styling

Author: marcburri

.github/CONTRIBUTING.md

@@ -5,23 +5,23 @@ For a detailed discussion on contributing to this and other tidyverse packages,
 
 ## Fixing typos
 
-You can fix typos, spelling mistakes, or grammatical errors in the documentation directly using the GitHub web interface, as long as the changes are made in the _source_ file. 
-This generally means you'll need to edit [roxygen2 comments](https://roxygen2.r-lib.org/articles/roxygen2.html) in an `.R`, not a `.Rd` file. 
+You can fix typos, spelling mistakes, or grammatical errors in the documentation directly using the GitHub web interface, as long as the changes are made in the _source_ file.
+This generally means you'll need to edit [roxygen2 comments](https://roxygen2.r-lib.org/articles/roxygen2.html) in an `.R`, not a `.Rd` file.
 You can find the `.R` file that generates the `.Rd` by reading the comment in the first line.
 
 ## Bigger changes
 
-If you want to make a bigger change, it's a good idea to first file an issue and make sure someone from the team agrees that it’s needed. 
-If you’ve found a bug, please file an issue that illustrates the bug with a minimal 
+If you want to make a bigger change, it's a good idea to first file an issue and make sure someone from the team agrees that it’s needed.
+If you’ve found a bug, please file an issue that illustrates the bug with a minimal
 [reprex](https://www.tidyverse.org/help/#reprex) (this will also help you write a unit test, if needed).
 See our guide on [how to create a great issue](https://code-review.tidyverse.org/issues/) for more advice.
 
 ### Pull request process
 
 *   Fork the package and clone onto your computer. If you haven't done this before, we recommend using `usethis::create_from_github("p-wegmueller/reviser", fork = TRUE)`.
 
-*   Install all development dependencies with `devtools::install_dev_deps()`, and then make sure the package passes R CMD check by running `devtools::check()`. 
-    If R CMD check doesn't pass cleanly, it's a good idea to ask for help before continuing. 
+*   Install all development dependencies with `devtools::install_dev_deps()`, and then make sure the package passes R CMD check by running `devtools::check()`.
+    If R CMD check doesn't pass cleanly, it's a good idea to ask for help before continuing.
 *   Create a Git branch for your pull request (PR). We recommend using `usethis::pr_init("brief-description-of-change")`.
 
 *   Make your changes, commit to git, and then create a PR by running `usethis::pr_push()`, and following the prompts in your browser.
@@ -32,13 +32,13 @@ See our guide on [how to create a great issue](https://code-review.tidyverse.org
 
 ### Code style
 
-*   New code should follow the tidyverse [style guide](https://style.tidyverse.org). 
-    You can use the [styler](https://CRAN.R-project.org/package=styler) package to apply these styles, but please don't restyle code that has nothing to do with your PR.  
+*   New code should follow the tidyverse [style guide](https://style.tidyverse.org).
+    You can use the [styler](https://CRAN.R-project.org/package=styler) package to apply these styles, but please don't restyle code that has nothing to do with your PR.
 
-*  We use [roxygen2](https://cran.r-project.org/package=roxygen2), with [Markdown syntax](https://cran.r-project.org/web/packages/roxygen2/vignettes/rd-formatting.html), for documentation.  
+*  We use [roxygen2](https://cran.r-project.org/package=roxygen2), with [Markdown syntax](https://cran.r-project.org/web/packages/roxygen2/vignettes/rd-formatting.html), for documentation.
 
-*  We use [testthat](https://cran.r-project.org/package=testthat) for unit tests. 
-   Contributions with test cases included are easier to accept.  
+*  We use [testthat](https://cran.r-project.org/package=testthat) for unit tests.
+   Contributions with test cases included are easier to accept.
 
 ## Code of Conduct
 

.github/workflows/pkgdown.yaml

@@ -48,4 +48,4 @@ jobs:
           clean: false
           branch: gh-pages
           folder: docs
-          
+

CODE_OF_CONDUCT.md

@@ -59,7 +59,7 @@ representative at an online or offline event.
 ## Enforcement
 
 Instances of abusive, harassing, or otherwise unacceptable behavior may be
-reported to the community leaders responsible for enforcement at marc.burri91@gmail.com. 
+reported to the community leaders responsible for enforcement at marc.burri91@gmail.com.
 All complaints will be reviewed and investigated promptly and fairly.
 
