This is a maintenance release that fixed a failing test.

This version is presented of aweek is presented as a more mature and stable version of aweek:

• combining objects is more predictable
• weeks/dates can be created from vectors with heterogeneous week_starts
• documentation is improved
• data frames can be created

NEW FUNCTIONS

• change_week_start() allows the user to change the week_start attribute of
  an aweek object, adjusting the weeks to match the new attribute accordingly.
• get_aweek() can generate aweek objects from a vector of week numbers. It has
  the ability to take into account different week start times.
• get_date() is similar to get_aweek(), but returns dates instead.
• as.aweek() allows users to create aweek object directly from characters with
  validation. It also allows for dates by passing to date2week().
• as.data.frame.aweek() is a new function that allows aweek objects to be
  directly incorporated into data frames.
• as.list.aweek() will now preserve the aweek structure in lists
• trunc.aweek() will truncate the day to the first day of the week.
• rep.aweek() allows repeating aweek characters.
• factor_aweek() allows the user to create aggregated aweek objects on the fly.

BREAKING CHANGES

There are a couple of breaking changes coming to aweek that will improve
stability by removing unclear coercion methods
(see #20).

• It is no longer possible to combine aweek objects if they do not share the
  same week_start attribute. This will result in an error informing the user
  to adjust the week_start attribute with the change_week_start() function.
• Factors will no longer coerce factors when combining aweek objects.
• Using date2week() with factor = TRUE and floor_day = FALSE now throws an
  error instead of a warning (as prophesized in #13).

DOCUMENTATION

• The vignette has been updated to include an explanation of the underlying
  week calculation from date.
• The aweek class documentation has been updated to detail the different
  elements of the object and the calculations.
• Package documentation package?aweek has been added for an introduction.

BUG FIX

• conversions will now retain the names of the object.

MISC

• More checks that weekdays and weeks are within the correct bounds have been
  added.
• as.Date.aweek has been simplified to no longer rely on regex since the aweek
  class is standard.
• The internal get_wday() has been vastly simplified with improved speed.

This version of aweek introduces the flexible global week_start option, "aweek.week_start". This allows users to set the following line at the top of their scripts or in their .Rprofile to set the default start of the week:

aweek::set_week_start("Monday")

This is equivalent of setting options(aweek.week_start = 1L) and will have the side effect of setting the default week_start option for both date2week() and week2date(), making it easier to maintain scripts using aweek.

This release fixes a bug where NA dates were not properly processed (#12)

This release fixes a test, allows subsetting and concatenation, and enforces stricter guidelines on character inputs.

• Subsetting and concatenating methods added to the aweek class (see #1)
• Documentation divided into smaller chunks
• as.POSIXlt() bug where tz was not being passed was fixed.
• date2week(): an error is now issued if users specify non-ISO 8601 dates OR
  don't specify a format option. (found: @scottyaz, #2)
• Best practices added to vignette
• Fix test that would fail every seven days on CRAN (see #5)

This is the initial CRAN version of aweek