preliminary commit

ok preliminary commit to use scverse-airr structure as per #261

TODO:

- wherever *.data is mentioned in dandelion, this needs to handle the awkward arrays
- create a slot/function that stores the cell_id
- update the tutorials
- create a new to_scirpy/from_scirpy function that basically transfers the .obs, .obsm and .uns
- create a to/from_ak function so that the rearrangement data can still be repopulated as a pandas dataframe.
- would be cool if i can also make this dask compatible - only issue is dask.dataframe.DataFrame.compute is always slow..? need to learn more about dask. vaex is also cool but not really compatible for contig->cell level wrangling.

dandelion/tools/_tools.py

@@ -634,7 +634,7 @@ def find_clones(
                 layout=layout_,
                 graph=graph_,
             )
-            vdj_data.update_metadata(reinitialize=True)
+            vdj_data.update_metadata(initialize=True)
         elif ("clone_id" in vdj_data.data.columns) and (key_added is not None):
             vdj_data.__init__(
                 data=dat_,
@@ -643,7 +643,7 @@ def find_clones(
                 graph=graph_,
             )
             vdj_data.update_metadata(
-                reinitialize=True,
+                initialize=True,
                 clone_key="clone_id",
                 retrieve=clone_key,
                 retrieve_mode="merge and unique only",
@@ -656,7 +656,7 @@ def find_clones(
                 graph=graph_,
                 clone_key=clone_key,
             )
-            vdj_data.update_metadata(reinitialize=True, clone_key=clone_key)
+            vdj_data.update_metadata(initialize=True, clone_key=clone_key)
         vdj_data.threshold = threshold_
 
     else: