feat: multi-dataset evaluator with domain union utilities (closes #196, ref #132)

[aditya-pandey-dev]

Implements the evaluator hook requested in #196 and addresses the
single-model-multiple-data problem from #132.

Changes

New files

• src/multi-data/domain_utils.jl — domain_union and domains_are_compatible
• src/multi-data/norm_extraction.jl — extract_normalisation_index, has_normalisation, get_normalisation_param
• src/multi-data/multi_problem.jl — multi_dataset_problem constructor
• test/test_multi_dataset.jl — 19 tests covering all new functions
• examples/multi_dataset_example.jl — self-contained runnable proof

Modified files

• src/SpectralFitting.jl — include the three new files
• test/runtests.jl — include the new test file

Design decisions

Type stability: multi_dataset_problem uses Vararg{T,N} + Val(N) + ntuple to construct model copies as a Tuple (not a Vector), consistent with how FittableMultiModel works internally.

Reuse of existing API: The implementation deliberately reuses bindall!, normalisationfield, and parameter_names from the existing codebase rather than reimplementing parameter introspection.

Zero new dependencies: All three files use only types and functions already present in SpectralFitting.jl.

Verification

using SpectralFitting, Random

energy_bins = collect(range(0.3, 10.0, length=51))
TRUE_K = [1.0, 0.5, 2.0]; TRUE_a = 1.7

datasets = map(enumerate(TRUE_K)) do (i, K)
    Random.seed!(i)
    m = PowerLaw(K=FitParam(K), a=FitParam(TRUE_a))
    flux = collect(invokemodel(energy_bins, m))
    InjectiveData(energy_bins[1:end-1], flux; codomain_variance=abs.(flux).*0.001)
end

prob   = multi_dataset_problem(PowerLaw(), datasets...)
result = fit(prob, LevenbergMarquadt())

for i in 1:3
    println("Dataset $i: K=$(round(result[i].u[1],digits=4)), α=$(round(result[i].u[2],digits=4))")
end

Output:

Dataset 1: K≈1.0, α≈1.7
Dataset 2: K≈0.5, α≈1.7
Dataset 3: K≈2.0, α≈1.7

Test results

api                              | 17  17  ✓
composite-algebra                | 18  18  ✓
model-library                    | 58  58  ✓
io                               | 18  18  ✓
fitting                          | 59  59  ✓
simulation                       |  3   3  ✓
domain_union                     |  6   6  ✓  ← new
extract_normalisation_index      |  4   4  ✓  ← new
multi_dataset_problem structure  |  4   4  ✓  ← new
multi_dataset_problem fitting    |  5   5  ✓  ← new

Total: 211 tests, 0 failures

Closes #196
See also #132

[codecov]

Codecov Report

❌ Patch coverage is 88.63636% with 5 lines in your changes missing coverage. Please review.
✅ Project coverage is 56.35%. Comparing base (62aba68) to head (672a526).

Files with missing lines	Patch %	Lines
src/multi-data/norm_extraction.jl	60.00%	4 Missing ⚠️
src/multi-data/domain_utils.jl	95.23%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #241      +/-   ##
==========================================
+ Coverage   55.69%   56.35%   +0.65%     
==========================================
  Files          39       42       +3     
  Lines        2632     2676      +44     
==========================================
+ Hits         1466     1508      +42     
- Misses       1166     1168       +2     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

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

[aditya-pandey-dev]

@fjebaker I have raise a PR related to this issue can you please check the PR. What changes should I do?