spb/metrika Public

- sysuse command with bundled samples (sales: 200-firm x 5-year panel;
  mtcars) shipped as ZQEngine resources, generated by
  scripts/make_samples.R
- metrika-cli docs: static documentation site rendered from the shared
  command registry (docs/index.html, 38 commands, light/dark); CLI
  restructured with a default 'run' subcommand
- Metal point-sprite renderer (CLAUDE.md §7 pane 4): packed 12-byte
  vertices, unified-memory buffer written once, scroll-pan /
  pinch-zoom / double-click-reset touching only a 24-byte uniform;
  takes over scatter plots past 100k points (METRIKA_METAL_THRESHOLD
  override for tests); verified rendering 2,000,000 points in-app
- METRIKA_AUTORUN debug hook for headless app driving
- MARKETING_VERSION 1.0.0
- 116 kit tests + 4 UI tests green

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

- ZQGradientBoosting: exact-greedy trees cloning xgboost's algorithm —
  gain 1/2[GL2/(HL+l) + GR2/(HR+l) - G2/(H+l)] - gamma, leaf -G/(H+l),
  midpoint splits between consecutive distinct values, missing rows
  default left, pre-sorted feature indices; squared loss, deterministic
  (no subsampling)
- engine: 'boost y x…, rounds(#) [eta() maxdepth() lambda()]' reporting
  training R2/RMSE with an in-sample caveat; model stored in the
  estimation state (Kind.boost) so predict routes through the trees
  (missing features follow the default direction); margins and GLM
  statistics refused after boost
- validation: per-observation prediction parity with R xgboost 3.2
  (exact method, base_score = mean) at 1e-4 (xgboost is float32
  internally); a stump finds the exact midpoint split with lambda 0;
  training loss decreases monotonically in rounds
- manual entry + coverage test
- 115 tests green (swift test and xcodebuild with GPU suites)

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

- PhiloxStream in ZQGPU: sequential variate stream over a dedicated
  counter block (bit 62 + per-stream 2^48 words, disjoint from bootstrap
  and permutation streams); Box-Muller normals, Marsaglia-Tsang gammas
  (moments verified against theory at 200k draws)
- ZQBayesianRegression: semi-conjugate Gibbs with Stata bayes default
  priors (coefficients N(0, 10000), variance InvGamma(.01, .01)); exact
  full conditionals via dense Cholesky; posterior mean/sd and
  equal-tailed 95% credible intervals; ZQStats now depends on ZQGPU for
  the shared RNG
- engine: 'bayes [, mcmcsize() burnin() seed() normalprior()]: reg …'
  with reproducible chains; manual entry included
- validation: with diffuse priors the posterior reproduces OLS (mean
  within 5% of a posterior SD, sd ratio in [0.9, 1.15], CrI brackets the
  estimate, sigma recovers the DGP); tight priors shrink toward zero;
  chains bit-reproducible per seed
- 111 tests green (swift test and xcodebuild with GPU suites)

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

- ZQElasticNet: cyclic coordinate descent with residual updates,
  glmnet-convention objective ((1/2n)RSS + lambda(alpha*l1 +
  (1-alpha)/2*l2)), internal predictor standardization (1/n variance),
  unpenalized intercept, coefficients reported on the original scale;
  lambdaMax helper
- matched glmnet's gaussian y-standardization quirk deliberately: the L1
  penalty is invariant to it but the effective ridge penalty scales by
  1/sd(y) — without this, alpha<1 fits diverge from glmnet by ~10%
- engine: 'elasticnet y x…, lambda(#) [alpha(#)]' and 'lasso' (alpha
  fixed at 1); missing lambda() errors with the data's lambda_max as a
  hint; predict works afterwards, margins refuses (no VCE)
- fixtures: glmnet 5.0 at thresh 1e-15 over deliberately correlated
  regressors; coefficients match at 1e-6 (documented tolerance for
  penalized iterative solvers) and the selection pattern (which
  coefficients are exactly zero) matches exactly
- ZQCoefficient gains a public initializer
- 104 tests green (swift test and xcodebuild with GPU suites)

