• [openai] - OpenAI LLM support (included by default)
• [anthropic] - Anthropic Claude LLM support
• [gemini] - Google Gemini LLM support
• [dev] - Development dependencies (testing, linting, formatting)

Basic Usage

from smart_cloud_tag import SmartCloudTagger

# Initialize the tagger
tagger = SmartCloudTagger(
    storage_uri="az://telehealthcanada",  # target bucket location
    tags={
        "protected_health_information": ["T", "F"],  # allowed values are T/F
        "document_type": ["chat_transcript", "lab_summary", "claim"],
    },  # tag schema 
)

# Preview tags before applying (optional)
preview_result = tagger.preview_tags()
print(f"Preview: {preview_result.summary}")

# Apply tags
result = tagger.apply_tags()
print(f"Applied tags to {result.summary['applied']} objects")

Configuration

Environment Variables

Create a .env file:

# LLM Provider API Key (used for all providers)
API_KEY=your_api_key_here

# AWS (if using S3)
AWS_ACCESS_KEY_ID=your_access_key
AWS_SECRET_ACCESS_KEY=your_secret_key
AWS_REGION=us-east-1

# Azure (if using Blob Storage)
AZURE_STORAGE_CONNECTION_STRING=your_connection_string

# Google Cloud (if using GCS)
GOOGLE_APPLICATION_CREDENTIALS=path/to/service-account.json

Supported Storage Providers

Provider	URI Format	Example
AWS S3	s3://bucket	s3://my-documents
Azure Blob	az://container	az://documents
Google Cloud	gs://bucket	gs://my-files

Supported LLM Providers

Provider	Default Model	Environment Variable
OpenAI	gpt-5	API_KEY
Anthropic	claude-3-opus-4.1	API_KEY
Google Gemini	gemini-1.5-pro	API_KEY

Advanced Usage

SmartCloudTagger Parameters

Parameter	Type	Required	Default	Description
storage_uri	str	Yes	-	Storage location URI (s3://, az://, or gs://)
tags	Dict[str, Optional[List[str]]]	Yes	-	Tag schema with keys and allowed values. If allowed values are missing for a key, LLM will deduce appropriate values
llm_model	str	No	Provider-specific	LLM model to use (see supported models below)
llm_provider	str	No	"openai"	LLM provider: "openai", "anthropic", or "gemini"
max_bytes	Optional[int]	No	5000	Maximum bytes to read from each object/file
custom_prompt_template	Optional[str]	No	config.py	Custom prompt template. Must include placeholders: {content}, {filename}, {tags} (see config.py for default)

Default Models by Provider

Provider	Default Model
OpenAI	gpt-5
Anthropic	claude-3-opus-4.1
Google Gemini	gemini-1.5-pro

Different LLM Providers

# Using Anthropic Claude
tagger = SmartCloudTagger(
    storage_uri="s3://my-bucket",
    tags=tags,
    llm_provider="anthropic"
)

# Using Google Gemini
tagger = SmartCloudTagger(
    storage_uri="s3://my-bucket", 
    tags=tags,
    llm_provider="gemini"
)

Development

Installation from Source

git clone https://github.com/yourusername/smart_cloud_tag.git
cd smart_cloud_tag
pip install -e ".[all]"

Note: Use quotes around ".[all]" to prevent shell expansion issues in zsh and other shells.

Running Tests

python -m pytest tests/

License

MIT License - see LICENSE file for details.

Support

• 📧 Email: [e-mail]
• 🐛 Issues: GitHub Issues