data(smartphone): compute display ppi from resolution + size (51,557 records)

[Seungpyo1007]

Summary

Computes display pixel density (ppi) for smartphones from fields already in the record — pure geometry, no new/external data.

• display.ppi = round(sqrt(width² + height²) / size_inch) for 51,557 smartphones that have a resolution and screen size but no ppi.
• Coverage ppi ~26% → ~81%.
• Consistency-validated against the 24k records that already carry a source ppi: 89% exact match, 10% off-by-1 (rounding), only 0.4% differ by more — and those are Apple's marketed/rounded figures (e.g. iPhone "326") vs the true geometric density.
• Guards: requires w,h ≥ 32 and size > 0; rejects results outside [20, 2000] ppi.

Same deterministic method PR #107 used for monitors & laptops, now applied to the smartphone catalogue. All verified: false. Validation passes. Dump regenerated — 51,557 changed smartphone detail pages.

Closes #1

[Seungpyo1007]

🤖 TechEngine · data validation

category      changed   verified
smartphone     51,557     false

• python -m app.validate → ✅ passed (0 errors), 108,575 records / 10 categories.
• display.ppi coverage 24,255 → 75,812 (~26% → ~81% of smartphones).
• Cross-checked vs existing source ppi: 89% exact, 10% ±1, 0.4% >1 (Apple marketed figures).
• Dump regenerated: 51,557 smartphone detail pages (list summaries unaffected).

[Seungpyo1007]

🤖 TechEngine · verify (Tier 0)

Derivation, not import — the density is computed from resolution and size_inch that already live in each record:

• Formula round(sqrt(w²+h²)/size_inch) is the physical definition of pixel density; validated to reproduce the source's own ppi within ±1 for 99.6% of records that have both.
• Bounds guard rejects implausible inputs (tiny resolutions, non-positive size) and outputs outside [20, 2000].
• The ≤0.4% larger deltas are where the manufacturer published a rounded/marketed ppi; for records lacking ppi entirely, the geometric value is the honest measure.

verified unchanged (false). Safe to rebase-merge.