 All community leaders are obligated to respect the privacy and security of the

DESCRIPTION

@@ -5,7 +5,7 @@ Version: 0.1.0.9000
 Authors@R: c(
     person(given = "Marc", family = "Burri", email = "marc.burri91@gmail.com", role = c("aut", "cre", "cph"),
            comment = c(ORCID = "0000-0001-8974-9090")),
-    person(given = "Philipp", family = "Wegmueller", 
+    person(given = "Philipp", family = "Wegmueller",
         email = "philipp.wemueller@seco.admin.ch", role = c("aut", "cph"))
   )
 Description: Provides tools to analyze revision properties in real-time time series data.
@@ -14,7 +14,7 @@ Maintainer: Marc Burri <marc.burri91@gmail.com>
 License: MIT + file LICENSE
 Encoding: UTF-8
 LazyData: true
-Depends: 
+Depends:
     R (>= 2.10)
 RoxygenNote: 7.3.3
 Roxygen: list (markdown = TRUE, roclets = c ("namespace", "rd", "srr::srr_stats_roclet"))
@@ -34,7 +34,7 @@ Imports:
     lubridate,
     KFAS,
     numDeriv
-Suggests: 
+Suggests:
     tsbox,
     testthat (>= 3.0.0),
     knitr,

R/data-vintages.R

@@ -31,7 +31,7 @@
 #' * JP:
 #'   * Japan
 #'   * Sources: Cabinet Office (Japan), OECD
-#'   
+#'
 #'
 #' @srrstats {G1.0} academic literature
 #'

R/graphs.R

@@ -320,7 +320,8 @@ plot_vintages <- function(
 #' @param ... Additional arguments passed to plot_vintages.
 #'
 #' @return A ggplot2 object.
-#' @srrstats {TS5.0} Implements default plot methods for implemented class system
+#' @srrstats {TS5.0} Implements default plot methods for implemented
+#'   class system
 #' @srrstats {TS4.2} Explicitly documents the type and class of return values
 #' @method plot tbl_pubdate
 #' @examples
@@ -338,7 +339,8 @@ plot.tbl_pubdate <- function(x, ...) {
 #' @param ... Additional arguments passed to plot_vintages.
 #'
 #' @return A ggplot2 object.
-#' @srrstats {TS5.0} Implements default plot methods for implemented class system
+#' @srrstats {TS5.0} Implements default plot methods for implemented
+#'   class system
 #' @srrstats {TS4.2} Explicitly documents the type and class of return values
 #' @method plot tbl_release
 #' @examples
@@ -359,9 +361,12 @@ plot.tbl_release <- function(x, ...) {
 #' @param ... Additional arguments passed to theme_reviser.
 #' @srrstats {G1.4a} Internal function documented with @noRd tag
 #' @srrstats {TS5.0} Implements plot methods for class system
-#' @srrstats {TS5.7} Includes model (input) values in plot with forecast (output) values (in-sample vs out-of-sample)
-#' @srrstats {TS5.8} Provides clear visual distinction between model and forecast values (different colors/samples)
-#' @srrstats {TS5.6} Indicates distributional limits (confidence intervals) on plot by default
+#' @srrstats {TS5.7} Includes model (input) values in plot with forecast
+#'   (output) values (in-sample vs out-of-sample)
+#' @srrstats {TS5.8} Provides clear visual distinction between model and
+#'   forecast values (different colors/samples)
+#' @srrstats {TS5.6} Indicates distributional limits (confidence
+#'   intervals) on plot by default
 #' @return ggplot object
 #'
 #' @keywords internal

