Rename SpatialData elements

[Fritze]

Is your feature request related to a problem? Please describe. The spatialdata_io.merscope reader has the two arguments 'region_name' and 'slide_name'. If set to None, file path directory is used. If set to a string, the order will always be slide_region_element. This can make element names more complicated and longer than needed.

Describe the solution you'd like A renaming function that works on individual elements within the SpatialData would be great, so e.g. spatialdata.rename(elements={"slide_region_transcripts": "just_transcripts"})

Describe alternatives you've considered One can re-built a SpatialData object from scratch by copying from the old one and thereby fix the names (described here). But ideally, one would do this without copying the data.

[sophiamaedler]

+1 this would be a useful function

[pranavmishra90]

Would also appreciate having the ability to easily rename tables. When trying to do a sd.concatenate([sdata1, sdata2]), you may run into a naming issue like so:

SpatialData object
├── Images
│     ├── '20004_cytassist_image': DataArray[cyx] (3, 2997, 3200)
│     ├── '20004_full_image': DataTree[cyx] (3, 19007, 71119), (3, 9503, 35559), (3, 4751, 17779), (3, 2375, 8889), (3, 1187, 4444)
│     ├── '20004_hires_image': DataArray[cyx] (3, 1604, 6000)
│     ├── '20004_lowres_image': DataArray[cyx] (3, 161, 600)
│     ├── '40001_cytassist_image': DataArray[cyx] (3, 3000, 3200)
│     ├── '40001_full_image': DataTree[cyx] (3, 16755, 45808), (3, 8377, 22904), (3, 4188, 11452), (3, 2094, 5726), (3, 1047, 2863)
│     ├── '40001_hires_image': DataArray[cyx] (3, 2195, 6000)
│     └── '40001_lowres_image': DataArray[cyx] (3, 220, 600)
├── Shapes
│     ├── '20004_square_002um': GeoDataFrame shape: (3399493, 1) (2D shapes)
│     ├── '20004_square_008um': GeoDataFrame shape: (433722, 1) (2D shapes)
│     ├── '20004_square_016um': GeoDataFrame shape: (151013, 1) (2D shapes)
│     ├── '40001_square_002um': GeoDataFrame shape: (3813253, 1) (2D shapes)
│     ├── '40001_square_008um': GeoDataFrame shape: (467052, 1) (2D shapes)
│     └── '40001_square_016um': GeoDataFrame shape: (160661, 1) (2D shapes)
└── Tables
      ├── 'square_002um_0': AnnData (3399493, 32285)
      ├── 'square_002um_1': AnnData (3813253, 32285)
      ├── 'square_008um_0': AnnData (433722, 32285)
      ├── 'square_008um_1': AnnData (467052, 32285)
      ├── 'square_016um_0': AnnData (151013, 32285)
      └── 'square_016um_1': AnnData (160661, 32285)

In my projects, I've used a naming schema where the first 5 numbers represents a particular sample (mouse or human subject ID). These are then used downstream for subsetting the sdata object (e.g. selecting a particular table with f{selected_sample}_square_{bin_size}um. This functionality will break, as there is an incremental integer appended to the name of the table. If we could rename / transform this element, it would be excellent

I've been hacking my way around this, but it would be a nice feature to add