Updates (CRAN)

photobiology 0.9.12 (CRAN release)

Changes from the previous CRAN release 0.9.11 are:

Enhancements:

Rewrite all “astronomical” calculations using different algorithms. The interface has also changed to some extent. For the solar position calculations, the returned values are slightly different, usually in the order of tenths of a degree. For day length the differences are of up to a few minutes for the twilight definition "none" which used to be the default. With the current default "sunlight", at high latitudes, days are longer than before. For latitudes < 72 degrees day length values returned should be within 1 min of the true values. At higher latitudes the errors can be up to 10 min. The range of dates supported is now much wider. In addition the calculation of the times of sunrise, solar noon and sunset is many times faster than earlier. The same is true for the calculation of local solar time for any geographic coordinates. One additional value for unit.out is accepted: "days". All unit constants are now recognized both in singular and plural.

The main changes to the interface of the functions has to do with the changed default for twilight, and the deletion of the redundant parameters lat and lon. The geocode parameter remains, but if present, the address variable is now also copied to the output. Multiple dates can be supplied as a vector, and multiple geocodes as a data frame with multiple rows. The tz parameter can be either a single value, or a vector of the same length as the number of rows in the geocode data frame. Vectorization is optimized for performance: only the parts of the calculations that need to be recalculated are recalculated.

Another important change is to allow calculation of all integral quantities on re-scaled data, with a warning (they would earlier trigger errors). Ratios and similar quantities which are not affected by linear re-scaling no longer trigger warnings or errors.

Revise cbindspct() so that it stores the name of "idfactor" as an attribute and set the retrieved value of the attribute as default in subset2mspct(). This
makes the conversion from a long form spectral object back into a collection of spectra ‘automatic’ in many additional use cases.

Bugs fixed:

normalize() sets "scaled" attribute to FALSE.
fscale() sets "normalized" attribute to FALSE.

photobiologyInOut 0.4.12 (CRAN release)

Changes from the previous CRAN release 0.4.11 are:

Enhancements:

Add function read_uvspec_disort(), rename read_libradtran_vesa() to read_uvspec_disort_vesa().

Bugs fixed:

Fix error in CITATION file.

photobiologyLamps 0.4.2 (first CRAN release)

Spectral data included in a single source_mspct object. Vectors of collection members’ names are provided to facilitate indexing and subsetting.

photobiologyLEDs 0.4.1 (first CRAN release)

Spectral data included in a single source_mspct object. Vectors of collection members’ names are provided to facilitate indexing and subsetting.

ggpmisc 0.2.12 (CRAN release)

Changes from the previous CRAN release 0.2.10 are:

Enhancements:

Add stat_dens2d_labels(), a statistic that resets label values to “” by default, or any character string supplied as argument, in regions of a panel with high
density of observations. Add stat_den2d_filter(), a statistic that filters-out/filters-in observations in regions of a panel with high density of observations.
These two statistics are useful for labelling or highlighting observations in regions of a panel with low density. Both stats use a compute_panel function.
Add stat_den2d_filter_g(), a statistic that filters-out/filters-in observations in regions of a group with high density of observations. This statistics is useful for highlighting observations. It uses a compute_group function. They use internally MASS:kde2d to estimate densities and default values for parameters are adjusted dynamically based on the number of observations.

Improve documentation using ‘ggrepel’ version (>= 06.3). stat_dens2d_labels() was developed in collaboration with the author/maintainer of package ‘ggrepel’. We thank everybody who contributed to the discussion in Github, including the user who raised the issue in first instance.

Please, report any problems you encounter, and suggest possible changes or additions to the functionality!

NOTE: These versions of the packages are available through CRAN.

RANDOM.ORG – True Random Number Service

Source: RANDOM.ORG – True Random Number Service

In most situations pseudo-random numbers produced by computer software  (“random” number generators) are good enough as long as we are careful when choosing the seed for the generator. Sometimes, it can be even an advantage to be able to reproduce sequences of pseudo-random numbers by setting the seed value. Frequently, the seed is obtained from the clock of the computer, e.g. using the seconds or milliseconds digits from current time. This is still not truly random, as random numbers cannot be generated by any deterministic process. True random numbers can be only be generated by a random physical process.

The site random.org is a service which provides true random numbers for free (at least if below a quota). R package random provides an interface to this service.

Updates

photobiology 0.9.11.9003 (test release)

Rewrite all “astronomical” calculations using different algorithms. The interface has changed to some extent. For the solar position calculations, the returned values are slightly different, usually in the order of tenths of a degree. For day length the differences are of up to a few minutes for the twilight definition "none" which used to be the default. With the current default "sunlight", at high latitudes, days are longer than before. For latitudes < 72 degrees day length now returned are within 1 min of the true values. At higher latitudes the errors can be up to 10 min. The range of dates supported is now much wider. In addition the calculation of the times of sunrise, solar noon and sunset is many times faster than earlier. The same is true for the calculation of local solar time for any geographic coordinates. This edit was long overdue.

The main changes to the interface of the functions has to do with the changed default for twilight, and the deletion of the redundant parameters lat and lon. The geocode parameter remains, but if present the address variable is copied to the output. Multiple dates can be supplied as a vector, and multiple geocodes as a data frame with multiple rows. The tz parameter can be either a single value, or a vector of the same length as the number of rows in the geocode data frame. Vectorization is optimized for performance: only the parts of the calculations that need to be recalculated are recalculated.

Please, test and report any problems. I hope there are not many bugs after the testing done by myself already caught quite a few. However, being this a full rewrite from scratch of several functions, problems may be still lurking.

Please, report any problems you encounter, and suggest possible changes or additions to the functionality!

NOTE: These versions of the packages are available through the r4photobiology repository.