Remove 'convert()'

NAMESPACE

@@ -4,16 +4,6 @@ S3method(clock_roll,Duration)
 S3method(clock_roll,Period)
 S3method(clock_roll,difftime)
 S3method(clock_roll,hms)
-S3method(convert,Date)
-S3method(convert,Duration)
-S3method(convert,Interval)
-S3method(convert,POSIXt)
-S3method(convert,Period)
-S3method(convert,character)
-S3method(convert,data.frame)
-S3method(convert,difftime)
-S3method(convert,hms)
-S3method(convert,numeric)
 S3method(na_as,Date)
 S3method(na_as,Duration)
 S3method(na_as,POSIXct)
@@ -30,12 +20,6 @@ S3method(round_time,difftime)
 S3method(round_time,hms)
 export(assign_date)
 export(chronotype)
-export(convert)
-export(convert_pt)
-export(convert_pu)
-export(convert_tu)
-export(convert_ut)
-export(convert_uu)
 export(fd)
 export(gu)
 export(le_week)

NEWS.md

@@ -12,17 +12,29 @@
 * The DESCRIPTION file now requires a lubridate version `>= 1.7.9`. `vctrs` support was added in this version, fixing some issues related to rounding `Duration`, `Period`, and `Interval` objects. Click [here](https://github.com/tidyverse/lubridate/pull/871) to learn more.
 * The DESCRIPTION file now requires a dplyr version `>= 0.2`. That's when the `dplyr` start to import the pipe operator. Click [here](https://github.com/tidyverse/dplyr/blob/master/NEWS.md#piping) to learn more.
 * To avoid any unknown compatibility issues, all packages on imports now requires the latest version of them at the moment of this release.
+* `convert()` and all `convert_*()` functions were removed. See a dedicated note about this below.
 * `round_time()` is now a S3 generic.
 * `shortest_interval()` was renamed to `shorter_interval()`.
 * `shorter_interval()` and `longer_interval()` now return only `Interval` objects.
+* `assign_date()` now return only `Interval` objects.
 * `sum_time()` now have different arguments and was divided in two functions: `sum_time()` (for non-vectorized sums) and `vct_sum_time()` (for vectorized sums).
-* The `sd()` function was renamed to `sdu()`. That was a tough, but necessary, call. Although we tried to preserve the original authors naming pattern for the MCTQ functions, the name `sd` provokes a dangerous name collision with the widely used `stats::sd` (standard deviation) function. That's why we named it as `sdu`. This is the only exception, all the other `mctq` functions maintain a strong naming resemblance with the original authors naming pattern.
+* The `sd()` function was renamed to `sdu()`. See a dedicated note about this below.
 * Several functions were optimized.
 
 ## Fixed
 
 * Several typos were fixed in the documentation.
 
+## Note about removing `convert()`
+
+`convert()` was created considering the user experience (sleep and chronobiology scientists). Since most of them don't have much experience with R and that time can have different types of representations (e.g., decimal hours, radian), `convert()` helped by trying to transpose those difficulties, posing as an "universal translator" (🖖).
+
+After much thought and consideration, we believe that the `convert()` feature may be out of the `mctq` scope. It can maybe be part of another package (a `lubritime` package perhaps? 😄). Other `mctq` tools, like `shorter_interval()` and `sum_time()`, could also be a part of that package (but are really necessary in `mctq` for the time being). Hence, we decided to remove `convert()` and to instruct the user to check the [lubridate](https://lubridate.tidyverse.org/) and [hms](https://hms.tidyverse.org/) packages for parsing/conversion.
+
+## Note about renaming `sd()`
+
+That was a tough, but necessary, call. Although we tried to preserve the original authors naming pattern for the MCTQ functions, the name `sd` provokes a dangerous name collision with the widely used `stats::sd` (standard deviation) function. That's why we named it as `sdu`. This is the only exception, all the other `mctq` functions maintain a strong naming resemblance with the original authors naming pattern.
+
 
 # mctq 0.0.0.9000
 

R/assign_date.R

@@ -13,10 +13,10 @@
 #' ## Class requirements
 #'
 #' The `mctq` package works with a set of object classes specially created to
-#' hold time values. These classes can be found in the [hms][hms::hms-package]
-#' and [lubridate][lubridate::lubridate-package] package. If your data do not
-#' conform to the object classes required, you can use the `mctq`
-#' [mctq::convert()] function to convert it.
+#' hold time values. These classes can be found in the
+#' [lubridate][lubridate::lubridate-package] and [hms][hms::hms-package]
+#' packages. Please refer to those package documentations to learn more about
+#' them.
 #'
 #' ## `ambiguity` argument
 #'
@@ -119,8 +119,15 @@ assign_date <- function(start, end, ambiguity = 0) {
                               lower = 0, upper = 86400)
     checkmate::assert_choice(ambiguity, c(0, 24 , NA))
 
-    start <- flat_posixt(convert(start, "posixct", quiet = TRUE))
-    end <- flat_posixt(convert(end, "posixct", quiet = TRUE))
+    start <- start %>%
+        hms::as_hms() %>%
+        as.POSIXct() %>%
+        flat_posixt()
+
+    end <- end %>%
+        hms::as_hms() %>%
+        as.POSIXct() %>%
+        flat_posixt()
 
     out <- dplyr::case_when(
         is.na(start) | is.na(end) ~ lubridate::as.interval(NA),