rnikutta / hypercubizer Goto Github PK

Contemplate whether this should be optional or the only mode.

Either multiprocessing or MPI, and with either do parallel I/O

E.g.:

$cat file.dat
# col 0        1            2        3           4
# lambda    i =  0         10        20          30
1.000E-02   1.819E-02   1.810E-02   1.782E-02   1.723E-02
1.500E-02   2.018E-02   2.008E-02   1.977E-02   1.912E-02
2.200E-02   2.053E-02   2.043E-02   2.012E-02   1.946E-02

Allow e.g. to read columns 1,2,3,4 and form a new parameter axis 'i' with values [0,30].

The values should be provided as arguments, not read from file (too ambiguous).

E.g. in CLUMPY fits files, the two HDUs 'imgdata' (Nx,Ny,Nlambda) and 'clddata' (Nx,Ny).
Requires that self.hypercubeshape be a sequence of shapes.