added column constraints for tidy, glance, augment

build.clj

@@ -1,7 +1,12 @@
 (ns build
   (:refer-clojure :exclude [test])
-  (:require [clojure.tools.build.api :as b] ; for b/git-count-revs
-            [org.corfield.build :as bb]))
+  (:require
+   [camel-snake-kebab.core :as csk]
+   [clj-yaml.core :as yaml]
+   [clojure.java.io :as io]
+   [clojure.pprint :as pp]
+   [clojure.tools.build.api :as b] ; for b/git-count-revs
+   [org.corfield.build :as bb]))
 
 (def lib 'scicloj/metamorph.ml)
 ; alternatively, use MAJOR.MINOR.COMMITS:
@@ -66,3 +71,25 @@
   (-> opts
       (assoc :lib lib :version version)
       (bb/deploy)))
+
+(defn build-glance-columns [ops]
+  (with-open [w (io/writer "resources/columms-glance.edn")]
+    (-> (slurp "https://raw.githubusercontent.com/alexpghayes/modeltests/main/data-raw/columns_glance.yaml")
+        (yaml/parse-string
+         :key-fn #(-> % :key  csk/->kebab-case-keyword))
+        (pp/pprint w))))
+
+
+(defn build-tidy-columns [opts]
+  (with-open [w (io/writer "resources/columms-tidy.edn")]
+    (-> (slurp "https://raw.githubusercontent.com/alexpghayes/modeltests/main/data-raw/columns_tidy.yaml")
+        (yaml/parse-string
+         :key-fn #(-> % :key  csk/->kebab-case-keyword))
+        (pp/pprint w))))
+
+(defn build-augment-columns [ops]
+  (with-open [w (io/writer "resources/columms-augment.edn")]
+    (-> (slurp "https://raw.githubusercontent.com/alexpghayes/modeltests/main/data-raw/columns_augment.yaml")
+        (yaml/parse-string
+         :key-fn #(-> % :key  csk/->kebab-case-keyword))
+        (pp/pprint w))))

deps.edn

@@ -14,7 +14,9 @@
 
   :build {:deps {io.github.clojure/tools.build {:mvn/version "0.9.6"}
                  slipset/deps-deploy {:mvn/version "0.2.1"}
-                 io.github.seancorfield/build-clj {:git/tag "v0.9.2" :git/sha "9c9f078"}}
+                 io.github.seancorfield/build-clj {:git/tag "v0.9.2" :git/sha "9c9f078"}
+                 clj-commons/clj-yaml {:mvn/version "1.0.27"}
+                 camel-snake-kebab/camel-snake-kebab {:mvn/version "0.4.3"}}
           :ns-default build}
 
 
@@ -29,7 +31,7 @@
   {:jvm-opts ["-Djdk.attach.allowAttachSelf"]
    :extra-paths ["test"]
    :extra-deps {
-                lambdaisland/kaocha {:mvn/version "0.15.957"}
+                lambdaisland/kaocha {:mvn/version "1.88.1376"}
                 scicloj/scicloj.ml.smile {:mvn/version "7.3"}
                 datacraft-sciences/confuse {:mvn/version "0.1.1"}
                 ch.qos.logback/logback-classic {:mvn/version "1.5.6"}}}

resources/columms-augment.edn

@@ -0,0 +1,50 @@
+{:.class "Predicted class.",
+ :.cluster "Cluster assignment.",
+ :.cochran.qe.loo
+ "Leave-one-out residual heterogeneity test statistics.",
+ :.col.prop "Column proportion (2 dimensional table only).",
+ :.conf.high "Upper bound on confidence interval for fitted values.",
+ :.conf.low "Lower bound on confidence interval for fitted values.",
+ :.cooksd "Cooks distance.",
+ :.cov.ratio "The covariance ratio.",
+ :.cred.high "Upper bound on credible interval for fitted values.",
+ :.cred.low "Lower bound on credible interval for fitted values.",
+ :.dffits
+ "Estimated change in standard deviations for the predicted effect after excluding the study",
+ :.expected "Expected count under the null hypothesis.",
+ :.fitted "Fitted or predicted value.",
+ :.fitted-j-0
+ "Population-level fitted values for the j-th longitudinal process.",
+ :.fitted-j-1
+ "Individual-level fitted values for the j-th longitudinal process.",
+ :.hat "Diagonal of the hat matrix.",
+ :.lower "Lower bound on interval for fitted values.",
+ :.moderator
+ "In meta-analysis, the moderators used to calculate the predicted values.",
+ :.moderator.level
+ "In meta-analysis, the level of the moderators used to calculate the predicted values.",
+ :.observed "Observed count.",
+ :.probability "Class probability of modal class.",
+ :.prop "Proportion of the total.",
+ :.remainder
+ "The remainder, or random, component of the decomposition.",
+ :.resid "The difference between observed and fitted values.",
+ :.resid-j-0
+ "Population-level residuals for the j-th longitudinal process.",
+ :.resid-j-1
+ "Individual-level residuals for the j-th longitudinal process.",
+ :.row.prop "Row proportion (2 dimensions table only).",
+ :.rownames "Rownames from the original data, if present.",
+ :.se.fit "Standard errors of fitted values.",
+ :.seasadj "The seasonally adjusted (or *deseasonalised*) series.",
+ :.seasonal "The seasonal component of the decomposition.",
+ :.sigma
+ "Estimated residual standard deviation when corresponding observation is dropped from model.",
+ :.std.resid "Standardised residuals.",
+ :.tau "Quantile.",
+ :.tau.squared.loo "leave-one-out tau-squared estimates",
+ :.trend "The trend component of the decomposition.",
+ :.uncertainty
+ "The uncertainty associated with the classification. Equal to one minus the model class probability.",
+ :.upper "Upper bound on interval for fitted values.",
+ :.weight "The final robust weights."}

