animuz-core 0.1.5

Core shared utilities for Animuz RAG system - LLM clients, pipelines, vector DB, and document ingestion

animuz-core

Core shared utilities for Animuz RAG (Retrieval-Augmented Generation) system.

Features

• LLM Clients: OpenAI, Anthropic Claude, Ollama
• RAG Pipelines: Simple and Agentic RAG implementations
• Vector Database: Qdrant integration with hybrid search (dense + sparse)
• Embedding Clients: Multiple providers (local server, Modal, S3/SageMaker)
• Document Ingestion: Azure Document Intelligence, Unstructured, PDF extraction, structured text parsing
• CloudWatch Logging: Structured JSON logging with watchtower

Requirements

• Python >= 3.10

Installation

Install the core package (minimal dependencies only):

pip install animuz-core

Then install only the extras you need:

# Single extra
pip install animuz-core[openai]

# Multiple extras
pip install animuz-core[openai,qdrant,aws]

# Everything
pip install animuz-core[all]

Works with uv too:

uv pip install animuz-core[openai,qdrant]

Available Extras

Extra	What it installs	Use when you need
openai	openai	OpenAI GPT models
anthropic	anthropic	Anthropic Claude models
ollama	ollama	Local LLMs via Ollama
qdrant	qdrant-client	Qdrant vector database
aws	boto3, aiobotocore, watchtower, sagemaker	S3, SageMaker embeddings, CloudWatch logging
azure	azure-ai-documentintelligence	Azure Document Intelligence for PDF ingestion
ingest	unstructured-client, PyMuPDF	Document parsing (Unstructured API, PDF extraction)
fastapi	fastapi	Streaming SSE endpoints
all	All of the above	Everything
dev	all + pytest, black, ruff, mypy	Development and testing

Usage

Unified RAG API

from animuz_core import RAG, RAGConfig, tool

@tool(description="Lookup a user by ID")
def lookup_user(user_id: str) -> str:
    return f"User {user_id}"

config = RAGConfig.from_env().with_defaults()
rag = RAG(config=config, tools=[lookup_user])

await rag.add_doc("docs/intro.md", user_chat_id="demo")
response = await rag.chat("What is this project?", user_chat_id="demo")

Example script:

python scripts/quickstart_rag.py

LLM Clients

from animuz_core.genai import OpenAIAgentClient, AnthropicAgentClient, OLlamaClient

# OpenAI agent with tool use
agent = OpenAIAgentClient(tools=my_tools)
response = await agent.chat(messages, model="gpt-4o")

# Anthropic agent with tool use
agent = AnthropicAgentClient(tools=my_tools)
response = await agent.chat(messages, model="claude-sonnet-4-20250514")

# Ollama (local)
client = OLlamaClient()
response = await client.chat(messages, model="llama3")

RAG Pipelines

from animuz_core.pipelines import AgenticRAG, SimpleRAG

# Agentic RAG - LLM decides when to call the retriever
pipeline = AgenticRAG(
    llm=agent,
    embedding_client=embedding_client,
    qdrant_client=qdrant_client,
)
result = await pipeline.run(query="What is RAG?", user_chat_id="tenant-123")

# Simple RAG - always retrieves then generates
pipeline = SimpleRAG(
    llm=client,
    embedding_client=embedding_client,
    qdrant_client=qdrant_client,
)
result = await pipeline.run(query="What is RAG?", user_chat_id="tenant-123")

Vector Database

from animuz_core.vectordb import QdrantDBClient

client = QdrantDBClient()
# Hybrid search with multi-tenant isolation
results = await client.search(
    dense_vector=dense_vec,
    sparse_vector=sparse_vec,
    user_chat_id="tenant-123",
    limit=5,
)

Embedding

from animuz_core.embedding import EmbeddingClient, ModalEmbeddingClient, S3EmbeddingClient

# Local embedding server
client = EmbeddingClient()
dense, sparse = await client.embed("Some text to embed")

# Modal-hosted embeddings
client = ModalEmbeddingClient()

# S3/SageMaker embeddings
client = S3EmbeddingClient()

Document Ingestion

from animuz_core.ingest import AzureDocAiClient, MyUnstructuredClient, Structured

# Azure Document Intelligence (PDFs)
azure_client = AzureDocAiClient()
text = await azure_client.extract("document.pdf")

# Unstructured API
unstructured = MyUnstructuredClient()
chunks = await unstructured.ingest("document.docx")

# Structured text (txt, md, csv)
structured = Structured()
chunks = structured.split("document.txt")

Top-level Imports

All main classes are re-exported from the package root:

from animuz_core import (
    RAG, RAGConfig, tool,
    AgenticRAG, SimpleRAG,
    OpenAIAgentClient, OpenAILLMClient, AnthropicClient, AnthropicAgentClient, OLlamaClient,
    QdrantDBClient,
    EmbeddingClient,
    AzureDocAiClient, MyUnstructuredClient, Structured,
)

Development

# Clone and install in editable mode with dev dependencies
git clone <repo-url>
cd animuz-core
pip install -e ".[dev]"

# Run tests
pytest tests/

# Run integration tests (requires external services + env vars)
pytest -m integration tests/integration/
pytest -m integration tests/integration/test_e2e_rag_wrapper_simple.py

# Format
black src/
ruff check src/

Publishing to PyPI

1. Bump the version in pyproject.toml and __init__.py.

2. Build the package:

uv pip install --upgrade build # python -m pip install --upgrade build
uv run python -m build # python -m build

3. (Optional) Verify the artifacts:

uv pip install --upgrade twine # python -m pip install --upgrade twine
uv run python -m twine check dist/* # python -m twine check dist/*

4. Upload to TestPyPI first:

uv run python -m twine upload -r testpypi dist/* # python -m twine upload -r testpypi dist/*

5. Upload to PyPI:

uv run python -m twine upload dist/* # python -m twine upload dist/*

Notes:

• Create a PyPI API token and set TWINE_USERNAME=__token__ and TWINE_PASSWORD=<your-token>.
• If you upload to TestPyPI, install with pip install -i https://test.pypi.org/simple animuz-core to verify.

Integration Test Setup (Qdrant)

Use Docker Compose to run Qdrant locally:

docker compose -f docker-compose-qdrant.yml up -d qdrant

Then set the Qdrant env vars (example):

export QDRANT_HOST=localhost
export QDRANT_PORT=6333

Environment Variables

The package reads configuration from environment variables (loaded via python-dotenv):

Variable	Used by
OPENAI_API_KEY	OpenAI client
ANTHROPIC_API_KEY	Anthropic client
QDRANT_HOST, QDRANT_PORT, QDRANT_COLLECTION_NAME	Qdrant client
QDRANT_CLOUD_API_KEY	Qdrant Cloud
EMBEDDING_HOST, EMBEDDING_PORT	Embedding client
AZURE_DOCAI_KEY, AZURE_DOCAI_ENDPOINT	Azure Document Intelligence
UNSTRUCTURED_ENDPOINT, UNSTRUCTURED_API_KEY	Unstructured client
S3_BUCKET_NAME, S3_DOWNLOAD_DIR	S3 operations

@tool decorator API

from animuz_core import tool

@tool(description="Search documents")
async def rag(query: str, user_chat_id: str) -> str:
    ...

agent = Agent(model="gpt-4o", tools=[rag])
response = await agent.chat(messages, system_prompt="You are helpful.")

License

MIT