Batch conversion of timestamps

[mhasself]

Is there a fast way to convert between arrays of G3VectorTime and numpy float arrays? I find myself writing this a lot:

  g3times = core.G3VectorTime([core.G3Time(t * core.G3Units.s) for t in timestamps])

and that's no fun. I am hoping for something like:

  g3times = core.G3VectorTime(timestamps * core.G3Units.s)
  timestamps = np.array(g3times) / core.G3Units.s

[nwhitehorn]

You would want to write something like on lines 195-205 of G3Vector.cxx, which implements an efficient converter for integers. I might have some time to do it in the next few days, but you are more than welcome to beat me to it.

[nwhitehorn]

Should be all fixed now -- you can do fancy stuff like this:

In [3]: a = core.G3VectorTime(numpy.asarray([53452354453424453,6572334542367,723423235444,834534543535453223],dtype='int64'))

In [4]: core.G3VectorTime(numpy.asarray(a))
Out[4]: spt3g.core.G3VectorTime([09-Dec-1986:14:45:44.534244530, 01-Jan-1970:18:15:23.345423670, 01-Jan-1970:02:00:34.232354440, 15-Jun-2234:15:30:35.354532230])

And then you can get the numbers back:

In [10]: numpy.asarray(a)
Out[10]: 
array([ 53452354453424453,      6572334542367,       723423235444,
       834534543535453223], dtype=int64)

The numpy array is also editable in place:

In [10]: numpy.asarray(a)
Out[10]: 
array([ 53452354453424453,      6572334542367,       723423235444,
       834534543535453223], dtype=int64)

In [11]: numpy.asarray(a)[2] = 5

In [13]: print(a[2])
01-Jan-1970:00:00:00.000000050

In [14]: numpy.asarray(a)
Out[14]: 
array([ 53452354453424453,      6572334542367,                  5,
       834534543535453223], dtype=int64)

[mhasself]

This is great, thanks!