py-mando

PyO3-based Python extension module for the Mando workspace. Exposes high-performance Rust components to Python for algorithm execution, data processing, and service integration.

Structure

py-mando/
├── Cargo.toml
├── pyproject.toml       # Maturin + Poetry config
├── src/
│   ├── lib.rs           # Module registration
│   ├── adapter.rs       # Backend service impls
│   ├── algo.rs          # Algorithm params (1,143 lines)
│   ├── algo_runner.rs   # HTTP algo runner/server
│   ├── polars.rs        # DataFrame ops (1,185 lines)
│   ├── data_point.rs    # DataPoint config/metadata
│   ├── validation.rs    # Input validation
│   ├── test.rs          # Test utilities
│   ├── util.rs          # General utilities
│   ├── bess_util.rs     # BESS-specific utils
│   ├── codegen.rs       # Client code generation
│   └── execution.rs     # Execution context
├── test/                # Python tests
└── example/notebooks/   # Jupyter notebooks (nbval)

Python API

Initialization

from py_mando import init, version
 
init()           # Sets up tracing, rustls
print(version()) # Returns Rust crate version

Key Classes

PyClass	Module	Purpose
AlgoParamsConfig	algo	Wraps JSON algorithm configuration
DataPointConfig	data_point	Data point metadata
PyMandoDataPointCalculation	data_point	Calculation definitions
AlgoRunner	algo_runner	HTTP service for running algorithms
ServerConfig	algo_runner	Server configuration

Re-exported from mando-lib

from py_mando import (
    DataPointId,
    DataPointFilter,
    FlowFilter,
    MandoRestConfig,
)

Adapters

Backend service implementations exposed to Python:

• Polars adapter — DataFrame-based operations
• PostgreSQL adapter — Direct DB access
• REST adapter — HTTP client for mando-bess API

Polars Operations

The polars module (1,185 lines) provides:

• Filtering and grouping
• Joins
• Time functions
• Resolution conversion
• DataFrame transformations

Algorithm Parameters

The algo module (1,143 lines) defines:

• Algorithm parameter configurations
• MOL (Machine Operations Learning) params
• Market aggregate definitions

Build Configuration

Setting	Value
Build tool	Maturin 1.9.2
Python ABI	abi3-py38 (stable across versions)
Python requirement	>= 3.10
Crate type	cdylib (shared library)

Python Dependencies

Package	Version
pandas	2.3.1
polars	1.31.0
PyArrow	21.0.0
PyYAML	6.0.2

Building

cd py-mando
maturin develop        # Dev build
maturin build --release  # Release wheel

Testing

cd py-mando
pytest test/
pytest --nbval example/notebooks/  # Notebook validation

Data Bridge

Zero-copy data exchange between Rust and Python using:

• pyo3-polars — Polars DataFrame bridge
• pyo3-arrow — Arrow array bridge
• pyo3-async-runtimes — Tokio async integration

This allows Python algorithms to operate on Rust-managed data without serialization overhead.