GitLab

It is not the size of a grid, so the name was a bit confusing.

The basic building blocks are now described in a way that should be
just enough to use...
Unfortunately I can't sink hours into this so it will have to grow as
the features expand.

No numerical testing is being performed so far (this is left for a time
when there is more to test...) but now it's unified and clean.

(when given a parameter name instead of an index)

... and make one of the new keys optional.

This only contains the latest version of MeteoIO compiled with netcdf and proj.
There is no web service yet!

Updated to a newer version of getopt that contains getopt_long and fixes the warnings experienced on VC++. Fixed a vc++ warning in RandomNumberGenerator.

Warnings such as warning: implicitly-declared
‘mio::Grid1DInterpolator::Grid1DInterpolator(const mio::Grid1DInterpolator&)’
is deprecated [-Wdeprecated-copy]

Warnings such as warning: implicitly-declared
‘mio::Meteo1DInterpolator::Meteo1DInterpolator(const mio::Meteo1DInterpolator&)’
is deprecated [-Wdeprecated-copy]

... to grid resampling files. Also some comments

Now dates before and after data span are (hopefully) enabled correctly.
It was tested with LinearResampling's EXTRAPOLATE keyword.

... meaning points inbetween valid other data points.

Once again, this is completely untested so far.

It does not look like the best spot to print this message... but for now
as long as all 2D interpolations return an info string it should be ok,
and it will not be printed if the data is read directly from disk
without interpolation.
In any case, it's cosmetics only.

With these, most occasions should already be covered (empty grid list,
only 1 grid, requested date outside loaded grids).
Also switch TZ to 0 for clarity.

Also tweak control program to produce more convenient test times.

Some work to make the grid getMeteoData() function return data.
Extend idea of GridTimeseriesResampling, but this remains untested.

This is a sketch for how to use meteo timeseries 1d
interpolations for grids. It compiles, but otherwise is completely
untested so far, just an idea.
Minor cleanup all around grid resampling.

The processing chain through an algorithm object factory has been
more or less implemented, and a test program for it was added.
This is an itermediate commit to build up upon.
Minor unrelated tweaks.

In a test case provided by Charles, aggregating as the maximum was wrong
when the aggregation period was a multiple of the sampling rate.

As IOUtils::seek() returns either an exact match OR the next element
greater than the provided date, the get_window_specs() method was
inaccurately assuming that it was always the next element and always
substracting 1 to be sure to contain the start point of interest. Now
the case of exact match is properly handled and skips the substracting 1
in this case.

Added a missing include

Removing the COPYING files because gitlab is too dumb to accept them when looking for the project license

Making the TRANSFORMWINDVECTOR filter accept both an arbitrary source (key: COORDPARAM_SRC) as well as destination coordinate system (key: COORDPARAM) for more flexibility. It can now also read a PROJ4 string from a file, if a filename is provided.

These future evolutions would be the temporal interpolations of gridded data.