Running Verification¶

Learn how to run the full verification suite with custom parameters and tolerances.

Default Verification¶

The default verification checks all core results from the monograph:

from src.verification.verify_all import verify_all

results = verify_all()
for name, result in results.items():
    print(f"{name}: {'PASS' if result.passed else 'FAIL'} (dev: {result.deviation:.4f}%)")

Custom Tolerance¶

Adjust the tolerance for deviation checks:

python run.py --mode verify --tolerance 0.001 --non-interactive

Or programmatically:

from src.verification.verify_all import verify_all

results = verify_all(tolerance=0.001)  # 0.1% tolerance

Enhanced Verification (v2.0)¶

The enhanced verification suite covers the v2.0 extensions:

4D spin manifold conformal invariance
Kähler surface Dolbeault correspondence
K3 surface hyperkähler verification
Tyukovsky equations (zero free parameters)
Einstein GR QNM corrections
Criticism response (b₂ = 22 uniqueness)

from src.verification.verify_enhanced import verify_enhanced

results = verify_enhanced()

Verification Results Reference¶

Constant	Computed	Observed	Deviation	Status
\(\Delta_{bC}\)	3.438710	3.443	0.125%	:verified:{ .verified }
\(\Delta_{Ch}\) (base)	3.437883	3.443	0.149%	:verified:{ .verified }
\(\Delta_{Ch}\) (full)	3.447040	3.443	0.117%	:verified:{ .verified }
\(b_{Ch}\)	0.376510	0.377	0.130%	:verified:{ .verified }

Output Formats¶

Verification results can be exported in all 7 report formats. The JSON format is particularly useful for automated processing:

from src.reporting.report_writer import ReportWriter

writer = ReportWriter(output_dir="output/reports")
writer.write(results, formats=["json", "csv", "pdf"])