Right now, I am using hyp3 for the following:
for a series of dates $t_0, t_1, ..., t_n$ and fixed spatial area (or tile) we produce a product for each date. Each product is independent of the others.
However, there is a process of aggregating these products in sequential order: take products $t_{j}$ and $t_{j-1}$ and then create a new product at $t_{j}$ incorporating information from the previous product corresponding to $t_{j-1}$.
We have an entrypoint that performs this agregation altogether i.e. <AGGREGATE ENTRYPOINT> --products <Product at t_0> <product at t_1> ... <product at t_n>.
I am trying to figure out map keyword as used here and am not sure if there is an easy way to do either of the above, i.e. process two "adjacent" products in time or aggregate all the products together over the given tile.
I think I might need to create a separate job and do the manual accounting of the products, but thought I would ask if there is a slicker way first.
Thanks in advance!
Right now, I am using hyp3 for the following:
for a series of dates$t_0, t_1, ..., t_n$ and fixed spatial area (or tile) we produce a product for each date. Each product is independent of the others.
However, there is a process of aggregating these products in sequential order: take products$t_{j}$ and $t_{j-1}$ and then create a new product at $t_{j}$ incorporating information from the previous product corresponding to $t_{j-1}$ .
We have an entrypoint that performs this agregation altogether i.e.
<AGGREGATE ENTRYPOINT> --products <Product at t_0> <product at t_1> ... <product at t_n>.I am trying to figure out
mapkeyword as used here and am not sure if there is an easy way to do either of the above, i.e. process two "adjacent" products in time or aggregate all the products together over the given tile.I think I might need to create a separate job and do the manual accounting of the products, but thought I would ask if there is a slicker way first.
Thanks in advance!