mango-time-series

All notable changes to the mango_time_series project will be documented in this file.

The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.

[Unreleased]

[1.0.3] - 2026-10-27

Changed

  • Allow numpy>=1.24.4,<3.0.0 (was <2.0.0)

  • Depend on mango>=1.0.3,<2.0.0 instead of the exact pin mango==1.0.2

Removed

  • Dropped Python 3.10 support (end of life since 2026-10); requires-python is now >=3.11

Security

  • Require mango>=1.0.3

  • pyarrow>=23.0.1,<26.0.0 instead of pinned pyarrow==22.0.0 (Arrow IPC use-after-free advisory; pyarrow 26 requires NumPy 2)

  • Refreshed uv.lock to patched versions of vulnerable dependencies

[1.0.1] - 2025-09-29

Added

  • Updated mango dependency version

  • Core time series data structures and classes

  • Exploratory analysis tools for time series data

  • Data validation and processing utilities

  • Moving averages and time feature extraction

  • Seasonal decomposition and stationarity tests

  • Heteroscedasticity analysis

  • Custom cross-validation for time series forecasting

  • Comprehensive logging and error handling

  • Enhanced time series analysis capabilities

  • Improved data processing and validation functions

  • Better integration with the mango ecosystem

Changed

  • Updated dependencies and package configuration

  • Enhanced test coverage and documentation

Features

  • Time series data loading and manipulation

  • Statistical analysis and visualization

  • Feature engineering for machine learning

  • Forecasting metrics and evaluation

  • Calendar and holiday handling

  • Efficient data processing with pandas and numpy