automatic backing of sdata.tables

[ArneDefauw]

I was testing the new mutlitple tables feature on the main branch, and noticed that automatic backing of sdata.tables to the backing zarr store no longer works (or I fail to get it working). This is how it used to work:

import anndata as ad
import numpy as np
import pandas as pd
from spatialdata import SpatialData
import spatialdata
import os

path=""

n_cells = 100
n_genes = 50

# Creating random data
X = np.random.randn(n_cells, n_genes)
obs = pd.DataFrame(index=[f'cell_{i}' for i in range(n_cells)])
var = pd.DataFrame(index=[f'gene_{i}' for i in range(n_genes)])

adata = ad.AnnData(X=X, obs=obs, var=var)

adata.obs['cell_type'] = ['type1' if i < n_cells // 2 else 'type2' for i in range(n_cells)]
adata.var['highly_variable'] = np.random.choice([True, False], size=n_genes)


sdata=SpatialData()
sdata.write( os.path.join( path, "sdata_dummy.zarr" ) )

sdata.tables[ "dummy_table" ] = spatialdata.models.TableModel.parse( adata,
)

assert sdata.table is not None # this is ok

sdata=SpatialData.read( os.path.join( path, "sdata_dummy.zarr" ) )
sdata

assert sdata.table is not None # this gives assertionerror

Is this related to https://github.com/scverse/spatialdata/issues/496?

[LucaMarconato]

Hi, thanks for reporting. Yes it is related to the bug you mentioned. I will fix this over the weekend before the next release. You can call the internal functions that we use when calling write() for the moment. Sorry for the inconvenience.

[melonora]

@LucaMarconato bump