R/jvn.R

@@ -46,15 +46,20 @@
 #'   \item{bic}{Bayesian Information Criterion}
 #'   \item{data}{Input data}
 #' }
-#' 
+#'
 #' @srrstats {G1.0} Primary reference: Jacobs & Van Norden (2011)
-#' @srrstats {G1.1} First implementation in R of the Jacobs-Van Norden state-space model
-#' @srrstats {G1.3} Statistical terminology clearly defined (news, noise, spillovers)
-#' @srrstats {G2.0} Input assertions on lengths (e, h, ar_order must be single values)
-#' @srrstats {G2.0a} Documents expectations on input lengths in parameter descriptions
+#' @srrstats {G1.1} First implementation in R of the Jacobs-Van Norden
+#'   state-space model
+#' @srrstats {G1.3} Statistical terminology clearly defined (news, noise,
+#'   spillovers)
+#' @srrstats {G2.0} Input assertions on lengths (e, h, ar_order must be
+#'   single values)
+#' @srrstats {G2.0a} Documents expectations on input lengths in parameter
+#'   descriptions
 #' @srrstats {G2.1} Input type assertions (checks for list, numeric types)
 #' @srrstats {G2.1a} Documents data type expectations for all inputs
-#' @srrstats {G2.2} Restricts multivariate input to univariate parameters (e, h, ar_order)
+#' @srrstats {G2.2} Restricts multivariate input to univariate parameters
+#'   (e, h, ar_order)
 #' @srrstats {G2.3a} Uses match.arg() equivalent for method parameter validation
 #' @srrstats {G2.3b} Documents case-sensitivity requirements
 #' @srrstats {G2.4} Type conversion mechanisms (round() for ar_order)
@@ -71,18 +76,21 @@
 #' @srrstats {TS1.0} Uses explicit time series class systems
 #' @srrstats {TS1.1} Documents types/classes of input data
 #' @srrstats {TS1.2} Validation routines for input classes (vintages_check)
-#' @srrstats {TS1.3} Pre-processing to validate and transform input (vintages_wide)
+#' @srrstats {TS1.3} Pre-processing to validate and transform input
+#'   (vintages_wide)
 #' @srrstats {TS1.4} Maintains time/date components of input data
 #' @srrstats {TS1.5} Ensures strict ordering of time index
 #' @srrstats {TS1.6} Catches ordering violations in pre-processing
-#' @srrstats {TS1.8} Explicit about monthly time intervals (frequency calculation)
+#' @srrstats {TS1.8} Explicit about monthly time intervals (frequency
+#'   calculation)
 #' @srrstats {TS2.0} Handles explicit vs implicit missing values
 #' @srrstats {TS2.1} Options for handling missing data
 #' @srrstats {TS4.0b} Returns unique class-defined format (jvn_model)
 #' @srrstats {TS4.2} Explicitly documents return value types/classes
 #' @srrstats {TS4.3} Return values include time scales
 #' @srrstats {TS4.6b} Forecasting returns first- and second-order moments
-#' @srrstats {TS4.6c} Error indication for forecast estimates (confidence intervals)
+#' @srrstats {TS4.6c} Error indication for forecast estimates (confidence
+#'   intervals)
 #' @srrstats {TS4.7c} Distinguishes model vs forecast values (sample column)
 #'
 #' @references Jacobs, Jan P.A.M. and Van Norden, Simon, "Modeling Data
@@ -167,6 +175,11 @@ jvn_nowcast <- function(
   }
 
   # Check input e
+  if (!is.numeric(e) || length(e) != 1 || is.na(e) || e %% 1 != 0) {
+    rlang::abort("'e' must be a single whole number greater than 0.")
+  }
+  e <- as.integer(e)
+
   if (e == 0) {
     rlang::abort("The initial release is already efficient, 'e' is equal to 0!")
   }
@@ -209,8 +222,22 @@ jvn_nowcast <- function(
     if (is.list(df) && !is.data.frame(df)) df <- df[[1]]
   }
 
+  required_release_cols <- paste0("release_", 0:(e - 1))
+  missing_release_cols <- setdiff(required_release_cols, colnames(df))
+  if (length(missing_release_cols) > 0) {
+    rlang::abort(paste0(
+      "'df' must contain release columns ",
+      paste(required_release_cols, collapse = ", "),
+      " for e = ",
+      e,
+      ". Missing: ",
+      paste(missing_release_cols, collapse = ", ")
+    ))
+  }
+
   # Arrange data
-  df_intern <- df[, 1:(e + 1)]
+  df_intern <- df %>%
+    dplyr::select("time", dplyr::all_of(required_release_cols))
 
   y_mat <- as.matrix(dplyr::select(df_intern, -"time"))
   rownames(y_mat) <- as.character(df$time)
@@ -698,7 +725,7 @@ jvn_nowcast <- function(
     frequency <- unique((round(as.numeric(diff(df$time)) / 30)))
     if (length(frequency) > 1) {
       rlang::abort(
-        "The time series seems not to be regular, 
+        "The time series seems not to be regular,
         please provide a regular time series!"
       )
     }
@@ -821,7 +848,7 @@ jvn_nowcast <- function(
 #'
 #' Constructs the state-space matrices Z, T, R, H, Q according to the
 #' Jacobs & Van Norden (2011) specification.
-#' 
+#'
 #' @srrstats {G1.4a} Internal function documented with @noRd tag
 #' @srrstats {G2.0} Input assertions on parameter dimensions
 #' @srrstats {G2.1} Type checking for parameters
@@ -1149,6 +1176,49 @@ jvn_negloglik <- function(params, model_struct, data, transform_se = TRUE) {
   )
 }
 