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

- ZQOLSResult/ZQGLMResult expose the full covariance matrix (column-major
  k x k aligned with coefficients); IV passes it through
- EstimationState carries vce, inference df, and the estimation-sample
  design (GLMs only — AMEs need it)
- margins, dydx(varlist): OLS/IV effects are the coefficients with their
  SEs; GLM average marginal effects with analytic delta gradients
  (logit p(1-p)(1-2p), probit -xb*phi, poisson exp) over the estimation
  sample; t or z inference per model kind
- factor/interaction dydx rejected with a clear message (discrete-change
  margins later); continuous terms of factor models work
- R fixtures mirror the exact formulas at the converged coefficients;
  logit/poisson AME and delta SE match at 1e-10
- 98 tests green (swift test and xcodebuild with GPU suites)

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

- ZQResampling.permutationIndices: permutation as argsort of 64-bit
  Philox keys — order-free and counter-addressable so a future GPU
  argsort path reproduces it exactly; dedicated stream offset (bit 63,
  applied after word expansion) keeps permutations disjoint from
  bootstrap draws under the same seed
- engine: 'permute, reps(#) [seed(#)]: reg …' — permutes the response
  within the estimation sample, parallel chunked refits, empirical
  two-sided p per coefficient (c, reps, p, SE(p)); degenerate _cons row
  omitted
- planner: GPU heuristic restricted to bootstrap until the permute
  argsort path lands
- graphics: standalone histogram/scatter/kdensity verbs; single-variable
  histogram (Sturges default, bins() option) and Epanechnikov kdensity
  with Silverman bandwidth (density integrates to 1 in tests)
- 81 tests green (swift test and xcodebuild with GPU suites)

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

- parser: (endog = instruments) varlist groups -> ZQIVSpec; digit-led
  sub-commands (2sls) reassembled from number+identifier tokens via
  column adjacency
- ZQFixedEffects: within estimator with Stata conventions (add-back
  means, reported _cons, df = N-K-G), within R-squared, panel-clustered
  VCE with G/(G-1) and t on G-1 df; v0.2 restriction: cluster variable
  must equal the panel variable
- ZQIV: 2SLS via thin-Q projection of the instrument matrix (Z'Z never
  formed), residuals from original regressors, Stata 'small' inference,
  classical/HC1/cluster VCE built on projected regressors
- engine: xtreg (requires xtset + fe), ivregress 2sls with dedicated
  listwise deletion across depvar/exog/endog/instruments; shared
  coefficient-table renderer extracted
- fixtures: z1/z2 instrument columns (drawn after existing draws, earlier
  golden values bit-identical), manual within/2SLS algebra in R
- 75 tests green (swift test and xcodebuild with GPU suites)

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

- MetrikaKit SPM package: ZQParser, ZQPlanner, ZQEngine, ZQData, ZQStats,
  ZQGPU, ZQGraphics, ZQPlugins (Swift 6, strict concurrency)
- ZQL parser: lexer, Pratt expressions, factor variables, prefix commands,
  column-cited errors with Levenshtein verb suggestions
- ZQData: DuckDB-backed load/save (parquet, csv, json, arrow)
- ZQStats: OLS via LAPACK QR, HC0-HC3 and cluster-robust SE, summarize,
  t/F distributions accurate in the far tails
- ZQGPU: Philox4x32-10 reference RNG, counter-addressable bootstrap
- ZQEngine: session actor with use/save/gen/replace/drop/keep/summarize/
  regress/count/list/graph/bootstrap/set seed/xtset/log
- SwiftUI app (xcodegen): console with history, variables sidebar,
  Swift Charts plots; sandboxed + hardened runtime entitlements
- Tests: 44 green (parser golden, R fixtures at 1e-10, Philox KAT,
  end-to-end engine); Tests/Fixtures/generate.R; Tests/Bench harness
- scripts: check_headers.sh + pre-commit hook, make_icns.sh, release.sh

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>