Skip to content

source product version and slicing checks #233

@johntruckenbrodt

Description

@johntruckenbrodt

Currently the processor does not do anything to filter duplicate source products. It often happens that source products (GRD/SLC) are reprocessed and, if the data archive is not optimally curated, multiple versions of the same acquisition may exist. In some cases the reprocessed products are also resliced so that both versions may not easily be compared by the acquisition times.

Consider the following case. A data take of products was first processed with manifest.safe metadata attribute sliceNumber=0, then reprocessed with sliceNumber>0 using the same slicing, and later reprocessed again this time using a different slicing and sliceNumber=0 again. Hence, three versions of the same data take may exist. At the time of writing, ASF only offers the version processed last and CDSE offers all three versions. In case CDSE is used as reference archive, s1ard has no way of determining which are the right products to use.

S1A_IW_GRDH_1SDV slice_num pref. ASF CDSE
*_20231207T064644_20231207T064709_051545_0638F1_9569 0 x x
*_20231207T064649_20231207T064714_051545_0638F1_4B74 6 x x
*_20231207T064649_20231207T064714_051545_0638F1_95E7 0 x
*_20231207T064709_20231207T064734_051545_0638F1_440B 0 x x
*_20231207T064714_20231207T064739_051545_0638F1_4799 7 x x
*_20231207T064714_20231207T064739_051545_0638F1_AC9A 0 x
*_20231207T064734_20231207T064759_051545_0638F1_9500 0 x x
*_20231207T064739_20231207T064804_051545_0638F1_C7A2 8 x x
*_20231207T064739_20231207T064804_051545_0638F1_FECD 0 x
*_20231207T064759_20231207T064824_051545_0638F1_8191 0 x x

The reprocessed product with the 0 slice numbers is the one that was processed last.
Also, all products have the s1sarl1:productTimelinessCategory value NRT-3h and thus cannot be differentiated by it:

S1A_IW_GRDH_1SDV slice_num proc.time timeliness
*_20231207T064739_20231207T064804_051545_0638F1_FECD 0 20231207T085207 NRT-3h
*_20231207T064649_20231207T064714_051545_0638F1_95E7 0 20231207T085214 NRT-3h
*_20231207T064714_20231207T064739_051545_0638F1_AC9A 0 20231207T085228 NRT-3h
*_20231207T064649_20231207T064714_051545_0638F1_4B74 6 20231207T155100 NRT-3h
*_20231207T064739_20231207T064804_051545_0638F1_C7A2 8 20231207T155140 NRT-3h
*_20231207T064714_20231207T064739_051545_0638F1_4799 7 20231207T155151 NRT-3h
*_20231207T064644_20231207T064709_051545_0638F1_9569 0 20231207T171137 NRT-3h
*_20231207T064734_20231207T064759_051545_0638F1_9500 0 20231207T171139 NRT-3h
*_20231207T064709_20231207T064734_051545_0638F1_440B 0 20231207T171145 NRT-3h
*_20231207T064759_20231207T064824_051545_0638F1_8191 0 20231207T171149 NRT-3h

So, ASF just keeps the products that were processed last. In this case this might not suffice. I have opened an issue at ASF about this.

Metadata

Metadata

Assignees

No one assigned

    Labels

    enhancementNew feature or request

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions