This fixes a regression introduced in #2127 where reference AnnData files (i.e. non-parsed files, added in Classic upload UX) have a default clustering X_umap fragment created. This ends up accidentally switching the user in the AnnData upload experience, even if they have existing normal files. Additionally, this fixes a logic bug where reference AnnData files never had their parse_status flag updated after uploading, meaning they could never be deleted normally by a user.
MANUAL TESTING
Note: there is an existing bug where reference AnnData files are still being validated. For testing, you will need to use a valid file.
Boot all services, sign in, and go to a new or empty study
Select the "Classic" upload UX and go to the AnnData tab
Upload a file (or use bucket path) and save, confirming the file saves
In a separate Rails console session, load the file and confirm that it did not create the X_umap cluster fragment:
BACKGROUND & CHANGES
This fixes a regression introduced in #2127 where reference AnnData files (i.e. non-parsed files, added in Classic upload UX) have a default clustering
X_umap
fragment created. This ends up accidentally switching the user in the AnnData upload experience, even if they have existing normal files. Additionally, this fixes a logic bug where reference AnnData files never had theirparse_status
flag updated after uploading, meaning they could never be deleted normally by a user.MANUAL TESTING
Note: there is an existing bug where reference AnnData files are still being validated. For testing, you will need to use a valid file.
X_umap
cluster fragment:file.ann_data_file_info.data_fragments => []
file.ann_data_file_info.obsm_key_names => []