Skip to content

Changelog

All notable changes to this project will be documented in this file. Format based on Keep a Changelog.

[Unreleased]

[0.1.0] - 2026-06-23

Added

  • Tabular classification task: ClassificationExtractor (22-dim Problexity), ClassificationEvaluator (kernel SVC + MCC)
  • Tabular regression task: RegressionExtractor (12-dim Problexity), RegressionEvaluator (kernel Ridge + R²)
  • PairwiseRecommender: OvO kNN meta-learner, shared across task types
  • Pre-trained default recommenders shipped in Qmes/_models/: kNN_top10 (classification, regret=0.0165), kNN_top10 (regression, regret=0.0150),
  • Public API: recommend, load_default_recommender, get_extractor, get_evaluator, get_recommender, evaluate_recommendation, preprocess_new_dataset, run_loo_evaluation, filter_degenerate_datasets, PairwiseRecommender, CIRCUIT_POOL, get_circuit_names
  • 7 candidate circuits: unit, SRx, RY, HERx, RY_CX, ZFM, HD
  • Pluggable architecture via abstract base classes: BaseExtractor, BaseEvaluator
  • Bundled Qsun quantum simulator (Qsun v1.0, MIT License)
  • Test suite: 81 tests across 5 files, CI passing on Python 3.10/3.11/3.12
  • Documentation: mkdocs-material site at https://tungduy1704.github.io/Qmes/