Tagged data from a DataArray with polynom_coefficients and "alias" RangeDimension

[refactoriel]

Haha, I get the feeling that I have weird use cases....

The following should return spike times in seconds, but, alas, they're saved as ints and the ticks of the range dimension is only using the raw linked data.

np.random.seed(1234)
timestamps = np.sort(np.random.choice(np.arange(1000000, dtype=int), 10000, replace=False))
fs = 40000

f = nixio.File.open("test.nix", "w")
b = f.create_block("block0", "foo")

da_ts = b.create_data_array("spike_times", "foo", dtype=np.uint64, data=timestamps)
da_ts.polynom_coefficients = (0, 1 / fs) 
da_ts.unit = "s" 
da_ts.label = "time"
dim = da_ts.append_range_dimension()
dim.link_data_array(da_ts, [-1])

pos = b.create_data_array("pos", "foo", data=[3.0, 4.0])
ext = b.create_data_array("ext", "foo", data=[0.4, 0.4])
mt = b.create_multi_tag("mt", "foo", positions=pos)
mt.extents = ext 
mt.references.append(da_ts)
# it should return the scaled values
print(da_ts[np.logical_and(da_ts[:] >= 3., da_ts[:] <= 3.4)])
try:
    print("try...")
    # but it doesn't
    print(mt.tagged_data(0, "spike_times")[:])
    print("success!")
except:
    print("aliasrangedim failed")
    # a normal range dim works, because the timestamps are stored twice
    da_ts.delete_dimensions()
    dim = da_ts.append_range_dimension(ticks=da_ts[:], label="time", unit="s")
    print(mt.tagged_data(0, "spike_times")[:])

I guess if I have to save the timestamps twice it's more efficient to actually use floats and and then create the alias range dimension...

Might be related to #482