Closed ychae closed 3 years ago
schematic should handle the generation of a google sheet when there is no manifest present and when there is a manifest present in a given dataset folder. schematic will return a url to the resulting google sheet. The DCA should only need to call schematic's populate manifest function, w/o having logic to decide whether an empty or pre-populated google sheet has to be generated.
@rrchai you can take a look at the updated PR 523 in schematic.
Instead of calling populateModelManifest, you should be able to call getModelManifest the same way as you do here.
The new files should be appended at the end of the manifest. Currently all fields other than filename and entityId are left blank (i.e. empty string).
To test in schematic for the dataset referenced above in this issue: schematic manifest -v INFO --config ./config-htan.yml get --data_type ScRNA-seqLevel1 --title scRNA-seqLevel1 -d syn23520241 --oauth --sheet_url
Output here: https://docs.google.com/spreadsheets/d/19BH8yUc8AtmVVPaYajdwUDMkVPOR-jmuNPm8WGP4TVs/edit#gid=0
@ychae thank you for the detailed bug description across these issues (e.g. #190, #192 and #193) :)
@milen-sage @rong I just tested and I see a lot blank rows and then once I scroll down I see the old file names. If I scroll over to the right though, I can see all of the SynapseIDs.
Blank file names to filled in old file names:
Filled in columns, including Synapse IDs for all files:
@ychae I see the biospecimen component there. Aren't these files single cell? If the files are single cell, could you try selecting single cell RNA seq in the template dropdown?
Biospecimen is not a file-based data type, and the expected behavior is different wrt to pulling updated list of files (i.e. files do not apply to Biospecimen data; e.g. the data model doesn't include the attribute Filename for Biospecimen).
@milen-sage it's really odd, but I selected the correct scRNA-seq Level 1 template and that's what it generated. Here's the screenshot of the dropdown values
:
@rrchai has accidentally submitted a Biospecimen manifest for the Vanderbilt folder while testing today. I restored Vanderbilt's manifest. @ychae could you test the app again?
Whoo-hoo! Looks good! Love it when it's an easy fix π
I can see the annotated old files and the filenames of the new files, with no annotations. Looks good to share with Marisol.
Thanks Rong! π Thanks Milen! π
@ychae Good catch!! And thank you @milen-sage for a quick fix.
I realized when I was testing submission function, I accidentally submitted some files from other project to the Vanderbilt scRNA_seq_level1. If I remember correctly, it only originally contains synapse_storage_manifest.csv
and FASTQ_raw.
@milen-sage Is it possible to restore the folder/delete other extra files?
Sorry about the mistake!
Yes, we can delete the unneeded entities.
closes the issue since it's been fixed. Feel free to reopen if the issue remains
Describe the bug From Marisol R at Vanderbilt:
One of the folders that contains the old (February) files and new (August) files are here: https://www.synapse.org/#!Synapse:syn23564801
To Reproduce Steps to reproduce the behavior:
Expected behavior The generated template should contain both the older and newer files that were uploaded for centers to annotate.
Additional Notes Ideally have this fix in place by early October 2021