Add parametric validation framework

[ilkilic]

Summary

Introduces a composable, parametric validation framework alongside the existing validation.py. Users can define custom electrophysiology validations by composing Protocol + Measurement + Criterion — without modifying BlueCelluLab source code.

Motivation

The current validation.py contains hardcoded validation tests (fixed protocols, fixed eFEL features, fixed pass/fail logic). Adding a new validation requires writing a new function from scratch. This framework makes it trivial to define parametric validations that can be configured at runtime.

What's added

bluecellulab/validation/
├── __init__.py          # Package exports
├── base.py              # ValidationOutcome dataclass + ValidationTest ABC
├── protocol.py          # Protocol ABC + StepProtocol (IDRest at X% rheobase)
├── measurement.py       # Measurement ABC + EfelMeasurement (any eFEL feature)
├── criterion.py         # Criterion ABC + GreaterThan (numpy-array-safe)
├── parametric_test.py   # ParametricValidation: composes protocol+measurement+criterion
├── plotting.py          # plot_trace utility (extracted from validation.py)
└── validation.py        # UNCHANGED — backward compatible

Usage

from bluecellulab.validation import (
    ParametricValidation, StepProtocol, EfelMeasurement, GreaterThan,
)

# Define a custom validation: spike count > 0 at 130% rheobase
test = ParametricValidation(
    validation_name="Spiking Validation",
    protocol=StepProtocol(threshold_percentage=130.0),
    measurement=EfelMeasurement(feature_name="Spikecount"),
    criterion=GreaterThan(threshold=0),
)

outcome = test.run(cell.template_params, rheobase, out_dir)
# outcome.name, outcome.passed, outcome.details, outcome.figures

Design decisions

• Existing validation.py is untouched — no breaking changes
• numpy-safe criteria — GreaterThan handles both scalar and array-valued measurements (e.g., AP_amplitude returns one value per spike)
• Composable — any combination of protocol, measurement, and criterion can be assembled without subclassing
• Extensible — new protocols, measurements, and criteria can be added independently

[codecov]

Codecov Report

❌ Patch coverage is 94.81268% with 18 lines in your changes missing coverage. Please review.

Files with missing lines	Patch %	Lines
bluecellulab/validation/protocol.py	67.92%	17 Missing ⚠️
bluecellulab/validation/criterion.py	94.73%	1 Missing ⚠️

Files with missing lines	Coverage Δ
bluecellulab/validation/__init__.py	100.00% <100.00%> (ø)
bluecellulab/validation/base.py	100.00% <100.00%> (ø)
bluecellulab/validation/measurement.py	100.00% <100.00%> (ø)
bluecellulab/validation/parametric_test.py	100.00% <100.00%> (ø)
bluecellulab/validation/plotting.py	100.00% <100.00%> (ø)
tests/test_validation/test_parametric_framework.py	100.00% <100.00%> (ø)
bluecellulab/validation/criterion.py	94.73% <94.73%> (ø)
bluecellulab/validation/protocol.py	67.92% <67.92%> (ø)

... and 2 files with indirect coverage changes

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.