sibi-flux 2026.2.4

Sibi Toolkit: A collection of tools for Data Analysis/Engineering.

SibiFlux: The Production Data Toolkit

SibiFlux is a production-grade data engineering toolkit designed for Python 3.11+. It provides a unified ecosystem for project scaffolding, declarative datacube management, automated resource generation, and agentic interoperability through the Model Context Protocol (MCP).

Key Features

• Declarative Datacubes: Define data structures in YAML and auto-generate specialized Python classes.
• Automated Init Engine: Transform modular YAML configs and .env files into strongly-typed Pydantic settings.py.
• High-Performance Storage: Unified fsspec and PyArrow storage registry with depot isolation and path traversal protection.
• Agentic Ready: Native MCP (Model Context Protocol) integration to expose your data pipelines directly to AI agents.
• Scalable Workloads: First-class support for Dask and distributed computing.
• Geospatial Power: Integrated OSMnx helpers and PBF handling for complex GIS workflows.

---

Tech Stack

• Core: Python 3.11+
• Configuration: Pydantic Settings v2
• CLI Framework: Typer
• Data Engineering: Pandas, PyArrow, Dask
• Storage Layer: fsspec, s3fs
• Databases: SQLAlchemy, Clickhouse, MySQL, PostgreSQL
• Web/API: FastAPI, Uvicorn, MCP
• Orchestration: Docker, Docker Compose, uv

---

Prerequisites

• Python 3.11+
• uv: Extremely fast Python package manager.
• Docker & Docker Compose: For local services (Postgres, Clickhouse).
• Poethepoet: Task runner (installed via dev dependencies).

---

Getting Started

1. Clone & Install

git clone https://github.com/lvalverdeb/sibi-flux.git
cd sibi-flux
uv sync --all-extras

2. Environment Setup

Copy the example environment file:

cp .env.linux .env

Initialize the project configuration:

# This generates conf/settings.py and conf/credentials/
sibi-flux env --env-file .env

3. Start Local Infrastructure

docker-compose up -d

4. Project Scaffolding

# Initialize a new project structure
sibi-flux init my_project --app

# Create a specific application within the project
sibi-flux create_app my_app

5. Datacube Workflow

# Propose cubes from a database domain
sibi-flux propose-cubes my_db_domain my_app

# Generate the app-specific extensions
sibi-flux create-cubes my_app

---

Architecture

Directory Structure

├── src/sibi_flux/
│   ├── init/             # Project bootstrapping & logic generation
│   ├── datacube/         # Datacube orchestration, discovery & mapping
│   ├── storage/          # Unified storage manager (fsspec + pyarrow)
│   ├── mcp/              # MCP Router & Resource registration
│   ├── osmnx_helper/     # Geospatial PBF handling & Graph loading
│   ├── config/           # Unified settings & config management
│   ├── parquet/          # Optimized parquet I/O with gatekeeping
│   ├── df_helper/        # Rich DataFrame manipulation utils
│   └── orchestration/    # Dask/Distributed workload management
├── conf/                 # Generated configuration (settings.py)
├── solutions/            # Reference implementations & examples
├── test_prj/             # Standard test project layout
└── Dockerfile            # Production-grade container image

Data Lifecycle

1. Declarative Phase: Developers define discovery rules and datacube mappings in conf/discovery_params.yaml.
2. Generation Phase: sibi-flux dc commands scan data sources and generate Python wrappers.
3. Runtime Phase:
   • ProjectService detects the root and loads environment-specific settings.py.
   • StorageManager establishes connections to S3/WebDAV/Local.
   • Datacube instances provide high-level abstractions over PyArrow/Dask.
4. Exposure Phase: The MCP Router registers datacubes as resources for external consumption (e.g., by LLM agents).

Unified Storage Registry

SibiFlux implements a robust storage registry that isolates distinct "depots" (e.g., bronze, silver, gold).

• Isolation: Each depot is mapped to a designated path.
• Performance: Automatic switching between fsspec and Native C++ PyArrow S3FileSystem.
• Security: Built-in path traversal protection (blocking .. in joined segments).

---

Environment Configuration

Required Variables

Variable	Description	Default
DB_URL	Primary database connection string	sqlite:///:memory:
FS_TYPE	Storage type (s3, file, webdav)	s3
STORAGE_PATH	Base path for all data artifacts	-

Modular YAMLs

The init engine populates conf/ with modular YAML files:

• storage.yaml: S3 layouts and filesystem definitions.
• databases.yaml: Database connection mapping.
• osmnx.yaml: Geospatial storage references.

---

Available Scripts (Poe)

Command	Description
poe dev	Start development server with reload
poe test	Run the full pytest suite
poe dc-sync	Synchronize datacube registry
poe dc-discover	Discover new data structures
poe test-snapshots	Run regression snapshot tests

---

Testing

SibiFlux uses complex snapshot-based testing for code generation verification.

# Run all tests
pytest tests/

# Update snapshots (Golden Masters)
UPDATE_SNAPSHOTS=1 pytest

---

Deployment

Docker

The project includes a production-ready multi-stage Dockerfile.

docker build -t sibi-flux:latest .

Docker Compose

The docker-compose.yml orchestrates:

• API: The main SibiFlux web service (FastAPI).
• Dask Scheduler: For distributed task execution.
• Postgres/Clickhouse: Local data stores for development.

---

Troubleshooting

• ImportError: Ensure you have run uv sync to install all dependencies.
• Connection Error: Check your .env variables and ensure Docker services are running.
• Storage Risk: If you see "Security Risk: Path traversal detected," ensure your dynamic paths do not contain .. or absolute prefixes.

---

Contributing

1. Follow the Conventional Commits standard.
2. Ensure all tests pass (poe test).
3. Add snapshots for any new code generation logic.