+#' Create stable default starting values for the JVN model
+#' @keywords internal
+#' @noRd
+jvn_default_params <- function(model_struct, transform_se = TRUE) {
+  info <- model_struct$param_info
+  params <- numeric(model_struct$n_params)
+
+  if (!is.null(info$ar_coef_idx)) {
+    ar_default <- 0.4 / seq_along(info$ar_coef_idx)
+    params[info$ar_coef_idx] <- ar_default
+  }
+
+  if (!is.null(info$sigma_e_idx)) {
+    params[info$sigma_e_idx] <- 0.2
+  }
+  if (!is.null(info$sigma_nu_idx)) {
+    params[info$sigma_nu_idx] <- 0.15
+  }
+  if (!is.null(info$sigma_zeta_idx)) {
+    params[info$sigma_zeta_idx] <- 0.1
+  }
+  if (!is.null(info$spill_nu_idx)) {
+    params[info$spill_nu_idx] <- 0.3
+  }
+  if (!is.null(info$spill_zeta_idx)) {
+    params[info$spill_zeta_idx] <- 0.2
+  }
+
+  if (transform_se) {
+    if (!is.null(info$sigma_e_idx)) {
+      params[info$sigma_e_idx] <- log(params[info$sigma_e_idx])
+    }
+    if (!is.null(info$sigma_nu_idx)) {
+      params[info$sigma_nu_idx] <- log(params[info$sigma_nu_idx])
+    }
+    if (!is.null(info$sigma_zeta_idx)) {
+      params[info$sigma_zeta_idx] <- log(params[info$sigma_zeta_idx])
+    }
+  }
+
+  params
+}
+
 #' Initialize Parameters with Data-Driven Starting Values
 #' @keywords internal
 #' @noRd
@@ -1163,7 +1233,7 @@ jvn_init_params <- function(model_struct, data, transform_se = TRUE) {
 
   if (nrow(data_clean) < ar_order + 10) {
     warning("Insufficient data for smart initialization. Using default values.")
-    return(jvn_init_params(model_struct, transform_se))
+    return(jvn_default_params(model_struct, transform_se))
   }
 
   # ===== 1. AR COEFFICIENTS AND SHOCK from final vintage =====
@@ -1301,7 +1371,7 @@ jvn_init_params <- function(model_struct, data, transform_se = TRUE) {
   # ===== 6. VALIDATION =====
   if (any(!is.finite(params))) {
     warning("Some starting values are non-finite. Falling back to defaults.")
-    return(jvn_init_params(model_struct, transform_se))
+    return(jvn_default_params(model_struct, transform_se))
   }
 
   # Ensure AR stationarity
@@ -1483,7 +1553,7 @@ print.jvn_model <- function(x, ...) {
 #' @param state String. The name of the state to visualize.
 #' @param type String. Type of estimate to plot: "filtered" or "smoothed".
 #' @param ... Additional arguments passed to theme_reviser.
-#' 
+#'
 #' @srrstats {TS5.0} Implements default plot methods for class system
 #' @srrstats {TS5.1} Time axis labeling (delegates to base method)
 #' @srrstats {TS5.2} Time on horizontal axis (delegates to base method)

R/kk.R

@@ -101,7 +101,8 @@
 #' @srrstats {TS4.6} Time Series Software which implements or otherwise
 #' enables forecasting should return either:
 #' @srrstats {TS4.6b} filtered/forecasted point estimates
-#' @srrstats {TS4.6c} Error indication for forecast estimates (confidence intervals)
+#' @srrstats {TS4.6c} Error indication for forecast estimates (confidence
+#'   intervals)
 #' @srrstats {TS4.7c} Distinguishes model vs forecast values (sample column)
 #' @srrstats {TS4.7} forecast values and models separately returned
 #' @srrstats {TS4.7a} only forecast values returned
@@ -160,6 +161,11 @@ kk_nowcast <- function(
   }
 
   # Check input e
+  if (!is.numeric(e) || length(e) != 1 || is.na(e) || e %% 1 != 0) {
+    rlang::abort("'e' must be a single whole number greater than 0.")
+  }
+  e <- as.integer(e)
+
   if (e == 0) {
     rlang::abort("The initial release is already efficient, 'e' is equal to 0!")
   }
@@ -269,6 +275,19 @@ kk_nowcast <- function(
     if (is.list(df) && !is.data.frame(df)) df <- df[[1]]
   }
 
+  required_release_cols <- paste0("release_", 0:e)
+  missing_release_cols <- setdiff(required_release_cols, colnames(df))
+  if (length(missing_release_cols) > 0) {
+    rlang::abort(paste0(
+      "'df' must contain release columns ",
+      paste(required_release_cols, collapse = ", "),
+      " for e = ",
+      e,
+      ". Missing: ",
+      paste(missing_release_cols, collapse = ", ")
+    ))
+  }
+
   # Define state and observable variable names
   z_names <- c(paste0("release_", e, "_lag_", (e):0))
   y_names <- c(paste0("release_", e:0, "_lag_", e:0))
@@ -873,7 +892,7 @@ kk_nowcast <- function(
     frequency <- unique((round(as.numeric(diff(df$time)) / 30)))
     if (length(frequency) > 1) {
       rlang::abort(
-        "The time series seems not to be regular, 
+        "The time series seems not to be regular,
         please provide a regular time series!"
       )
     }