resources/columms-glance.edn

@@ -0,0 +1,153 @@
+{:adj.r.squared
+ "Adjusted R squared statistic, which is like the R squared statistic except taking degrees of freedom into account.",
+ :agfi "Adjusted goodness of fit.",
+ :aic "Akaike's Information Criterion for the model.",
+ :ai-cc
+ "Small sample corrected Akaike's Information Criterion for the model.",
+ :alpha "Estimated correlation parameter for geepack::geeglm.",
+ :alternative "The alternative hypothesis. Usually character.",
+ :autocorrelation "Autocorrelation.",
+ :avg.silhouette.width "The average silhouette width for the dataset.",
+ :betweenss "The total between-cluster sum of squares.",
+ :bic "Bayesian Information Criterion for the model.",
+ :cfi "Comparative fit index.",
+ :chi.squared
+ "The Pearson Chi-Square goodness of fit statistic for multiway tables.",
+ :chisq "Model chi squared.",
+ :cochran.qe
+ "In meta-analysis, test statistic for the Cochran's Q_e test of residual heterogeneity.",
+ :cochran.qm
+ "In meta-analysis, test statistic for the Cochran's Q_m omnibus test of coefficients.",
+ :conf.high "Upper bound on confidence interval.",
+ :conf.low "Lower bound on confidence interval.",
+ :converged
+ "Logical indicating if the model fitting procedure was succesful and converged.",
+ :convergence "Convergence code.",
+ :crit "Minimized criterion",
+ :cv.crit "Cross-validation score",
+ :den.df "Degrees of freedom of the denominator",
+ :deviance "Deviance of the model.",
+ :df "Degrees of freedom used by the model.",
+ :df.null "Degrees of freedom used by the null model.",
+ :df.residual "Residual degrees of freedom.",
+ :dw.original "Durbin-Watson statistic of original fit.",
+ :dw.transformed "Durbin-Watson statistic of transformed fit.",
+ :edf "The effective degrees of freedom.",
+ :estimator "Estimator used.",
+ :events "Number of events.",
+ :fin-tol "The achieved convergence tolerance.",
+ :function.count "Number of calls to `fn`.",
+ :g "The optimal number of mixture components.",
+ :g.squared "The likelihood ratio/deviance statistic.",
+ :gamma "Estimated scale parameter for geepack::geeglm.",
+ :gradient.count "Number of calls to `gr`.",
+ :h
+ "H statistic for computing confidence interval of major axis slope.",
+ :h.squared "Value of the H-Squared statistic.",
+ :hypvol
+ "If the other model contains a noise component, the value of the hypervolume parameter. Otherwise `NA`.",
+ :i.squared "Value of the I-Squared statistic.",
+ :independence "Whether the model assumed dyadic independence.",
+ :is-conv "Whether the fit successfully converged.",
+ :iter "Iterations of algorithm/fitting procedure completed.",
+ :iterations "The number of iterations performed before convergence.",
+ :k-hkb "Modified HKB estimate of the ridge constant.",
+ :k-lw "Modified L-W estimate of the ridge constant.",
+ :lag.order "Lag order.",
+ :lambda "Choice of lambda corresponding to `spar`.",
+ :lambda.-1se
+ "The value of the penalization parameter lambda that results in the sparsest model while remaining within one standard error of the minimum loss.",
+ :lambda.min
+ "The value of the penalization parameter lambda that achieved minimum loss as estimated by cross validation.",
+ :lambda-gcv "choice of lambda that minimizes GCV.",
+ :log-lik
+ "The log-likelihood of the model. [stats::logLik()] may be a useful reference.",
+ :max.cluster.size "Max number of elements in clusters.",
+ :max.hazard "Maximal estimated hazard.",
+ :max.time "The maximum observed event or censoring time.",
+ :maxit "Number of iterations performed.",
+ :mcmc.burnin "The burn-in period of the MCMC estimation.",
+ :mcmc.interval "The interval used during MCMC estimation.",
+ :mcmc.samplesize "The sample size used during MCMC estimation.",
+ :measure "The measure used in the meta-analysis.",
+ :median "median survival.",
