"no filtering according to a range restriction will take place". What does that mean exactly?
This debug trace output refers to a SAP internal feature used with data ageing in ECC on HANA and S/4 HANA.
You can safely ignore this for your use case.
By the way: partitioning by calendar week is likely a mistake. If the data between any two weeks is not tremendously different in the type and volume of entries, then choosing this many partitions likely will give you a much worse compression than necessary. And that of course will lead to higher memory consumption - which I assume was the reason for looking into partitioning in the first place.
Coming back to the original topic: if I got you right, then the filter specification in the model was done on a different column (SALES_DATE). This is something that SAP HANA cannot figure out itself.
If you want to exploit partition pruning but also want to allow users to provide filters based on other time columns you have to perform a mapping to an explicit filter in your model.