fix: deterministic expenv behaviour

[iraedeus]

No description provided.

[Copilot]

Pull Request Overview

This PR aims to improve determinism across the experimental environment by standardizing seed usage, updating estimator naming, enhancing the error summarizer, and adding an end-to-end reproducibility test.

• Consolidate and propagate RNG seeding in dataset generators and executors for reproducible experiments
• Rename LMomentsEstimator.name from "LM-EM" to "ELM" and adjust related tests
• Refactor error summarizer to use NumPy functions and handle empty-result cases

Reviewed Changes

Copilot reviewed 12 out of 14 changed files in this pull request and generated 3 comments.

Show a summary per file

File	Description
tests/experimental_env/test_reproducibility_expenv.py	Added full reproducibility scenario test
tests/core/test_problem.py	Added Hypothesis tests for Problem
tests/core/test_mixture_distribution.py	Extended unit and integration tests for mixture distribution
tests/core/test_distribution.py	Extended unit and integration tests for distribution
experimental_env/preparation/dataset_generator.py	Moved RNG seeding into base RandomDatasetGenerator
experimental_env/mixture_generators/utils.py	Tweaked default Weibull parameters
experimental_env/mixture_generators/abstract_generator.py	Removed default seeding from abstract generator
experimental_env/experiment/experiment_executors/standart_executor.py	Dropped seed forwarding to standard mixture generator
experimental_env/experiment/experiment_executors/random_executor.py	Dropped seed forwarding to random mixture generator
experimental_env/experiment/experiment_executors/abstract_executor.py	Added seed-based RNG initialization but missing NumPy import
experimental_env/experiment/estimators.py	Renamed L-moments estimator identifier from "LM-EM" to "ELM"
experimental_env/analysis/analyze_summarizers/error_summarizer.py	Switched to NumPy for stats and added empty-list guard

Comments suppressed due to low confidence (3)

experimental_env/preparation/dataset_generator.py:63

• ConcreteDatasetGenerator no longer seeds the RNG, breaking deterministic behavior. Re-add random.seed(self._seed) and np.random.seed(self._seed) inside its __init__ to maintain reproducibility.

    def __init__(self, seed: int = 42):

tests/core/test_problem.py:48

• [nitpick] The docstring is in Russian, which may be unclear to other maintainers. Consider translating it to English for consistency.

        """Тест инициализации Problem."""

experimental_env/experiment/experiment_executors/abstract_executor.py:37

• The module references np.random but does not import NumPy. Add import numpy as np at the top to avoid a NameError.

        np.random.seed(self._seed)