+ :method "Which method was used.",
+ :min.hazard "Minimal estimated hazard.",
+ :min.time "The minimum observed event or censoring time.",
+ :missing-method "Method for eliminating missing data.",
+ :model
+ "A character string denoting the model at which the optimal BIC occurs.",
+ :n "The total number of observations.",
+ :n.clusters "Number of clusters.",
+ :n.factors "The number of fitted factors.",
+ :n.max "Maximum number of subjects at risk.",
+ :n.start "Initial number of subjects at risk.",
+ :nevent "Number of events.",
+ :nexcluded "Number of excluded observations.",
+ :ngroups "Number of groups in model.",
+ :nobs "Number of observations used.",
+ :norig "Number of observation in the original dataset.",
+ :npar "Number of parameters in the model.",
+ :npasses "Total passes over the data across all lambda values.",
+ :null.deviance "Deviance of the null model.",
+ :nulldev "Null deviance.",
+ :num.df "Degrees of freedom of the numerator.",
+ :number.interaction "Number of interactions.",
+ :offtable "Total number of person-years off table.",
+ :p.value "P-value corresponding to the test statistic.",
+ :p.value.cochran.qe
+ "In meta-analysis, p-value for the Cochran's Q_e test of residual heterogeneity.",
+ :p.value.cochran.qm
+ "In meta-analysis, p-value for the Cochran's Q_m omnibus test of coefficients.",
+ :p.value.original "P-value of original Durbin-Watson statistic.",
+ :p.value.-sargan "P-value for Sargan test.",
+ :p.value.transformed
+ "P-value of autocorrelation after transformation.",
+ :p.value.weak.instr "P-value for weak instrument F-test.",
+ :p.value.-wu.-hausman "P-value for Wu-Hausman test.",
+ :parameter
+ "Parameter field in the htest, typically degrees of freedom.",
+ :pen.crit "Penalized criterion.",
+ :power "Power achieved by the analysis.",
+ :power.reached "Whether the desired power was reached.",
+ :pseudo.r.squared
+ "Like the R squared statistic, but for situations when the R squared statistic isn't defined.",
+ :r.squared
+ "R squared statistic, or the percent of variation explained by the model. Also known as the coefficient of determination.",
+ :records "Number of observations",
+ :residual.deviance "The residual deviance of the model",
+ :rho "Spearman's rho autocorrelation",
+ :rho-2
+ "McFadden's rho squared with respect to a market shares (constants-only) model.",
+ :rho-20
+ "McFadden's rho squared with respect to an equal shares (no information) model.",
+ :rmean "Restricted mean (see [survival::print.survfit()]).",
+ :rmean.std.error "Restricted mean standard error.",
+ :rmsea "Root mean square error of approximation.",
+ :rmsea.conf.high "95 percent upper bound on RMSEA.",
+ :rscore "Robust log-rank statistic",
+ :score "Score.",
+ :sigma "Estimated standard error of the residuals.",
+ :sigma-2-j
+ "The square root of the estimated residual variance for the j-th longitudinal process.",
+ :spar "Smoothing parameter.",
+ :srmr "Standardised root mean residual.",
+ :statistic "Test statistic.",
+ :statistic.-sargan "Statistic for Sargan test.",
+ :statistic.weak.instr "Statistic for weak instrument F-test.",
+ :statistic.-wu.-hausman "Statistic for Wu-Hausman test.",
+ :tau "Quantile.",
+ :tau.squared
+ "In meta-analysis, estimated amount of residual heterogeneity.",
+ :tau.squared.se
+ "In meta-analysis, standard error of residual heterogeneity.",
+ :theta "Angle between OLS lines `lm(y ~ x)` and `lm(x ~ y)`.",
+ :timepoints "Number of timepoints.",
+ :tli "Tucker Lewis index.",
+ :tot.withinss "The total within-cluster sum of squares.",
+ :total "Total number of person-years tabulated.",
+ :total.variance
+ "Total cumulative proportion of variance accounted for by all factors.",
+ :totss "The total sum of squares.",
+ :value "Minimized or maximized output value.",
+ :within.r.squared "R squared within fixed-effect groups."}